VOLUME 52

NUMBER 8

IETPAK

(ISSN 0018-926X)

EDITORIAL

**A Note From the Outgoing Editor-in-Chief . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. W. Glisson
**

PAPERS

1926

Bandwidth Enhancement and Further Size Reduction of a Class of Miniaturized Slot Antennas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. Behdad and K. Sarabandi Miniature Built-In Multiband Antennas for Mobile Handsets . . . . . . . . . . . . . Y.-X. Guo, M. Y. W. Chia, and Z. N. Chen Miniature Reconﬁgurable Three-Dimensional Fractal Tree Antennas. . . . . . . . . . . . . . . . . J. S. Petko and D. H. Werner Investigations on Miniaturized Endﬁre Vertically Polarized Quasi-Fractal Log-Periodic Zigzag Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. K. Sharma and L. Shafai Compact Wide-Band Multimode Antennas for MIMO and Diversity . . . . . . . . . . . . . . C. Waldschmidt and W. Wiesbeck Ground Inﬂuence on the Input Impedance of Transient Dipole and Bow-Tie Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. A. Lestari, A. G. Yarovoy, and L. P. Ligthart Adaptive Crossed Dipole Antennas Using a Genetic Algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .R. L. Haupt Modeling and Investigation of a Geometrically Complex UWB GPR Antenna Using FDTD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .K.-H. Lee, C.-C. Chen, F. L. Teixeira, and R. Lee Radiation Properties of an Arbitrarily Flanged Parallel-Plate Waveguide . . . . . . D. N. Chien, K. Tanaka, and M. Tanaka Scan Blindness Free Phased Array Design Using PBG Materials. . . . . L. Zhang, J. A. Castaneda, and N. G. Alexopoulos Fractile Arrays: A New Class of Tiled Arrays With Fractal Boundaries . . . . D. H. Werner, W. Kuhirun, and P. L. Werner A New Millimeter-Wave Printed Dipole Phased Array Antenna Using Microstrip-Fed Coplanar Stripline Tee Junctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Y.-H. Suh and K. Chang Physical Limitations of Antennas in a Lossy Medium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Karlsson Minimum Norm Mutual Coupling Compensation With Applications in Direction of Arrival Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .C. K. E. Lau, R. S. Adve, and T. K. Sarkar A Phase-Space Beam Summation Formulation for Ultrawide-band Radiation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Shlivinski, E. Heyman, A. Boag, and C. Letrou

1928 1936 1945 1957 1963 1970 1976 1983 1992 2000 2008 2019 2027 2034 2042

(Contents Continued on Page 1925)

(Contents Continued from Front Cover) Theoretical Considerations in the Optimization of Surface Waves on a Planar Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. F. Mahmoud, Y. M. M. Antar, H. F. Hammad, and A. P. Freundorfer Generalized System Function Analysis of Exterior and Interior Resonances of Antenna and Scattering Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L. Li and C.-H. Liang MIMO Wireless Communication Channel Phenomenology. . . . . . . . . . . . . . .D. W. Bliss, A. M. Chan, and N. B. Chang Service Oriented Statistics of Interruption Time Due to Rainfall in Earth-Space Communication Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Matricciani Full-Wave Analysis of Dielectric Frequency-Selective Surfaces Using a Vectorial Modal Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Coves, B. Gimeno, J. Gil, M. V. Andrés, A. A. San Blas, and V. E. Boria On the Interaction Between Electric and Magnetic Currents in Stratiﬁed Media . . . . . D. Llorens del Río and J. R. Mosig Scattering by Arbitrarily-Shaped Slots in Thick Conducting Screens: An Approximate Solution . . . . . . . . . . J. R. Mosig Double Higher Order Method of Moments for Surface Integral Equation Modeling of Metallic and Dielectric Antennas ´ s and Scatterers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. Djordjevic and B. M. Notaroˇ Loop-Tree Implementation of the Adaptive Integral Method (AIM) for Numerically-Stable, Broadband, Fast Electromagnetic Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . V. I. Okhmatovski, J. D. Morsey, and A. C. Cangellaris A Single-Level Low Rank IE-QR Algorithm for PEC Scattering Problems Using EFIE Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. M. Seo and J.-F. Lee Accelerated Gradient Based Optimization Using Adjoint Sensitivities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. K. Nikolova, R. Saﬁan, E. A. Soliman, M. H. Bakr, and J. W. Bandler A Theoretical Study of the Stability Criteria for Hybridized FDTD Algorithms for Multiscale Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. Marrone and R. Mittra

COMMUNICATIONS

2057 2064 2073 2083 2091 2100 2109 2118 2130 2141 2147 2158

Full-Wave Analysis of a Waveguide Printed Slot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Montisci and G. Mazzarella Dual Polarized Wide-Band Aperture Stacked Patch Antennas . . . . . . . . . . . . . . . . . .K. Ghorbani and R. B. Waterhouse Resonant Frequency of Equilateral Triangular Microstrip Antenna With and Without Air Gap. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D. Guha and J. Y. Siddiqui Effect of a Cavity Enclosure on the Resonant Frequency of Inverted Microstrip Circular Patch Antenna. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D. Guha and J. Y. Siddiqui Design and Development of Multiband Coaxial Continuous Transverse Stub (CTS) Antenna Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R. Isom, M. F. Iskander, Z. Yun, and Z. Zhang Near-Field, Spherical-Scanning Antenna Measurements With Nonideal Probe Locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R. C. Wittmann, B. K. Alpert, and M. H. Francis Resonance Series Representation of the Early-Time Field Scattered by a Coated Cylinder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. Vollmer and E. J. Rothwell High Order Symplectic Integration Methods for Finite Element Solutions to Time Dependent Maxwell Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .R. Rieben, D. White, and G. Rodrigue

CORRECTIONS

2168 2171 2174 2177 2180 2184 2186 2190

**Corrections to “Phased Arrays Based on Oscillators Coupled on Triangular and Hexagonal Lattices” . . . R. J. Pogorzelski
**

CALLS FOR PAPERS

2196

Special Issue on Multifunction Antennas and Antenna Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2197

**IEEE ANTENNAS AND PROPAGATION SOCIETY
**

All members of the IEEE are eligible for membership in the Antennas and Propagation Society and will receive on-line access to this TRANSACTIONS through IEEE Explore upon payment of the annual Society membership fee of $24.00. Print subscriptions to this TRANSACTIONS are available to Society members for an additional fee of $50.00. For information on joining, write to the IEEE at the address below. Member copies of Transactions/Journals are for personal use only. J. L. VOLAKIS, President 2003 S. D. GEDNEY J. S. HERD A. Q. MARTIN *D. H. SCHAUBERT P. F. WAHID 2004 W. C. CHEW *T. MILLIGAN J. ROCKWAY J. A. MOELLERS J. L. YOUNG ADMINISTRATIVE COMMITTEE R. W. ZIOLKOWSKI, Vice President 2005 T. S. BIRD S. C. HAGNESS M. THORBURN N. ENGHETA *P. L. E. USLENGHI Honorary Life Member: R. C. HANSEN *Past President Committee Chairs and Representatives Antenna Measurements (AMTA): E. S. GILLESPIE Applied Computational EM Society (ACES): A. F. PETERSON Asian Representative: H. NAKANO Awards and Fellows: A. SCHELL CAEME Board: D. H. SCHAUBERT Chapter Activities: D. R. JACKSON Committee on Communications and Information Policy: A. I. ZAGHLOUL Committee on Man and Radiation: R. D. NEVELS Computer Applications in EM Education (CAEME): M. F. ISKANDER Constitution and Bylaws: R. I. WOLFSON Distinguished Lecturers: P. H. PATHAK Education: S. R. RENGARAJAN EAB/TAAC: E. K. MILLER Energy Committee: J. F. LINDSEY, III European Representatives: P. J. B. CLARRICOATS, A. G. ROEDERER AP Transactions web site: http://aps.ee.olemiss.edu Finance: M. SHIELDS Historian: K. STEPHAN IEEE Magazine Committee: W. R. STONE IEEE Press Liaison: R. J. MAILLOUX IEEE Transactions Committee: P. L. E. USLENGHI International Radio Consultative Cmte (CCIR): E. K. SMITH Institutional Listings: C. C. ALLEN Joint Committee on High-Power Electromagnetics: C. E. BAUM Long-Range Planning: A. C. SCHELL Magazine Editor: W. R. STONE Meetings: S. A. LONG Meetings Coordination: S. A. LONG Meetings Workshops: W. G. SCOTT Membership: S. D. GEDNEY New Technology: A. I. ZAGHLOUL Nominations: L. W. PEARSON Professional Activities (PACE): J. MICHAEL JOHNSON Publications: R. J. MARHEFKA RAB/TAB Transnational Committee Liaison: P. S. KILDAL R&D Committee—Aerospace: G. HYDE R&D Committee—Defense: A. C. SCHELL R&D Committee—Engineering: A. C. SCHELL Society of Social Implications of Technology: M. J. PIKET-MAY Standards Board: M. H. FRANCIS, D. V. THIEL Standards—Antennas: M. H. FRANCIS Standards—Propagation: D. V. THIEL Superconductivity Committee: K. MEI, J. WILLIAMS TAB Magazines Committee: W. R. STONE TAB New Technology Directions Committee: A. I. ZAGHLOUL TAB Public Relations Committee: W. R. STONE TAB Transactions Committee: P. L. E. USLENGHI USAB Aerospace Policy Committee: G. HYDE USAB Defense R&D Committee: A. C. SCHELL USAB Enginerring R&D Committee: A. C. SCHELL Transnational Committee: J. MOSIG USNC/URSI: P. L. E. USLENGHI M. SHIELDS, Secretary-Treasurer 2006 M. ANDO J. T. BERNHARD *M. F. ISKANDER M. SALAZAR-PALMA D. H. WERNER

AP Transactions Manuscript Central web site: http://tap-ieee.manuscriptcentral.com Chapter Chairs

For current Chapter Chairs information, visit the APS website (http://www.ieee.aps.org) and look under “Local Chapters.”

**IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION
**

is a publication devoted to theoretical and experimental advances in antennas including design and development, and in the propagation of electromagnetic waves including scattering, diffraction, and interaction with continuous media; and applications pertinent to antennas and propagation, such as remote sensing, applied optics, and millimeter and submillimeter wave techniques. See inside back cover for Editorial Board.

**THE INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS, INC.
**

Ofﬁcers

ARTHUR W. WINSTON, President MICHAEL R. LIGHTNER, Vice President, Publication Services and Products MARC T. APTER, Vice President, Regional Activities W. CLEON ANDERSON, President-Elect JAMES T. CARLO, President, IEEE Standards Association MOHAMED EL-HAWARY, Secretary PEDRO A. RAY, Treasurer RALPH W. WYNDRUM, JR., Vice President, Technical Activities MICHAEL S. ADLER, Past President JOHN W. STEADMAN, President, IEEE-USA JAMES M. TIEN, Vice President, Educational Activities HAROLD L. FLESCHER, Director, Division IV — Electromagnetics and Radiation

Executive Staff

DANIEL J. SENESE, Executive Director DONALD CURTIS, Human Resources ANTHONY DURNIAK, Publications Activities JUDITH GORMAN, Standards Activities CECELIA JANKOWSKI, Regional Activities BARBARA COBURN STOLER, Educational Activities MATTHEW LOEB, Corporate Strategy & Communications RICHARD D. SCHWARTZ, Business Administration W. THOMAS SUTTLE, IEEE-USA MARY WARD-CALLAN, Technical Activities SALLY A. WASELIK, Information Technology

**IEEE Periodicals Transactions/Journals Department
**

Staff Director: FRAN ZAPPULLA Editorial Director: DAWN MELLEY Production Director: ROBERT SMREK Managing Editor: WILLIAM A. COLACCHIO Associate Editor: DAWN L. MENENDEZ

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION (ISSN 0018-926X) is published monthly by the Institute of Electrical and Electronics Engineers, Inc. Responsibility for the contents rests upon the authors and not upon the IEEE, the Society/Council, or its members. IEEE Corporate Ofﬁce: 3 Park Avenue, 17th Floor, New York, NY 10016-5997. IEEE Operations Center: 445 Hoes Lane, P.O. Box 1331, Piscataway, NJ 08855-1331. NJ Telephone: +1 732 981 0060. Price/Publication Information: Individual copies: IEEE Members $20.00 (ﬁrst copy only), nonmembers $50.00 per copy. (Note: Postage and handling charge not included.) Member and nonmember subscription prices available upon request. Available in microﬁche and microﬁlm. Copyright and Reprint Permissions: Abstracting is permitted with credit to the source. Libraries are permitted to photocopy for private use of patrons, provided the per-copy fee indicated in the code at the bottom of the ﬁrst page is paid through the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. For all other copying, reprint, or republication permission, write to Copyrights and Permissions Department, IEEE Publications Administration, 445 Hoes Lane, P.O. Box 1331, Piscataway, NJ 08855-1331. Copyright © 2004 by The Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Periodicals Postage Paid at New York, NY and at additional mailing ofﬁces. Postmaster: Send address changes to IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, IEEE, 445 Hoes Lane, P.O. Box 1331, Piscataway, NJ 08855-1331. GST Registration No. 125634188. Printed in U.S.A.

Digital Object Identiﬁer 10.1109/TAP.2004.835169

1926

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**A Note from the Outgoing Editor-in-Chief
**

T HAS BEEN an honor and a privilege for me to serve as the Editor-in-Chief of the IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION for the past three years. I want to express my sincere thanks to the Antennas and Propagation (AP) Society Administrative Committee for giving me the opportunity to serve the Society in this capacity. Although I was initially reluctant to accept the great responsibility that goes with this position, it has been a unique and rewarding experience. The TRANSACTIONS is the leading journal in its ﬁeld and the AP Society is justiﬁably proud of it. I can only hope that we have been successful during my tenure as Editor-in-Chief (EIC) in maintaining the outstanding quality of the Transactions that APS members expect and deserve. During the last three years the annual number of pages published in the TRANSACTIONS has been increased from 1,960 pages in 2001 to 3,400 pages in 2003 to decrease the publication backlog. Page budgets that are required to be set far in advance along with an increasing number of submissions have made it difﬁcult to reduce the publication backlog. A special thanks is due to the APS AdCom for authorizing an extraordinary expenditure of funds to increase the number of pages published in 2003 to help reduce the backlog. Additionally, the IEEE has recently approved more ﬂexible rules with regard to page budgets that will hopefully make it easier to avoid large backlogs in the future. The transition to an all-electronic process from submission to publication is also well underway. Galley proof delivery to authors has become electronic. Corrections can be noted on the galley proof ﬁles with Adobe Acrobat software and returned electronically as well. A new electronic copyright form has recently been included in Manuscript Central. The promise of a completely paper-free process is expected to be achieved soon, possibly within the next year. When you have the chance, please thank Wilson Pearson, the previous EIC, and Anthony Martin for leading the way in implementing the electronic submission process. As I turn over the reins to the new Editor-in-Chief, Dr. Trevor S. Bird, I want express my deepest thanks to all of you who have served as reviewers for the TRANSACTIONS over the past three years. The review process is critical to maintaining the quality of the Transactions. I also particularly want to thank all the Associate Editors who have worked so hard. Their names are still listed on the inside back cover in this issue, but I list them here again to emphasize their outstanding service: • Jørgen Bach Andersen • Yahia Antar • Jennifer Bernhard • Trevor Bird • Filippo Capolino

I

• Lawrence Carin • Christos Chrisodoulou • Cynthia Furse • Stephen Gedney • George Hanson • Michael Jensen • Leo Kempel • Chi Chung Ko • Karl Langenberg • Louis Medgyesi-Mitschang • Kathleen Melde • Krzysztof Michalski • Eric Michielssen • Michal Okoniewski • Hsueh-Yuan Pao • Sembiam Rengarajan • Antoine Roederer • Kamal Sarabandi • Ari Sihvola • Rainee Simons • Parveen Wahid Thanks are also due to the Editors of the special issues that have appeared and are currently in preparation: Magdy Iskander and Jim Mink, for the Special Issue on Wireless Information Technology and Networks; Rick Ziolkowski and Nader Engheta, for the Special Issue on Metamaterials; and Per-Simon Kildal, Ahmed Kishk, and Stefano Maci, for the upcoming Special Issue on Artiﬁcial Magnetic Conductors, Soft/Hard Surfaces, and other Complex Surfaces. Special thanks are also extended to former Associate Editor Roena Rabelo Vega and current Associate Editor Dawn L. Menendez at IEEE Headquarters who have worked hard on producing a quality product. Dawn, in particular, has done a tremendous job in getting our publication schedule back on track. Finally, I want to give a special thanks to Sharon Martinez, who has served admirably as my Editorial Assistant. She has kept us organized, worked to keep Reviewers and Associate Editors on schedule, answered author questions, and generally kept things running efﬁciently. Without her help your Editor’s Ofﬁce would have dissolved into chaos after the ﬁrst few months. In closing, I again thank the AP Society for the opportunity to serve, and I hope that you will all support Trevor as he takes on his new role. He has done a superb job as an Associate Editor and I know he is committed to serving the AP Society and the TRANSACTIONS. There is no doubt that the TRANSACTIONS will be in good hands.

Digital Object Identiﬁer 10.1109/TAP.2004.834953

ALLEN W. GLISSON, Outgoing Editor-in-Chief The University of Mississippi Department of Electrical Engineering University, MS 38677-1848 USA

0018-926X/04$20.00 © 2004 IEEE

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

1927

Allen W. Glisson (S’71–M’78–SM’88–F’02) received the B.S., M.S., and Ph.D. degrees in electrical engineering from the University of Mississippi, in 1973, 1975, and 1978, respectively. In 1978, he joined the faculty of the University of Mississippi, where he is currently a Professor and Chair of the Department of Electrical Engineering. His current research interests include the development and application of numerical techniques for treating electromagnetic radiation and scattering problems, and modeling of dielectric resonators and dielectric resonator antennas. He has been actively involved in the areas of numerical modeling of arbitrarily shaped bodies and bodies of revolution with surface integral equation formulations. He has also served as a consultant to several different industrial organizations in the area of numerical modeling in electromagnetics. Dr. Glisson is a Member of Sigma Xi Research Society and the Tau Beta Pi, Phi Kappa Phi, and Eta Kappa Nu Honor Societies. He is a Member of several professional societies within the IEEE, Commission B of the International Union of Radio Science (URSI), and the Applied Computational Electromagnetics Society. He was a U.S. delegate to the 22nd, 23rd, and 24th General Assemblies of URSI. He was selected as the Outstanding Engineering Faculty Member in 1986, 1996, and 2004. He received a Ralph R. Teetor Educational Award in 1989 and the Faculty Service Award in the School of Engineering in 2002. He received a Best Paper Award from the SUMMA Foundation and twice received a citation for excellence in refereeing from the American Geophysical Union. He is the recipient of the 2004 Microwave Prize awarded by the Microwave Theory and Techniques Society. He has served as a member of the IEEE Antennas and Propagation Society Administrative Committee and is currently a member of the IEEE Press Liaison Committee. He currently serves on the Board of Directors of the Applied Computational Electromagnetics Society and has recently served as Co-Editor-in-Chief of the Applied Computational Electromagnetics Society Journal. He has also served as an Associate Editor for Radio Science and as the Secretary of Commission B of the U.S. National Committee of URSI. From August 2001 to July 2004 he was the Editor-in-Chief of the IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION.

1928

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Bandwidth Enhancement and Further Size Reduction of a Class of Miniaturized Slot Antennas
**

Nader Behdad, Student Member, IEEE, and Kamal Sarabandi, Fellow, IEEE

Abstract—In this paper, new methods for further reducing the size and/or increasing the bandwidth (BW) of a class of miniaturized slot antennas are presented. This paper examines techniques such as parasitic coupling and inductive loading to achieve higher BW and further size reduction for this class of miniaturized slot antennas. The overall BW of a proposed double resonant antenna is shown to be increased by more than 94% compared with a single resonant antenna occupying the same area. The behavior of miniaturized slot antennas, loaded with series inductive elements along the radiating section is also examined. The inductive loads are constructed by two balanced short circuited slot lines placed on opposite sides of the radiating slot. These inductive loads can considerably reduce the antenna size at its resonance. Prototypes of a double resonant antenna at 850 MHz and inductively loaded miniaturized antennas at around 1 GHz are designed and tested. Finally the application of both methods in a dual band miniaturized antenna is presented. In all cases measured and simulated results show excellent agreement. Index Terms—Slot antennas, electrically small antennas, parasitic antennas, multifrequency antennas.

I. INTRODUCTION

C

URRENT advancements in communication technology and signiﬁcant growth in the wireless communication market and consumer demands demonstrate the need for smaller, more reliable and power efﬁcient, integrated wireless systems. Integrating entire transceivers on a single chip is the vision for future wireless systems. This has the beneﬁt of cost reduction and improving system reliability. Antennas are considered to be the largest components of integrated wireless systems; therefore antenna miniaturization is a necessary task in achieving an optimal design for integrated wireless systems. The subject of antenna miniaturization is not new and has been extensively studied by various authors [1]–[4]. Early studies have shown that for a resonant antenna, as size decreases, bandwidth (BW) and efﬁciency will also decrease [1]. This is a fundamental limitation which, in general, holds true independent of antenna architecture. However, research on the design of antenna topologies and architectures must be carried out to achieve maximum possible BW and efﬁciency for a given antenna size. Impedance matching for small antennas is also challenging and often requires external matching networks; Therefore antenna topologies and structures which inherently

Manuscript received May 22, 2003; revised September 30, 2003. This work was supported in part by the Engineering Research Centers program of the National Science Foundation (NSF) under Award EEC-9986866 and by the U.S. Army Research Ofﬁce under Contract DAA-99-1-01971. The authors are with The Radiation Laboratory, Department of Electrical Engineering and Computer Science, The University of Michigan, Ann Arbor, MI 48109-2122 USA (e-mail: behdad@engin.umich.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832330

allow for impedance matching are highly desirable. The fundamental limitation introduced by Chu [1] and later re-examined by McLean [4] relates the radiation Q of a single resonance antenna with its BW. However, whether such limitation can be directly extended to multiresonance antenna structures or not is unclear. In fact, through a comparison with ﬁlter theory, designing a relatively wideband antenna may be possible using multipole (multiresonance) high Q structures. In this paper we examine the applicability of multiresonance antenna structures to enhance the BW of miniaturized slot antennas. Different techniques have been used for antenna miniaturization such as: miniaturization using optimal antenna topologies [5]–[7] and miniaturization using magneto-dielectric materials [8], [9]. In pursuit of antenna miniaturization while maintaining ease of impedance matching and attaining relatively high efﬁciency, a novel miniaturized slot antenna was recently presented [6]. Afterwards, a similar architecture in the form of a folded antenna geometry was presented in order to increase the BW of the previously mentioned miniaturized slot antenna [7]. Here we re-examine this topology [6] and propose modiﬁcations that can result in further size reduction or BW enhancement without imposing any signiﬁcant constraint on impedance matching or cross polarization level. In Section II, a dual-resonant antenna topology is examined for BW enhancement. This miniaturized antenna shows a BW which is 94% larger than that of a single-resonant miniaturized antenna with the same size. Using series inductive elements distributed along the antenna aperture results in the increase of inductance per unit length of the line. Therefore the guided wavelength of the resonant slot line is shortened. Thus, the overall length of the antenna is decreased. In Section III, this technique is ﬁrst demonstrated using a standard resonant slot antenna and then incorporated in the miniaturized antenna topology of [6] to further reduce the resonant frequency without increasing the area occupied by the antenna. The aforementioned techniques for BW enhancement and further size reduction can be used individually or in combination. The combined application of the techniques of Sections II and III is presented in Section IV by demonstrating the design of a dual band miniaturized slot antenna. II. MINIATURIZED SLOT ANTENNA WITH ENHANCED BW A. Design Procedure In this section the design of coupled miniaturized slot antennas for BW enhancement is studied. The conﬁguration of the proposed coupled slot antenna is shown in Fig. 1(b) where

0018-926X/04$20.00 © 2004 IEEE

BEHDAD AND SARABANDI: BW ENHANCEMENT AND FURTHER SIZE REDUCTION

1929

Fig. 1. Geometry of single- and double-element miniaturized slot antennas. (a) Single-element miniaturized slot antenna. (b) Double-element miniaturized slot antenna.

two miniaturized slot antennas are arranged so that they are parasitically coupled. Each antenna occupies an area of about [Fig. 1(a)] and achieves miniaturization by the virtue of a special topology described in detail in [6]. However, this antenna demonstrates a small BW (less than 1%). A close examination of the antenna topology reveals that the slot-line trace of the antenna only covers about half of the rectangular printed-circuit board (PCB) area. Therefore another antenna, with the same geometry, can be placed in the remaining area without signiﬁcantly increasing the overall PCB size. Placing two antennas in close proximity of each other creates strong coupling between the antennas which, if properly controlled, can be employed to increase the total antenna BW. As seen in Fig. 1(b), only one of the two antennas is fed by a microstrip line. The other antenna is parasitically fed through capacitive coupling mostly at the elbow section. The coupling is a mixture of electric and magnetic couplings that counteract each other. At the elbow section, where the electric ﬁeld is large, the slots are very close to each other; therefore, it is expected that the electric ﬁeld coupling is the dominant coupling mechanism and the electric ﬁelds (magnetic currents) in both antennas will be in phase and thus the radiated far ﬁeld is enhanced. The two coupled antennas are designed to resonate at the , where is the center fresame frequency, and are the resonant frequencies of the two quency and spectral response of the coupled antennas. In this case the antenna will show two nulls, the separation of which is a function of the separation between the two antennas, , and their overlap distance . In order to quantify this null separation a coupling coefﬁcient is deﬁned as (1) where and are the frequencies of the upper and lower nulls . Hence can easily be adjusted by varying and in [Fig. 1(b)], and decreases as is increased and is decreased. A full-wave electromagnetic simulation tool can be used to extract as a function of and in the design process. BW maximization is accomplished by choosing a coupling coefﬁcient (by remains below dB over the choosing and ) such that

entire frequency band. Here the resonant frequencies of both anMHz and is used as the tennas are ﬁxed at and tuning parameter. However, it is also possible to change slightly, in order to achieve a higher degree of control for tuning the response. The input impedance of a microstrip-fed slot antenna, for a given slot width, depends on the location of the microstrip feed relative to one end of the slot and varies from zero at the short circuited end to a high resistance at the center. Therefore an off-center microstrip feed can be used to easily match a slot antenna to a wide range of desired input impedances. The optimum location of the feed line can be determined from the full-wave simulation. In the double antenna example the feed line consists of a 50 transmission line connected to an open-circuited 75 line crossing the slot [Fig. 1(b)]. The 75 line is extended by beyond the strip-slot crossing to couple the maximum energy to the slot and also to compensate for the imaginary part of the input impedance. Using this 75 line as the feed, allows for compact and localized feed of the antenna and tuning the location of the transition from 50 to 75 provides another tuning parameter for obtaining a good match. B. Fabrication and Measurement A double-element antenna (DEA) and two different singleelement antennas (SEAs) (SEA 1 and SEA 2) were designed, fabricated, and measured. SEA 1 is the constitutive element of DEA and SEA 2 is an SEA with the same topology as SEA 1 [see Fig. 1(a)] but with the same area as the DEA. SEA 2 is used to compare the BW of the double resonant miniaturized antenna with that of the single-resonant miniaturized antenna with the same size. All antennas were simulated using IE3D [12] which is a full wave simulation software based on method of moments (MoM) and fabricated on a Rogers RO4350B substrate with , and a thickness of 500 m, a dielectric constant of with a copper ground plane of loss tangent of 33.5 23 cm . The return losses of the SEAs as well as the DEA are presented in Fig. 2. SEA 1 shows a BW of 8 MHz or nearly 0.9% and SEA 2 shows a BW of 11.7 MHz or 1.31% whereas the BW of the DEA is 21.6 MHz (2.54%) which indicates a factor of 1.94 increase over a SEA (SEA) with the

1930

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Return losses of DEA and SEA 2: SEA with the same size as DEA. (a) Return losses of the DEA and SEA of the same size (SEA 2). (b) Return loss of the SEA that constitutes the DEA (SEA 1).

same area. Choosing a different substrate with different thickness and dielectric constant, can increase the overall BW of both antennas. However, it is also expected that the BW ratio of the DEA to the SEA remains the same. The overall size of the DEA which shows a 25% increase in area when is . The Q compared to the size of the SEA 1 of each antenna has also been calculated using the method presented in [10] and compared with the fundamental limit on the Q of small antennas [4] in Table II. Demonstrably the quality factors of both SEAs are well above the minimum theoretical limit. Since Q is only deﬁned for single resonant structures, no value for Q is reported for the DEA in Table II. In calculating the minimum Q for the slot antennas using the Chu limit, it is necessary to ﬁnd the radius of the smallest sphere than encloses the antenna. At ﬁrst, it may not be clear whether this sphere should only cover the aperture or, in addition to that, some portion of the ground plane too (because of the electric currents that exist in the ground plane). This becomes clear by applying the equivalence theorem to this problem, which shows that the magnetic currents responsible for radiation, exist only on the aperture and according the the derivation of the Chu limit, the smallest sphere that encloses these radiating magnetic currents should be used. The gain of the double resonant antenna was measured at three different frequencies and is presented in Table I. Radiation patterns of the antenna were measured MHz and found to be similar to each at other. Fig. 3 shows the co- and cross-polarized E- and H-plane MHz. The E- and H-plane radiradiation patterns at ation patterns of this antenna are expected to be dual of those of a short electric dipole. Fig. 3(a) shows the H-Plane radiation pattern which is similar to the E-Plane radiation pattern of . an electric dipole with deeps instead of nulls at This can be attributed to the ﬁniteness of the ground plane where some radiation comes from the electric currents on the antenna ground plane at the edges of the substrate. Fig. 3(b), however, does not show a uniform radiation pattern like the H-Plane radiation of a short electric dipole. This is because of the 180 difference in phase between the normal component of the electric ﬁelds at the top and bottom of the antenna ground plane. The H-Plane pattern is expected to have deep nulls at these angles; therefore, this does not signiﬁcantly affect the

**TABLE I COMPARISON BETWEEN
**

THE DEA AND ITS SHOWN IN FIG. 1

CONSTITUTIVE SEAS

TABLE II COMPARISON BETWEEN MEASURED Q AND THE MINIMUM ATTAINABLE Q. 3 CALCULATED USING THE FOSTER REACTANCE THEOREM [10]. 33 CALCULATED USING THE CHU-MCLEAN FORMULA FOR A SINGLE-RESONANT ANTENNA [4]

H-Plane pattern. Table I shows the radiation characteristics of the DEAs and SEAs. It is seen that the gain-BW product of the proposed double-antenna is signiﬁcantly higher than that of the single antenna. III. IMPROVED ANTENNA MINIATURIZATION USING DISTRIBUTED INDUCTIVE LOADING A. Design Procedure A microstrip-fed slot antenna has the length of , where is the wavelength in the slot, at its ﬁrst resonance. The electric current distribution can be modeled by the voltage distribution transmission line short circuited at both ends. The over a resonant length of a transmission line can be made smaller if the inductance per unit length of the line is increased. This can be accomplished by inserting a number of series inductors in the transmission line. For slot-lines, insertion of series lumped elements is not possible. Besides, series lumped elements have a low Q which adversely affects antenna efﬁciency (gain). To realize a slot line with higher inductance per unit length, an array of distributed, short circuited, narrow slot-lines can be placed

BEHDAD AND SARABANDI: BW ENHANCEMENT AND FURTHER SIZE REDUCTION

1931

Fig. 3. Far ﬁeld radiation patterns of the double-element miniaturized slot antenna at 852 MHz. (a) H-plane and (b) E-plane.

Fig. 4. Loaded and unloaded straight slot antennas. (a) Geometry of a microstrip-fed straight slot antenna. (b) Geometry of a microstrip-fed straight slot antenna loaded with an array of series inductive elements.

along the radiating segment of the slot antenna as shown in Fig. 4(b). The impedance of a short circuited slot line is obtained by (2) where is the propagation constant, is the characteristic impedance, and is the length of the short circuited slot-line. The characteristic impedance of a slot-line is inversely proportional to its width [11] therefore by using wider series slots, more inductance can be obtained for a ﬁxed length of short circuited transmission line. The best location to put series inductors in a slot is near its end where the amplitude of magnetic current is small. Putting them at the center of the slot where the magnetic current is at its maximum, strongly degrades radiation efﬁciency. It can easily be seen that by increasing the number and value of inductors, the length of transmission line necessary to satisfy the boundary conditions at both ends of the slot decreases. The size reduction may also be explained by considering the electric current distribution in the conductor around the slot. There are two components of electric current in the ground plane of the slot, one that circulates around the slot and one that is

Fig. 5. Geometry of a miniaturized slot antenna loaded with series distributed inductors (slits).

perpendicular to it. The latter is described by the continuity of the electric current and displacement current at the slot discontinuity. Putting a discontinuity (a slit) normal to the circulating current path forces the current to circle around

1932

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 6. Simulated and measured return losses of the straight slot antennas and miniaturized slot antennas with and without inductive loading. (see Figs. 4 and 5). (a) Return losses of straight loaded and unloaded slot antennas. (b) Return losses of ordinary and loaded miniaturized slot antennas.

Fig. 7.

Far ﬁeld radiation patterns of loaded straight slot antenna shown in Fig. 4(b). (a) H-plane and (b) E-plane.

the discontinuity. Hence the electric current traverses a longer path length than the radiating slot length which in turn lowers the resonant frequency. Fig. 4(b) shows a slot antenna loaded with a number of narrow slits which act as an array of series inductors. These slits are designed to have a length smaller and carry a magnetic current with a direction normal than to that of the main radiator. Placing them only on one side of the radiating slot results in asymmetry in phase and amplitude of the current along the slot which could create problems in matching and worsen cross polarization. In order to circumvent this problem, two series slits are placed on the opposite sides of the main slot. These slits carry magnetic currents with equal amplitudes and opposite directions. Since the lengths of these narrow slits are small compared to the wavelength and since they are closely spaced, the radiated ﬁelds from the opposite slits cancel each other and they do not contribute to the radiated far ﬁeld. Matching is performed by using an off-centered open circuited microstrip feed. The optimum location and length

of the microstrip line are found by trial and error, using full wave simulations. For both straight slots (with and without series inductors) the lengths of the extended microstrip lines , where is the wavelength in the are found to be microstrip lines at their respective resonance frequencies. Fig. 5 shows a miniaturized slot antenna (similar to the topology in [6]) loaded with series inductive slits to further reduce its resonant frequency. The antenna without the series inductors is already small, and adding series inductive elements further reduces the resonant frequency or equivalently the electrical dimensions of the antenna. Instead of using identical inductive elements along the radiating slot, differently sized inductive slits are used to cover most of the available area on the PCB in order to maintain the area occupied by the antenna. The antenna is matched to a microstrip transmission line in a manner similar to the straight slots described earlier. The feed line is composed of a 75 open-circuited microstrip line connected to a 50 feed line. In this case, the open circuited

BEHDAD AND SARABANDI: BW ENHANCEMENT AND FURTHER SIZE REDUCTION

1933

Fig. 8.

Far ﬁeld radiation patterns of loaded miniaturized slot antenna shown in Fig. 5. (a) H-plane and (b) E-plane.

B. Fabrication and Measurement The straight slots with and without series inductors were simulated using IE3D and fabricated on a 500 m thick Rogers RO4350B substrate. Fig. 6(a) shows the simulated and measured return losses for the slot antennas with and without inductive dB loading. This ﬁgure shows the resonance frequency and BW of 2.2 GHz, and 235 MHz (10.7%) for the straight slot. The loaded slot with the same length as that of the unloaded slot has a resonance frequency of 1.24 GHz and a BW of 63 MHz (5%). This result indicates a 44% reduction in the resonant frequency and a similar reduction in the BW, as expected. The overall size can still be reduced by using longer short circuited slits, if they could be designed in a compact fashion. The radiation patterns of the small slot antenna were measured in the anechoic chamber of the University of Michigan and are presented in Fig. 7. It is seen that the cross polarization components in the far ﬁeld region in both E- and H-planes are negligible, thereby conﬁrming the fact that the radiation from the magnetic currents in the inductive loadings with opposite directions cancel each other in the far ﬁeld region. The miniaturized loaded and unloaded slot antennas were also fabricated using RO4350B substrate. Fig. 6(b) shows the simulated and measured return losses of the loaded and unloaded miniaturized antennas. It is shown that, by inserting the series inductors, the resonant frequency of the antenna shifts down from 1116 to 959 MHz (14% reduction). In this design, the overall PCB size is unchanged. Fig. 8 shows the E- and H-plane co- and cross-polarized radiation patterns of the loaded miniaturized antenna. It is seen that the cross polarization level is negligible at broadside. The gains of the loaded and unloaded miniaturized slot antennas (antenna in Fig. 5) were also measured in the anechoic chamber using a standard log-periodic reference antenna and were found to be 0.8 and 0.7 dB, respectively. Table III shows a comparison between Q of the miniaturized antennas presented in this section and the fundamental limit on Q of small antennas with the same size [1], [4]. It is observed that the Q of these antennas are well above the Chu limit.

Fig. 9. Geometry of the dual band inductively loaded miniaturized slot antenna of Section IV.

Fig. 10. Measured and simulated return losses of the miniaturized dual band slot antenna of Section IV.

microstrip line is extended beyond the slot-strip crossing by and for the miniaturized antenna and the loaded miniaturized antenna respectively.

1934

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE III COMPARISON BETWEEN BW, MEASURED Q AND THE MINIMUM ATTAINABLE Q OF THE MINIATURIZED ANTENNAS IN SECTION III

TABLE IV COMPARISON BETWEEN BW, MEASURED Q AND THE MINIMUM ATTAINABLE Q OF THE DUAL BAND MINIATURIZED ANTENNA

IV. DUAL BAND MINIATURIZED SLOT ANTENNA In this section the techniques introduced in the previous sections are used in the design of a dual band miniaturized slot antenna. The geometry of this antenna is shown in Fig. 9. and ) The resonant frequencies of the slot antennas ( are used as and the value of the coupling coefﬁcient design parameters to achieve the desired response. Increasing the vertical displacement, , and decreasing the horizontal separation, , causes to increase or equivalently result in a larger separation between the two frequency bands. Small changes in the resonance lengths of the slots result in slight and which can be used as a means of ﬁnechanges in tuning the response. Note, however, that resonant frequencies , and should be close to each other so that coupling takes place. The separation between the two bands is limited values cannot be obtained by practical values of . Large easily, since both electric and magnetic couplings are present and add destructively. In addition to this problem, matching the antenna at the two bands becomes increasingly difﬁcult is deﬁned as a as the separation increases. A parameter measure of separation between the two frequency bands (3) , where is the center frequency. In practice by changing and , values of up to 10% can easily be obtained. This architecture is particularly useful for wireless applications that use two separate frequency bands (different bands for transmit and receive for example) that are close to each other but still cannot be covered with the available BW of these types of miniaturized antennas. In order to achieve a higher miniaturization level for the given size, series inductive elements are also placed along slots to reduce the resonance frequencies of each element. Fig. 10 shows the simulated and measured return losses of this dual band antenna. The discrepancies between the simulated and measured results are due to the ﬁniteness of the ground plane as described MHz and in [6]. The measured results indicate an MHz or equivalently a %. A good match at both bands is obtained by using an off center open-circuited microstrip feed where the microstrip line is extended by 7 cm over

the slot-strip transition. Table IV shows a comparison between dB BW, measured Q, and minimum atthe antenna size, tainable Q for the two bands. It is seen that the Q of both bands are well above the Chu limit. The overall size of the structure at the is 5.73 cm 5.94 cm or equivalently lowest frequency of operation. Radiation patterns of the antenna at the two bands are measured and found to be similar to those of the SEA topology (Fig. 8). V. CONCLUSION Two approaches are introduced for increasing the BW and reducing the size of miniaturized slot antennas. Placing two similar slot antennas in close proximity of each other creates a double resonant structure, the response of which is a function of relative spacing between the two antennas. The coupled miniaturized antenna can be designed to have a BW which is larger by 94% than the BW of a single resonant antenna with the same area or to behave as a dual band antenna. For a ﬁxed resonant frequency, adding series inductive elements to a slot antenna reduces its size. The size reduction is a function of number and values of the inserted inductive elements. Using series inductive elements does not adversely affect impedance matching and the cross polarization level. This technique is also used in combination with other miniaturization techniques to further decrease the size of the radiating structure. The technique is applied to a straight as well as a miniaturized slot antenna and for a given antenna size, signiﬁcant reduction in resonant frequencies are observed. Finally, both techniques are applied to the design of a miniaturized dual band antenna. Series inductors are used to reduce the resonant frequencies of each resonator. A large coupling coefﬁcient is used to achieve a large separation between the two response of the parasitically coupled antenna. nulls in the , and are used as design parameters in The values of order to obtain a miniaturized dual band slot antenna with relatively good simultaneous matching. ACKNOWLEDGMENT The authors would like to thank the anonymous reviewers for their constructive comments.

BEHDAD AND SARABANDI: BW ENHANCEMENT AND FURTHER SIZE REDUCTION

1935

REFERENCES

[1] L. J. Chu, “Physical limitations on omni-directional antennas,” J. Appl. Phys., vol. 19, pp. 1163–1175, Dec. 1948. [2] H. A. Wheeler, “Fundamental limitations of small antennas,” in Proc. IRE., vol. 35, Dec. 1947, pp. 1479–1484. [3] R. C. Hansen, “Fundamental limitations in antennas,” Proc. IEEE, vol. 69, pp. 170–182, Feb. 1981. [4] J. S. McLean, “A re-examination of the fundamental limits on the radiation Q of electrically small antennas,” IEEE Trans. Antennas Propagat., vol. 44, pp. 672–676, May 1996. [5] H. K. Kan and R. B. Waterhouse, “Small square dual-spiral printed antennas,” Electron. Lett., vol. 37, pp. 478–479, Apr. 2001. [6] K. Sarabandi and R. Azadegan, “Design of an efﬁcient miniaturized UHF planar antenna,” in Proc. IEEE Int. Antennas Propagat. & URSI Symp., Boston, MA, July 8–13, 2001. [7] R. Azadegan and K. Sarabandi, “Miniaturized folded-slot: An approach to increase the bandwidth and efﬁciency of miniaturized slot antennas,” in Proc. IEEE Int. Antennas Propagat. & URSI Symp., San Antonio, TX, June 16–21, 2002. [8] K. Sarabandi and H. Mosallaie, “Antenna miniaturization with enhanced bandwidth and radiation characteristics: A novel design utilizing periodic magneto-dielectric,” in Proc. IEEE Int. Antennas Propagat. & URSI Symp., San Antonio, TX, June 16–21, 2002. [9] T. Ozdemir, P. Frantzis, K. Sabet, L. Katehi, K. Sarabandi, and J. Harvey, “Compact wireless antennas using a superstrate dielectric lens,” in Proc. IEEE Trans. Antennas Propagat. & URSI Symp., Salt Lake City, Utah, July 2000. [10] W. Geyi, P. Jarmuszewski, and Y. Qi, “The foster reactance theorem for antennas and radiation Q,” IEEE Trans. Antennas Propagat., vol. 48, pp. 401–408, Mar. 2000. [11] J. J. Lee, “Slotline impedance,” IEEE Trans. Microwave Theory and Techniques, vol. 39, pp. 666–672, Apr. 1991. [12] Electromagnetic Simulation and Optimization Software. IE3D.

Nader Behdad (S’97) was born in Mashhad, Iran, in 1977. He received the Bachelor of Science degree from Sharif University of Technology, Tehran, Iran, and the Master of Science degree from the University of Michigan, Ann Arbor, in 2000 and 2003, respectively, where he is currently working toward the Ph.D. degree on bandwidth enhancement and miniaturization of printed antennas in the Department of Electrical Engineering and Computer Science. From 2000 to 2001, he was with the Electronics Research Center, Sharif Unviersity of Technology, as an antenna design Engineer working on design of antennas for wireless local loop (WLL) systems. Since January 2002, he has been working as a Research Assistant in the Radiation Laboratory, University of Michigan. Mr. Behdad is the recipient of the Best Student Paper Award in the Antenna Applications Symposium held in Monticelo, IL, in September 2003 and winner of the Second Prize in the student paper competition of the USNC/URSI National Radio Science Meeting, Boulder, CO, in January 2004.

Kamal Sarabandi (S’87–M’90–SM’92–F’00) received the B.S. degree in electrical engineering from Sharif University of Technology, Tehran, Iran, in 1980, the M.S. degree in electrical engineering/mathematics, and the Ph.D. degree in electrical engineering from The University of Michigan, Ann Arbor, in 1986 and 1989, respectively. He is Director of the Radiation Laboratory and a Professor in the Department of Electrical Engineering and Computer Science, The University of Michigan. He has 20 years of experience with wave propagation in random media, communication channel modeling, microwave sensors, and radar systems and is leading a large research group consisting of four research scientists, ten Ph.D. students, and two M.S. students. Over the past ten years he has graduated 15 Ph.D. students. He has served as the Principal Investigator on many projects sponsored by NASA, JPL, ARO, ONR, ARL, NSF, DARPA, and numerous industries. He has published many book chapters and more than 105 papers in refereed journals on electromagnetic scattering, random media modeling, wave propagation, antennas, microwave measurement techniques, radar calibration, inverse scattering problems, and microwave sensors. He has had more than 220 papers and invited presentations in national and international conferences and symposia on similar subjects. His research areas of interest include microwave and millimeter-wave radar remote sensing, electromagnetic wave propagation, and antenna miniaturization. Dr. Sarabandi is a Member of the International Scientiﬁc Radio Union (URSI) Commission F and of The Electromagnetic Academy. He received the Henry Russel Award from the Regent of The University of Michigan (the highest honor the University of Michigan bestows on a faculty member at the assistant or associate level). In 1999, he received a GAAC Distinguished Lecturer Award from the German Federal Ministry for Education, Science, and Technology. He also received a 1996 Teaching Excellence Award from the Department of Electrical Engineering and Computer Science, and the 2003/2004 College of Engineering Research Excellence Award, The University of Michigan. He is a Vice President of the IEEE Geoscience and Remote Sensing Society (GRSS), a past Chairman of the Awards Committee of the IEEE GRSS from 1998 to 2002, and a Member of the IEEE Technical Activities Board Awards Committee from 2000 to 2002. He is an Associate Editor of the IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION and the IEEE SENSORS JOURNAL. He is listed in American Men & Women of Science, Who’s Who in America, and Who’s Who in Electromagnetics.

1936

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Miniature Built-In Multiband Antennas for Mobile Handsets
**

Yong-Xin Guo, Member, IEEE, Michael Yan Wah Chia, Member, IEEE, and Zhi Ning Chen, Member, IEEE

Abstract—In this paper, we propose a new design for built-in handset antennas in that metal strips as additional resonators are directly connected with a feed strip. With the new design scheme, a quad-band antenna for covering GSM900, DCS1800, PCS1900, and UMTS2000 bands and a ﬁve-band antenna for covering GSM900, DCS1800, PCS1900, UMTS2000, and ISM2450 bands for use in mobile built-in handsets are experimentally carried out. Compared with the parasitic form with a shorted strip placed away from the main radiator in the open literature, the size of the proposed antennas can be reduced by an order of 10 20%, which is desirable since the size of mobile phones is becoming smaller according to consumer preferences. Moreover, the impedance matching for each band of the new antennas becomes easy. The new quad-band and ﬁve-band built-in handset antennas are developed within the limits of a 36 16 8 mm3 volume. The antennas are also analyzed using the ﬁnite-difference time-domain technique. A good agreement is achieved between measurement and simulation. Index Terms—Antennas, built-in antennas, handset antennas, planar inverted-F antennas (PIFAs) antennas, small antennas.

I. INTRODUCTION

T

HE SIZES AND weights of mobile handsets have rapidly been reduced due to the development of modern integrated circuit technology and the requirements of the users. Conventional monopole-like antennas have remained relatively large compared to the handset itself. Thus, built-in antennas are becoming very promising candidates for applications in mobile handsets. Most built-in antennas currently used in mobile phones include microstrip antennas, inverted-F shaped wire-form antennas (IFAs), and planar inverted-F antennas (PIFAs). Microstrip antennas are small in size and light in weight. However, at the lower band for mobile applications such as GSM900, half-wavelength microstrip antennas are too large to be incorporated into a mobile handset. Basic IFA and PIFA elements, which have a length equal to a quarter wavelength of the center frequency in the operating band, are narrow in bandwidth. In addition to reduced antenna sizes, it is envisaged that next generation mobile phones will require the capability to tune to a number of frequency bands for cellular applications, wireless local area networks (WLAN) and other wireless communications. The trend in the development of wireless personal communication systems has been in the pursuit of a single system that can accommodate the needs of all users. To develop compact, highly efﬁcient and broadband

Manuscript received February 24, 2003; revised November 2, 2003. The authors are with the Institute for Infocomm Research, Singapore 117674, Singapore (e-mail: yxguo@ieee.org). Digital Object Identiﬁer 10.1109/TAP.2004.832375

antennas, which are well suited for integration with multiband multifunctional personal communication transceiver systems, is a big challenge for antenna engineers. Currently, many mobile telephones use one or more of the following frequency bands: the GSM band centered at 900 MHz, the DCS band centered at 1800 MHz, and the PCS band centered at 1900 MHz. Many interesting designs based on the IFA and PIFA concepts for achieving dual-band operations have been available in open literatures [1]–[11]. Triple-band built-in antennas to operate at GSM900, DCS1800, and PCS1900 bands were demonstrated [12]–[14]. These tri-band antennas consist of a main radiator operating at a low frequency band and a ﬁrst high band and a shorted parasitic radiator operating at a second high band. The parasitic radiator lies in a plane parallel to and away from the main radiator and therefore occupies valuable space in mobile phones that are constantly shrinking in size. Moreover, the parasitic-feed technique used for introducing one more mode may have problems in tuning of the antenna. More recently, some customers may need the designed mobile antennas can also include the UMTS2000 band for 3G mobile applications or 2450 MHz ISM band for indoor cordless phones, WLAN and Bluetooth applications. Antenna designs for covering GSM900, DCS1800, and ISM2450 bands can be found in the literatures [15], [16]. Furthermore, a quad-band built-in antenna for covering GSM900, DCS1800, GSM1900, and UMTS2000 was reported in [17]. In this paper, we propose a new design in that metal strips as additional resonators are directly connected with a feed strip. With this direct-feed scheme, the forgoing problems relating to the parasitic-feed technique for an additional resonance in a conventional multiple-band antenna can be alleviated. As an example, a quad-band antenna for covering the GSM900, DCS1800, PCS1900, and UMTS2000 was achieved, which was initially presented in [18], [19]. Herein, we would like to report extensive results on this quad-band antenna. Further, by the addition of a second metal strip, a ﬁve-frequency band operation to cover GSM900, DCS1800, PCS1900, UMTS2000, and ISM2450 can be implemented. In the wireless industry, there are several ways to fabricate handset antennas, such as: 1) molded intrusion design (MID) technology; 2) using a conductive pattern, screen-printed on an adhesive ﬂexible ﬁlm; 3) thin ﬁlm technology; and 4) metal cutting. It is possible to fabricate our newly designed antennas using MID technology or metal cutting. The simulations were performed using Remcom software XFDTD5.3, which is based on the FDTD method. This paper is organized as follows. Section II presents a simple and efﬁcient measurement setup, which is very important for measuring small handset antennas. In Section III, a

0018-926X/04$20.00 © 2004 IEEE

GUO et al.: MINIATURE BUILT-IN MULTIBAND ANTENNAS FOR MOBILE HANDSETS

1937

Fig. 1. Geometry of the test bed.

starting point for the quad-band and ﬁve-band antennas is provided. After that, the built-in quad-band handset antenna design and parametric study are extensively described in Section IV. Then, Section V shows the built-in ﬁve-band handset antenna. Finally, the entire work is summarized in Section VI.

II. MEASUREMENT SETUP The measurement methods of mobile handset antennas are of much concern by many investigators [20]–[23]. During the development of such a handset antenna, the antenna under test (DUT) is connected to a network analyzer via a coaxial cable. Errors in the measured resonant frequency, bandwidth, radiation pattern, and antenna gain can be expected owing to the feed cable placed in the near ﬁeld of the antenna and the coaxial cable acting as a secondary radiator driven by the surface currents ﬂowing on its outer surface of the shield. Ferrite chokes on the exterior of the cables can reduce the cable-related effect signiﬁcantly [20]. However, ferrites typically work well as chokes up to 1 GHz. The use of sleeve-like baluns on the cable, located near the handset, can reduce the effect of the RF cable on the antenna measurement as well [21], [22]. As shown in [22], the design of such a balun in multiband operation is very complicated. Moreover, it is commonly known to us in wireless industry that a semi-rigid coaxial transmission line is placed on the ground plane with its centre conductor connected to the antenna while its shield soldered onto the ground place as in [23]. With this arrangement, the effects of the feeding coaxial cable on the antenna can be reduced to an acceptable level. Thus this feeding scheme can best model the real mobile phone where a separate RF transceiver, residing inside a metal enclosure, is employed to drive the antenna that is mounted very close to the RF transceiver. In this paper, the test printed circuit board (PCB) with the DUT for use in a mobile telephone is shown in Fig. 1. The PCB is in a rectangular shape. It has a ground plane and a microstrip line etched on the back. In practical use the PCB will have a number of electronic components mounted thereon, which are necessary for the operation of the mobile telephone, but which are omitted here for brevity. The rectangular ground plane has a

Fig. 2. (a) Existing dual-band internal handset antenna and (b) dual-band internal handset antenna with a ﬁne-tuning stub.

length of and a width of . The substrate is Duroid RO4003 with a thickness of 1.5 mm and dielectric constant of 3.38. The width of the microstrip line is 3.5 mm to keep its characteristic impedance at 50 . The DUT is placed at the top of the ground plane. A coaxial cable is connected to network analyzer HP8753E at the bottom of the board. The gain and radiation patterns are measured using Orbit/FR system in an anechoic chamber. The above mentioned measurement setup for handset antenna design was validated and conﬁrmed in terms of measured and simulated input reﬂection, near-ﬁeld currents on the ground plane and far-ﬁeld radiation patterns in house. III. STARTING POINT The antenna shown in Fig. 2(a) is the one as in [12], which was used as the starting point of this work. The antenna comprises a folded radiating patch in the ﬁrst layer, a ground plane in the second layer, a supporting foam in-between, a short-circuited strip, and a feed strip. The patch is connected to the ground plane via a vertical short-circuited strip and is fed via a feed strip connected to a 50- transmission line etched on the back of the ground plane. The folded PIFA is spaced from the ground plane by a dielectric substrate of foam. At the ﬁrst layer, the long bent portion of the antenna is tuned to have a relatively low band resonance frequency, such as 900 MHz, and the short

1938

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 4. Proposed quad-band internal handset antenna.

To reduce the resonant frequency further, we may extend the arm length or bend at the open end of the folded antenna in Fig. 2(a). However, the bending at the open end in Fig. 2(a) may be a little difﬁcult to manufacture. Fig. 2(b) shows one possible variation of the folded antenna in Fig. 2(a) with an additional strip stub bent perpendicularly toward the ground plane. The and width . All other stub has dimensions of length parameters are kept same as those in Fig. 2(a). The measured and simulated return losses of the antenna in Fig. 2(b) is shown in Fig. 3(b). The size of the stub for the results shown in Fig. 3(b) and . The measured bandwidths is for 6 dB return loss are 76 MHz (942–1018 MHz) for the lower band and 239 MHz (1752–1991 MHz) for the upper band, respectively. The corresponding simulated results are 122 MHz (928–1050 MHz) and 242 MHz (1808–2050 MHz). In this case, the resonant frequencies can meet the requirement of GSM900 and DCS1800 with a plastic cover being added.

Fig. 3. (a) Measured and simulated return losses of the antenna in Fig. 2(a) and (b) measured and simulated return losses of the antenna in Fig. 2(b).

IV. QUAD-BAND ANTENNA A. Antenna Structure The proposed antenna in this work is shown in Fig. 4. Compared with the dual-band folded patch in Fig. 2(b), a new radiating strip is added. The new radiating strip as an additional resonator is directly connected to the feed strip and positioned at a plane perpendicular to the ground plane and the original folded patch, whereas in the previous triple-band designs [12]–[14], the additional parasitic radiator has a small distance from the original folded patch. Due to that, the designs in [12]–[14] may occupy valuable space in mobile phones that are constantly shrinking in size. Compared with the designs in [12]–[14], the size of the newly proposed antenna can be reduced by an order of 10 20%, which is desirable since the size of mobile phones is becoming smaller according to consumer preferences. The new additional strip is like a PIFA antenna and is tuned to have a second high resonance frequency, such as 2100 MHz. The new quad-band antenna was volume. The developed within the limits of a and a rectangular ground plane has a length of width of . The dimensions of the new antenna , , , , are , , , , ,

part of the antenna is tuned to have a high band resonance frequency, such as 1800 MHz. The ground plane has dimensions of length 80 mm, and width 36 mm. The dielectric constant of foam is around 1.07. The dimensions of the dual-band antenna , , , , are , , , , , . and The measured and simulated return losses of the dual-band antenna in Fig. 2(a) is shown in Fig. 3(a). The measured bandwidths, deﬁned for 6 dB return loss (SWR 3), are 82 MHz (972–1054 MHz) for the lower band and 263 MHz (1769–2032 MHz) for the upper band, respectively. The corresponding simulated bandwidths are 120 MHz (965–1085 MHz) and 270 MHz (1830–2100 MHz). Reasonable agreement between measurement and simulation is achieved. The wide bandwidth at the upper band in this design may also be due to one resonance generated by the ground plane, which has a half-wave length with the center frequency being around 1.8 GHz. If the inﬂuence of the plastic casing as in real situation is considered with a rough 5% reduction of the resonant frequency [12], it is observed that the corresponding frequencies are still a little out of the GSM900 and DCS1800 bands.

GUO et al.: MINIATURE BUILT-IN MULTIBAND ANTENNAS FOR MOBILE HANDSETS

1939

Fig. 5. Measured and simulated return losses of the antenna in Fig. 4.

agreement between measurement and simulation is obtained. Referring to Fig. 5, it is observed that there are some differences for the null depth in the simulated and measured return losses of the upper band, which may come from that the antenna size cannot be modeled very accurately by the FDTD method due to its meshing scheme. The simulated bandwidths with the plastic cover as in real case are 126 MHz (883–1009 MHz) at the lower band and 573 MHz (1659–2232 MHz) at the upper band, respectively. The antenna has a capacity for covering the GSM900, DCS1800, PCS1900, and UMTS2000 bands. With regard to Fig. 5, the return loss has one distinct minimum at a low frequency band and two minima at two high frequency bands relatively close to each other. It is very clear to observe that the wide bandwidth of the higher band of the new antenna is due to the introduced strip connected to the feed. Note that the wide bandwidth at the upper band in this design may also come from one resonance generated by the ground plane, which has a half-wave length with the center frequency being around 1.8 GHz. The -plane far-ﬁeld radiation patterns of the new quadband antenna at 935, 1795, 1935, and 2100 MHz are depicted in Fig. 7(a)–(d), respectively. They are similar to those of other integrated antennas for mobile handsets [1]. Referring to Fig. 7, the overall shape of the radiation patterns can be suitable for mobile communications terminals. The measured values of the gain of the quad-band antenna are shown in Fig. 6. Also, the measured gains for the dual-band antenna as in Fig. 2(b) are included for comparison. The measured gains are varying from 0.4 to 3.6 dBi. C. Parametric Study

Fig. 6. Measured gains of the dual-band antenna in Fig. 2(b) and the quad-band antenna in Fig. 4.

, . B. Measured and Simulated Results

,

, and

The return losses and radiation properties of the new antenna as shown in Fig. 4 were investigated using measurement and simulation. The simulation was performed using the commercial software XFDTD5.3, which is based on the FDTD method. Fig. 5 shows the measured and simulated return losses of the new antenna presented in Fig. 4. In the actual design, we need to consider around 5% frequency-shifting due to the effect of the plastic cover [7]. Thus, the simulated result with the plastic cover is also provided. In the simulation, 2-mm thick and 1-mm dielectric sheet with dielectric constant spacing between the cover and the antenna is used to simulate the actual effect of the plastic cover [7], [24]. The measured bandwidths without the plastic cover according to 6 dB return loss matching are 78 MHz (933–1010 MHz) at the lower band and 516 MHz (1772–2288 MHz) at the upper band, respectively. The corresponding simulated results without the plastic cover are 130 MHz (924–1054 MHz) at the lower band and 486 MHz (1824–2310 MHz) at the upper band, respectively. A good

In this section, effects of varying key antenna parameters, i.e. the substrate thickness, the ground plane size, and the additional strip position and width are considered on the antenna bandwidth. Again, for all the simulations in this section, 2-mm thick dielectric sheet with dielectric constant and 1-mm spacing between the cover and the antenna are used to simulate the actual effect of the plastic cover as before [7], [24]. 1) Effects of the Substrate Thickness: The ﬁrst variation is performed by varying the height of the substrate. Other param, , eters of the antenna are as follows: , , , , , , , , , , , and . It can be seen from Fig. 8, that this variation has a large impact on the impedance matching of the upper band. With the substrate height increasing, the bandwidths for the upper band increase signiﬁcantly, while the bandwidths of the lower band increase slightly. The lower cutoff frequency (f1L), upper cutoff frequency (f1U) and the absolute bandwidth (BW1) for the lower band and the lower cutoff frequency (f2L), upper cutoff frequency (f2U) and the absolute bandwidth (BW2) for the upper band are tabulated in Table I for reference, respectively. Additionally, the resonant frequencies of the lower band shift up with the substrate height being increased, which may be due to the decreased capacitance between the folded radiating patch and the ground plane.

1940

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 7.

Radiation patterns of the antenna in Fig. 4 at xz plane: (a) 925 MHz, (b) 1795 MHz, (c) 1935 MHz, and (d) 2100 MHz.

Fig. 8. Simulated return losses with variation of H for the antenna in Fig. 4.

2) Effects of the Ground Plane Size: For small PIFA-like antennas, the ﬁnite ground plane can be considered as a radiator. Therefore, it is necessary to study the effects of the ground plane size on the impedance characteristics of the new antenna. With regard to Fig. 9 it is observed that the ground plane length has a large effect on the upper band with the length varying from 60 to 100 mm, while the lower band almost keeps unchanged.

Only small variations are seen for the bandwidths of both the lower and upper bands with the ground plane width varying from 36 to 44 mm. Other parameters of the antenna are as fol, , , , lows: , , , , , , , , and . 3) Effects of the Additional Strip Position and Its Size: Fig. 10 shows the effects of the additional strip position and its width on the new antenna. Other parame, , ters of the antenna are as follows: , , , , , , , , , , and . Referring to the Fig. 10, it can be seen that the additional strip position and its width mainly affect the impedance matching of the upper band. The matching will become deteriorated when the additional strip bottom approaches very near the ground plane as high capacitance may be introduced in this case. The bandwidths of the lower band almost keep constant with the additional strip position and its size varying. The antennas with other additional strip positions and the strip sizes were also simulated

GUO et al.: MINIATURE BUILT-IN MULTIBAND ANTENNAS FOR MOBILE HANDSETS

1941

TABLE I SIMULATED BANDWIDTHS WITH DIFFERENT SUBSTRATE THICKNESS H OF THE QUAD-BAND ANTENNA

frequency of the introduced second high band as it acts like a quarter-wavelength IFA antenna. V. FIVE-BAND ANTENNA A. Antenna Structure Fig. 11 depicts a ﬁve-band antenna for covering the GSM900, DCS1800, PCS1900, UMTS2000, and ISM2450 bands by adding a second additional strip and connecting it to the feed strip. The second metal strip is parallel to the ground plane and the original dual-band PIFA patch but orthogonal to the ﬁrst additional radiating strip with a horizontal separation d3 and a height h1. The additional separation and position arrangement between the ﬁrst radiating strip and the second additional strip reduces the mutual coupling therebetween. The new ﬁve-band antenna was still developed within the limits of volume. The rectangular ground plane has a and a width of . a length of , The dimensions of this antenna are , , , , , , , , , , , , , , , and . B. Measured and Simulated Results Fig. 12 shows the measured and simulated return losses of the proposed ﬁve-band antenna presented in Fig. 11. The measured bandwidths for 6 dB return loss are 78 MHz (932–1010 MHz) at the GSM900 band, 456 MHz (1818–2274 MHz), and 115 MHz (2523–2638 MHz), respectively. The corresponding simulated results are 130 MHz (920–1050 MHz) at the lower band, 552 MHz (1760–2312 MHz) at the ﬁrst high band, and 80 MHz (2480–2560 MHz), respectively. A reasonable agreement between measurement and simulation is obtained. The antenna has a capacity for covering the GSM900, DCS1800, PCS1900, UMTS2000, and ISM2450 bands after the frequency shifting is considered due to the plastic cover. Again, it is observed that there are some differences for the null depth in the simulated and measured return losses of the ﬁrst high band, which may come from that the antenna size cannot be modeled very accurately by the FDTD method due to its

Fig. 9. Simulated return losses with variation of L for the antenna in Fig. 4.

Fig. 10.

d = 2 mm.

Simulated return losses with variation of W9 for the antenna in Fig. 4,

and tabulated in Table II. Again, the lower cutoff frequency (f1L), upper cutoff frequency (f1U) and the absolute bandwidth (BW1) for the lower band and the lower cutoff frequency (f2L), upper cutoff frequency (f2U) and the absolute bandwidth (BW2) for the upper band are listed in Table II, respectively. With regard to Table II, the bandwidth is deﬁned according to return loss of 6 dB. Moreover, it is easy to realize that the length of the additional strip mainly affects its resonant

1942

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE II SIMULATED BANDWIDTHS WITH DIFFERENT W9 AND d2 OF THE QUAD-BAND ANTENNA

Fig. 11.

Geometry of proposed ﬁve-band internal handset antenna.

meshing scheme. It is obvious that the band around 2450 MHz is due to the introduced second strip shown in Fig. 11. The radiation patterns were also measured and similar to those for the antenna presented in Fig. 4, thus are not shown for brevity. The measured gain at 2.45 GHz is around 1.5 dBi. VI. CONCLUSION In this paper, we have proposed a new design in that a new metal strip as an additional resonator is directly connected with a feed strip and positioned at a plane perpendicular to a ground plane. With the new design scheme, a quad-band antenna

Fig. 12.

Measured and simulated return losses for the antenna in Fig. 11.

for covering GSM900, DCS1800, PCS1900, and UMTS2000 bands and a ﬁve-band antenna for covering GSM900, DCS1800, PCS1900, UMTS2000, and ISM2450 bands for use in mobile handsets have been experimentally carried out. Compared with the parasitic form, the size of the proposed antennas can be reduced by an order of 10 20%. Moreover, the impedance matching for each band becomes easy. The new quad-band and ﬁve-band antennas have been developed within the limits volume. The antennas have also been of a

GUO et al.: MINIATURE BUILT-IN MULTIBAND ANTENNAS FOR MOBILE HANDSETS

1943

analyzed using the FDTD technique. A good agreement has been achieved between measurement and simulation.

ACKNOWLEDGMENT The authors would like to thank the two anonymous reviewers for their comments which improved this paper.

[21] C. Icheln, J. Ollikainen, and P. Vainikainen, “Reducing the inﬂuence of feed cables on small antenna measurements,” Electron. Lett, vol. 35, no. 15, pp. 1212–1214, July 1999. [22] C. Icheln and P. Vainikainen, “Dual-frequency balun to decrease inﬂuence of RF feed calbes in small antenna measurements,” Electron. Lett., vol. 36, no. 21, pp. 1760–1761, Oct. 2000. [23] J. Haley, T. Moore, and J. T. Bernhard, “Experimental investigation of antenna-handset-feed interaction during wireless product testing,” Microwave and Opt. Technol. Lett., vol. 34, no. 3, pp. 169–172, Aug. 2002. [24] H. S. Hwang, “private communication,” unpublished.

REFERENCES

[1] K. Hirasawa and M. Haneishi, Eds., Analysis, Design, and Measurement of Small and Low-Proﬁle Antennas. Norwood, MA: Artech House, 1992. [2] Z. D. Liu, P. S. Hall, and D. Wake, “Dual-frequency planar inverted-F antenna,” IEEE Trans. Antennas Propagat., vol. 45, pp. 1451–1457, Oct. 1997. [3] C. R. Rowell and R. D. Murch, “A compact PIFA suitable for dualfrequency 900/1800-MHz operation,” IEEE Trans. Antennas Propagat., vol. 46, pp. 596–598, Apr. 1998. [4] M. Sanad and N. Hassas, “Compact wide-band microstrip antennas for PCS and cellular phones,” in Proc. IEEE Conf. Antennas Propagation for Wireless Commun., Nov. 1998, pp. 152–155. [5] R. Mittra and S. Dey, “Challenges in PCS antenna design,” in Proc. IEEE Antennas Propagation Symp. Dig., Orlando, FL, July 1999, pp. 544–547. [6] S. Tarvas and A. Isohatala, “An internal dual-band mobile phone antenna,” in Proc. IEEE Antennas Propagation Symp. Dig., Salt Lake City, UT, July 2000, pp. 266–269. [7] M. Yang and Y. Chen, “A novel U-shaped planar microstrip antenna for dual-frequency mobile telephone communications,” IEEE Trans. Antennas Propagat., vol. 49, pp. 1002–1004, June 2001. [8] A. Taﬂove and L. Vasilyeva, “Elongate Radiator Conformal Antenna for Portable Communication Devices,” U.S. patent 6 292 144, Sept. 2001. [9] M. Martinez-Vazquez, M. Geissler, D. Heberling, A. Martinez-Gonzalez, and D. Sanchez-Hernandez, “Compact dual-band antenna for mobile handsets,” Microwave and Opt. Technol. Lett., vol. 32, no. 2, pp. 87–88, Jan. 2002. [10] R. Chair, K. M. Luk, and K. F. Lee, “Measurement and analysis of miniature multilayer patch antenna,” IEEE Trans. Antennas Propagat., vol. 50, pp. 244–250, Feb. 2002. [11] K. L. Wong, “A short course note on planar antennas for wireless communications,” in Proc. IEEE Antennas and Propagation Society Int. Symp., San Antonio, TX, June 2002, pp. 6–14. [12] D. Manteuffel, A. Bahr, D. Heberling, and I. Wolff, “Design consideration for integrated mobile phone antennas,” in Proc. 11th Int. Conf. Antennas and Propagation, Manchest, U.K., April 2001, pp. 252–256. [13] Z. Ying, Multi Frequency-Band Antenna PCT application WO01/91233, May 2001. [14] I. Egorov, “Antenna,” U.S. patent application 09/908 817, July 2001. [15] W. P. Dou and Y. M. W. Chia, “Novel meandered planar inverted-F antenna for triple-frequency operation,” Microwave Opt. Technol. Lett., vol. 27, pp. 58–60, Oct. 2000. [16] C. T. P. Song, P. S. Hall, H. Ghafouri-Shiraz, and D. Wake, “Triple band planar inverted F antennas for handheld devices,” Electron Lett., vol. 36, pp. 112–114, Jan. 2000. [17] M. Martinez-Vazquez and O. Litschke, “Design considerations for quadband antennas integrated in personal communications devices,” in Proc. Int. Symp. Antennas (JINA), vol. 1, Nice, France, Nov. 2002, pp. 195–198. [18] Y. X. Guo, M. Y. W. Chia, and Z. N. Chen, “Compact multi-band antennas for wireless communications,” in Proc. Progress in Electromagnetics Research Symp., Singapore, Jan. 2003, p. 130. [19] , “Miniature built-in quad-band antennas for mobile handsets,” IEEE Antennas Wireless Propagat. Lett., vol. 2, pp. 30–32, 2003. [20] S. Saario, D. V. Thiel, J. W. Lu, and S. G. O’Keefe, “An assessment of cable radiation effects on mobile communications antenna measurements,” in Proc. IEEE Antennas Propagt. Symp. Dig., Montreal, Canada, July 1997, pp. 550–553.

Yong-Xin Guo (M’01) received the B.Eng. and M.Eng. degrees from Nanjing University of Science and Technology, Nanjing, China, and the Ph.D. degree from City University of Hong Kong, all in electronic engineering, in 1992, 1995, and 2001, respectively. From 1995 to 1997, he was a Teaching and Research Assistant and then a Lecturer in the Department of Electronic Engineering, Nanjing University of Science and Technology. From January 1998 to August 1998, he was a Research Associate in the Department of Electronic Engineering, City University of Hong Kong. Since September 2001, he has been with the Institute for Infocomm Research, Singapore, as a Scientist. He also holds an appointment of Adjunct Assistant Professor at the National University of Singapore. He has published over 50 technical papers in international journals and conferences. He holds one Chinese Patent and one pending U.S. patent. His current research interests include design and modeling of microstrip and dielectric resonator antennas for handsets and other wireless communications, UWB antennas, LTCC ﬁlters and baluns and LTCC optoelectronic transceiver for radio-over-ﬁber application, and numerical methods in electromagnetics. Dr. Guo has served as a reviewer for IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION since 2002. He also served as one session chair for 2003 Asia and Paciﬁc Microwave Conference (APMC2003), Korea.

Michael Yan Wah Chia (M’94) was born in Singapore. He received the B.Sc. (1st Class Honors) and Ph.D. degrees from Loughborough University, U.K. He joined the Center for Wireless Communications (CWC), Singapore, in 1994 as a Member of Technical Staff (MTS), was promoted to Senior MTS, then Principal MTS, and ﬁnally Senior Principal MTS. He was also holding the appointment as Division Director of Radio Division. In 1999, he started UWB work at CWC, which was later merged to I R. Currently, he is the Division Director of Communications and Devices Division (which consists of ﬁve Departments), Institute for Infocomm Research (I R) of ASTAR (Agency of Science Technology And Research). Concurrently, he has been appointed as an Adjunct Associate Professor at the National University of Singapore. Externally, he has also served in the Public Service Funding Panel of ASTAR-2002, Radio Standard Committee and UWB Task Force Committee of Infocomm Development Authority(IDA) of Singapore. He was also on the Technical Program Committee International Workshop on UWB System (IWUWBS) 2003, Finland, and has been listed in Marquis’s Who’s Who in Engineering since 2002. He has published 28 international journal papers and 50 international conference papers. He has 10 patents both ﬁled and granted. Some of the patents has been commercialized and licensed to companies. His main research interest are ultraeide-band(UWB) system, antenna, transceiver, radio over ﬁber, RFIC, linearization and communication, and radar system architecture. Dr. Chia was awarded Overseas Research Studentship (ORS) and British Aerospace Studentship from the U.K.

1944

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Zhi Ning Chen (M’99) was born in China and received the B.Eng., M.Eng., and Ph.D. degrees in electrical engineering from the Institute of Communications Engineering (ICE), China, and the Ph.D. degree from the University of Tsukuba, Japan. From 1988 to 1995, he was with the ICE and was appointed Teaching Assistant, Lecturer, and then promoted to Associate Professor. After that, he was with the Southeast University (SEU), Nanjing, China, as a Postdoctoral Fellow and then appointed Associate Professor. From 1995 to 1997, he undertook his research in the City University of Hong Kong, China, as a Research Assistant, Research Associate, Senior Research Associate, and then Research Fellow. From 1997 to 1999, he pursued research at the University of Tsukuba, Japan, with the Fellowship awarded by Japan Society for Promotion of Science (JSPS). He visited SEU in 2000 and 2001, as a Visiting Scholar awarded by the Ministry of Education, China. In 2001, he visited the University of Tsukuba, Japan, again under the Invitation Fellowship Program (senior level) of JSPS. In 1999, he joined the Centre for Wireless Communications (CWC) (later known the Institute for Communications Research (ICR) and now the Institute for Infocomm Research (I R)) as a Member of Technical Staff (MTS), Senior MTS, and then promoted Principal MTS. Currently, he is working as a Lead Scientist and a Manager for Department of Radio Systems. He is concurrently teaching and supervising postgraduate students at the National University of Singapore (NUS), as an Adjunct Associate Professor. Since 1990, he has authored and coauthored over 110 technical papers published in international journals and presented at international conferences. One of his patents has been granted and four are pending. His main research interests include applied computational electromagnetics, and antenna theory and designs. Currently, he is focused on small, broadband, lightweight antennas for wireless systems and ultrawide-band (UWB) radio systems, and metamaterials and their applications. He managed a research project on Small and Multiband Antennas during 2000 to2003. Dr. Chen was a Member of the technical program and organizer/chair of UWB technology workshop at IEEE Radio and Wireless Conference (RAWCON), 2003. He also organized and chaired special session on Antennas for UWB Wireless Communication Systems at the IEEE International Symposium on Antennas and Propagation (AP-S), 2003 and IEEE Asia and Paciﬁc Microwave Conference (APMC), Korea, 2003.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

1945

**Miniature Reconﬁgurable Three-Dimensional Fractal Tree Antennas
**

Joshua S. Petko, Student Member, IEEE, and Douglas H. Werner, Senior Member, IEEE

Abstract—This paper introduces a design methodology for miniature multiband as well as reconﬁgurable (i.e., tunable) antennas that exploits the self-similar branching structure of three-dimensional (3-D) fractal trees. Several fundamental relationships, useful for design purposes, are established between the geometrical structure of the fractal tree antenna and its corresponding radiation characteristics. In particular, it will be shown that the density and elevation angle of the branches play a key role in the effective design of miniature 3-D fractal tree antennas. Several design examples are considered where fractal trees are used as end-loads in order to miniaturize conventional dipole or monopole antennas. Multiband and reconﬁgurable versions of these miniature antennas are also proposed, where either reactive LC traps or RF switches are strategically placed throughout the branches and/or along the trunk of the trees. Included among these designs is a miniature reconﬁgurable dipole antenna that achieves a 57% size reduction for the center frequency of the lowest intended band of operation and has a tunable bandwidth of nearly 70%. Index Terms—Fractal antennas, fractal tree antennas, miniature anennas, reconﬁgurable antennas.

Fig. 1. First 4 iterations of the four-branch class of fractal tree antennas [7]. Also included is a pictorial representation of the nomenclature used to describe fractal trees. TABLE I PARAMETERS FOR GENERATING FOUR-BRANCH, 30 FRACTAL TREES

I. INTRODUCTION

A

S PART OF AN effort to further improve modern communication system technology, researchers are now studying many different approaches for creating new and innovative antennas. One technique that has received a lot of recent attention involves combining aspects of the modern theory of fractal geometry with antenna design. This rapidly growing area of research is known as fractal antenna engineering [1], [2]. One particular class of antenna conﬁgurations that have been studied recently is based on fractal trees [2]–[10]. Fractal tree structures can be exploited in antenna designs to produce multiband characteristics [1], [3]–[5], [9] or to achieve miniaturization [6]–[10]. A set of dipole antennas that use fractal tree structures as end loads to achieve a resonant frequency lower than a standard dipole of comparable length have been recently studied in [6]–[10]. This paper begins by discussing ways to improve antenna miniaturization techniques that employ fractal tree geometries as end loads by increasing the density of branches (i.e., by using trees with a higher fractal dimension). Several miniaturization schemes for fractal tree antennas are introduced, which are based on various combinations of different branch lengths or angles. The addition of a center stub is also considered as a means for improving existing designs for miniature fractal tree

Manuscript received February 10, 2003; revised September 26, 2003. The authors are with the The Pennsylvania State University, Department of Electrical Engineering, University Park, PA 16802 USA (dhw@psu.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832491

antennas. Finally, the unique self-similar wire branch structure of the three-dimensional (3-D) fractal tree is exploited to develop new design methodologies for reconﬁgurable miniature dipole and monopole antennas. Several design examples are considered where these miniature fractal tree antennas are made reconﬁgurable by the introduction of strategically placed reactive loads or RF switches. Among these designs is a reconﬁgurable miniature dipole antenna that achieves a 57% size reduction with respect to its conventional counterpart (i.e., a half-wave dipole designed for the center frequency of the lowest operating band) and has a tunable bandwidth of nearly 70%. II. FRACTAL TREE RADIATION STUDIES A. Dense Fractal Tree Generators Fractals are objects which have a self-similar structure repeated throughout their geometry [11], [12]. This self-similar structure may be produced by the repeated application of a generator, and in the case of fractal trees, the generator is deﬁned as a junction from which several smaller branches, called child branches, split from a parent branch. Every branch, with the exception of the ﬁrst and ﬁnal branches, has a generator connected

0018-926X/04$20.00 © 2004 IEEE

1946

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE II PARAMETERS FOR GENERATING SIX- AND EIGHT-BRANCH FRACTAL TREES

Fig. 2. First two iterations of the six-branch and eight-branch classes of fractal tree antennas.

to it at each end: one from which it is a child and the other to which it is the parent. Three different families of 3-D fractal trees are ﬁrst evaluated for their suitability as miniature dipole antennas. These classes of antennas will be referred to throughout the paper as four-branch, six-branch, and eight-branch fractal trees. The four-branch class of antennas, shown in Fig. 1, has been adapted from [7] and is used as a benchmark of comparison for the six-branch and the eight-branch classes introduced in this paper. Table I shows the generation parameters for the four-branch fractal trees shown in Fig. 1. The fractal generators of these antennas have several properties in common. First, the child branches are half the length of the parent branches from which they separate. In addition, child branches bend 30 from the direction the parent branch is aimed. Also, all child branches have equal angles separating them (i.e., for the four-branch class, there are 90 between each child branch, for the six-branch class, there are 60 between each branch, and for the eight-branch class, there are 45 between each branch). Finally, one of these child branches must continue the arc path of its parent and the parent branch before that. The generation parameters are listed in Table II for both the six- and eight-branch cases. The similarities between four-, six-, and eight-branch antenna classes are evident from Figs. 1 and 2. Several of the other fractal tree design parameters are also held constant for the sake of comparison. First, the distance from tip to tip for each fractal tree dipole antenna is ﬁxed at 7.5 cm (or 3.75 cm from tip to source). More speciﬁcally, the base or trunk length of the ﬁrst stage is 2.5 cm, the second stage is 2.143 cm,

the third stage is 2 cm, and the fourth stage is 1.936 cm. In addition, the diameter of the wire assumed for each antenna is . Finally, all of the fractal tree dipole antennas are assumed to be center-fed. The resulting radiation characteristics of these fractal tree antennas were evaluated using a numerically rigorous approach for based on the method of moments (MoM). Fig. 3 plots the the ﬁrst two stages of the six- and eight-branch antenna classes and compares them to the ﬁrst four stages for the four-branch class. The results show that the six-branch antennas have a resonant frequency 100 MHz lower than the four-branch antenna at the same fractal stage of growth, whereas the eight-branch antenna has a resonant frequency approximately 150 MHz lower. The six- and eight-branch classes follow the same trend as the four-branch in the sense that the resonant frequency decreases with an increase in the fractal stage of growth; however, there is no valid geometry for the six- and eight-branch antenna classes beyond stage 3, because wire segments will intersect. Nevertheless, at these lower stages of growth, the six- and eight-branch antenna classes are more effective than the four-branch and are easier to fabricate than higher order four-branch antennas because the six- and eight-branch classes have less junctions. Despite the complicated geometry of these fractal tree structures, the radiation patterns exhibited by these types of antennas are very similar to those of typical dipole antennas. Fig. 4 shows the radiation patterns produced by the four-, six-, and eight- branch fractal tree dipoles compared to the radiation pattern of a conventional half-wave dipole. The radiation patterns for these fractal tree antennas have 150 dB) negligible cross-polarization components (i.e., and are nearly identical to the radiation pattern of the half-wave dipole antenna. B. Fractal Tree Generators of Varying Angle In this section, several different 3-D fractal trees are evaluated for their suitability as miniature dipole antennas at their second and third stages of growth. These antennas are related to the four-branch class of antennas adapted from [7]; however, the elevation angle that the child branches bend from the parent branch is not held constant at 30 but is varied over a range of angles from 10 to 90 . All other independent design parameters are assumed to be the same for each antenna. For instance, the antennas are all center fed dipoles with fractal tree loads placed on both ends, each generator has only four equally spaced branches, child branches are half the length of

PETKO AND WERNER: MINIATURE RECONFIGURABLE 3-D FRACTAL TREE ANTENNAS

1947

Fig. 3.

Comparison of S

versus frequency for four-branch, six-branch, and eight-branch fractal trees. The S

**was calculated with respect to 50
**

line.

Fig. 5. Third iteration of four-branch fractal tree antennas with the elevation angle at 10 , 30 , 45 , 60 and 90 respectively. TABLE III PARAMETERS FOR GENERATING FOUR-BRANCH FRACTAL TREES WITH VARYING ELEVATION ANGLE

Fig. 4. Radiation patterns of dense fractal tree antennas compared to those of a conventional half-wave dipole.

the parent branches, and the arc length of wire from tip to source is 3.75 cm. Fig. 5 shows how modifying this angle can dramat-

ically change the shape of the antenna. The generation parameters for each case considered in Fig. 5 are listed in Table III. Next, we consider variations of a second stage and a third stage four-branch fractal tree antenna. In both cases, the elevation angle is varied from 10 to 90 and each antenna that results is individually simulated. From this data, the value of voltage standing wave ratio (VSWR) that corresponds to the resonant frequency of each fractal tree antenna is obtained and plotted as shown in Fig. 6. The results for both stages show that fractal tree antennas with small elevation angles have a lower VSWR but a higher resonant frequency than those with large elevation angles. The resonant frequency continues to move lower as the elevation angle increases until it reaches approximately 50 . From that point, the resonant frequency begins to increase again. Throughout the entire range of elevation angles that were considered for each stage, the VSWR increases as the elevation angle increases. Trends uncovered by the elevation angle study can be used to design effective miniature fractal tree dipoles that

1948

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 6. Fractal tree dipole VSWR versus resonant frequency for elevation angles ranging from 10 to 90 . The VSWR was calculated with respect to 50

line.

also possess good VSWR performance characteristics. The radiation patterns, not shown in this case, have very similar characteristics to the patterns illustrated in Fig. 4. III. FRACTAL TREE ANTENNA DESIGNS A. Hybrid Fractal Tree Antenna While the eight-branch class of fractal tree antennas considered in Section II does not have a valid third stage of growth, modiﬁcations can be performed on the generator in order to make a valid self-avoiding third stage possible. As shown by the generation parameters listed in Table IV, the child branches at 0 , 90 , 180 , and 270 are 50% as long as the parent branch, which is the same as the previous antenna classes; however, the child branches at 45 , 135 , 225 , and 315 are only 40% as long as the parent branch. This difference in scale creates a hybrid pattern of short and long wires, which also impacts the length of wires at later stages of growth (see Fig. 7). All other factors in the design remain unchanged, and the distance of the longest possible path from the source to a wire end is 3.75 cm (7.5 cm end-to-end). This class of antenna is not as dense as the eight-branch class for stages one and two; however, its advantage is its self-avoiding geometry at stage three. results (see Fig. 8) show a reduction in the resonant The frequencies for the hybrid class of antennas when compared to the four-branch class of antennas for the same stage of growth: 100 MHz difference for the ﬁrst stage, 80 MHz difference for the second, and 60 MHz difference for the third. Fig. 8 demonstrates that the resonant frequency of the third stage hybrid antenna is about the same as the resonance of the fourth stage four-branch antenna, illustrating that denser fractal tree structures can effectively reduce resonant frequency in a similar manner as less dense fractal tree structures at higher stages of growth. As in the previous cases considered, the radiation patterns (not shown) are nearly indistinguishable from the pattern of a conventional half-wave dipole antenna and possess essentially no cross-polarized ﬁeld components.

TABLE IV PARAMETERS

FOR GENERATING EIGHT-BRANCH HYBRID SELF-AVOIDING FRACTAL TREES

SCALED

Fig. 7. First three iterations for the eight-branch hybrid scaled self-avoiding class of fractal tree antennas.

B. Center-Stubbed Fractal Tree Antennas It will be demonstrated here that fractal tree antennas can be designed to have a reduced resonant frequency and low reﬂection properties by incorporating a center stub. As shown previously in Section II-B, fractal tree dipole antennas can be designed to have low values of reﬂection by using generator branching schemes with sufﬁciently small elevation angles. Also, the resonant frequency of fractal tree dipole antennas may be reduced by considering geometries which have a more dense conﬁguration of branches. In this manner a center stubbed fractal tree antenna can take into account both of these design

PETKO AND WERNER: MINIATURE RECONFIGURABLE 3-D FRACTAL TREE ANTENNAS

1949

**Fig. 8. Comparison of respect to 50
**

line.

S

for eight-branch hybrid fractal tree dipoles with standard four-branch and eight-branch fractal trees. The

S

was calculated with

considerations by increasing the density of the fractal branching structure while, at the same time, keeping the elevation angles small with respect to the generator. Here we consider two generators, one with and one without a center stub. Each generator has four branches that bend out with elevation angles of 45 that are 50% as long as the previous branch. The branches of each generator are equally spaced in the phi direction; however, the generator is rotated an additional 45 so that they are offset from the path of the previous two branches. For this reason, the rotational angles are offset from 0 , 90 , 180 and 270 to 45 , 135 , 225 , and 315 respectively. A fractal tree antenna with a generator having only these 4 branches is compared to a fractal tree antenna with a generator having these 4 branches and an additional branch as a center stub. This center stub continues as an extension of the previous branch, but is rotated by 45 . In this way the child branches of the center stub are offset by 45 from the branches of the level below it. All other parameters in the design remain unchanged. The distance of the longest possible path from the source to an end of a wire is 3.75 cm (7.5 cm end-to-end). These structures are shown in Fig. 9 and the corresponding generation parameters are listed in Table V. The MoM simulation results indicate that fractal tree antennas with center stubs, when compared to fractal trees without a center stub, exhibit a downward shift in the resonant frequency with only a minimal increase in the reﬂection at resonance. The characteristics (see Fig. 10) show that for each of the ﬁrst three stages of growth there is a 60 MHz reduction in resonant frequency for fractal tree antennas with a center stub compared to those without a center stub. The results plotted in Fig. 10 also show that for each of the ﬁrst three stages of growth there is only (between 3 dB and 1 dB) a minimal increase in the level of for tree antennas with a center stub as opposed to those without a center stub. Thus, fractal tree antennas with center stubs can be very effective designs because of the similar efﬁciency and the lower resonant frequency. Finally, the radiation patterns for miniature fractal tree antennas with center stubs were also

Fig. 9. An example of a four-branch, center-stubbed fractal tree antenna.

found to be comparable to the radiation pattern produced by a conventional half-wave dipole antenna. C. Six Branch 50 –30 Fractal Tree Antenna In this section we consider an approach for designing miniature fractal tree dipole antennas that have relatively low values . This is achieved by using a combination of different of branches with small elevation angles to increase the density or space-ﬁlling property of the end-loads. One particular design considered here has 6 branches, with a 60 rotational angle between each branch of the generator. The elevation angles of the generator alternate between 30 and 50 , and the 50 branch is aligned with the 0 rotational angle. Fig. 11 shows these structures in detail and Table VI lists the corresponding generation parameters. Finally, all other design parameters are kept the same as before and the length of the antenna is 3.75 cm from tip to source. plots shown in Fig. 12 indicate that there The resulting is a signiﬁcant reduction in the resonant frequency when compared to the four-branch 30 fractal tree antenna evaluated in [7]. The second iteration six-branch 50 –30 antenna has a resonant frequency of 920 MHz, which is the same as the third

1950

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE V PARAMETERS FOR GENERATING FOUR-BRANCH, 45 FRACTAL TREES WITH AND WITHOUT CENTER STUB

Fig. 10. S versus frequency for a four-branch, 45 fractal tree dipole antenna (stage 1, stage 2, and stage 3) with and without a center stub. The calculated with respect to 50

line.

S

was

four-branch fractal tree antenna. The radiation patterns for a stage 1, stage 2, and stage 3 six-branch 50 –30 fractal tree dipole antenna are again similar to those illustrated in Fig. 4 with negligible cross-polarization. IV. RECONFIGURABLE FRACTAL TREE ANTENNAS A. Self-Reconﬁgurable Reactive Loaded Fractal Tree Monopoles It has been shown that reactive loads acting as traps can be used to make a typical monopole antenna resonant at more than one frequency [13], [14]. These reactive loads behave as an open circuit at some frequencies and a short circuit at others, effectively making an antenna self-reconﬁgurable (i.e., reconﬁgurable without the need for RF switches). This concept can also be applied to fractal tree monopoles to produce an antenna that not only is resonant at more than one frequency but also is miniature in size due to the presence of the space-ﬁlling end-load structure. The ﬁrst case study evaluates a monopole version of the third stage four-branch fractal tree, this time including parallel LC reactive load elements on the base (i.e., trunk) of the antenna. By placing the reactive loads below the fractal tree end-load, the antenna can be designed so that the feed effectively sees the entire antenna at the lowest resonant frequency and only parts of the base at higher resonant frequencies. Both a dual-band and a tri-band version of the third stage four-branch fractal tree

Fig. 11.

First three stages of a six-branch 50 –30 fractal tree dipole antenna.

TABLE VI PARAMETERS FOR GENERATING SIX BRANCH 50 –30 FRACTAL TREES

iteration for the four-branch fractal tree antenna. For the third iteration six-branch 50 –30 antenna, the resonant frequency is 790 MHz, 70 MHz lower than the fourth iteration of the

PETKO AND WERNER: MINIATURE RECONFIGURABLE 3-D FRACTAL TREE ANTENNAS

1951

Fig. 12. S versus frequency for a stage 1, stage 2, and stage 3, six-branch 50 –30 fractal tree antenna. Also shown for comparison is the S stages of a standard four-branch fractal tree antenna. The S was calculated with respect to 50

line. TABLE VII LOAD COMPONENT VALUES FOR THE DUAL- AND TRI-BAND FOUR-BRANCH FRACTAL TREE MONOPOLE ANTENNAS

for the ﬁrst four

Fig. 13. Load locations on dual- and tri-band four-branch fractal tree monopole antennas.

antenna are considered here. For the dual-band antenna, one resonant load is placed near the top of the fractal tree’s base, 1.95 cm above the ground plane along the 2-cm tall trunk. For the tri-band antenna, an additional load is placed on the base at a height of 1.65 cm above the ground plane. Fig. 13 illustrates the load locations on the dual- and tri-band monopole antennas and Table VII lists the corresponding load compoversus frequency are shown in nent values. Plots of the Fig. 14 for the single-band unloaded fractal tree monopole, the dual-band fractal tree monopole with one reactive load, and the

tri-band fractal tree monopole with two reactive loads. The reactive loads also have another advantage in further reducing the minimum resonant frequencies of the antenna structure. The unloaded fractal tree structure has one primary resonance at 910 MHz, the dual-band antenna has two primary resonances at 800 and 2460 MHz, and the tri-band antenna has three primary resonances at 550, 2300, and 5240 MHz. Also, fundamental or primary resonant frequencies can be distinguished from secondary resonant frequencies in that there is only one main lobe present in the radiation pattern at a primary resonant frequency. Fig. 15 shows the radiation patterns at each resonant frequency within the 400 to 6400 MHz band. In the plot for the unloaded fractal tree, we see that there is a primary resonance at 910 MHz, characterized by a single lobe in the radiation pattern, and a secondary resonance at 6100 MHz, characterized by two lobes in the radiation pattern. For the fractal tree with one load, the primary resonances are conﬁrmed to be at 800 MHz and 2460 MHz while there is a secondary resonance at 6230 MHz. Finally for the fractal tree with two loads, primary resonances occur at 550, 2300, and 5240 MHz. The next case study considers the third stage of a four-branch, 45 fractal tree monopole with center stub. In this instance, series LC traps are placed inside the fractal tree end-load structure instead of, as in the previous instance, parallel LC traps being placed on the trunk or base of the tree. The capability of placing

1952

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 14. S versus frequency for self-reconﬁgurable four-branch fractal tree monopoles. The response of the single-band unloaded fractal tree monopole is compared to a dual-band and a tri-band version of the same antenna with one and two reactive loads respectively. The S was calculated with respect to 50

line.

Fig. 17. Matching network for the tri-band four-branch, center stubbed fractal tree monopole antenna.

Fig. 15. Radiation patterns for self-reconﬁgurable four-branch fractal tree monopoles.

Fig. 16. Load locations on a tri-band four-branch, center stubbed fractal tree monopole.

LC traps in the end-load structure is an important advantage of the fractal tree antenna. Resonant traps can be placed on the base of any end-loaded monopole antenna; however, by placing

the traps only on the base, the antennas will exhibit a characteristically large separation between the lowest operating band and any higher frequency bands. Having the ability to place the traps in the branches of the fractal tree structure provides more ﬂexibility in the design of miniature multiband monopole antennas by allowing the resonances to be placed closer together. Using ﬁve distinct reactive LC traps, a tri-band version of the third stage four-branch, 45 fractal tree monopole with center stub is presented and discussed here. Four traps are placed near the top of the ﬁrst stage of the outer four branches. A ﬁfth trap is placed near the bottom of the ﬁrst stage of the center stub. Fig. 16 shows the position of the traps and Table VIII provides the corresponding component values. The antenna has three primary resonances at 330 MHz, 800 MHz, and 2220 MHz. The separation between the ﬁrst and third resonant frequencies is only 1890 MHz, 2800 MHz less than the separation between the ﬁrst and third resonant frequencies of the tri-band four-branch fractal tree shown in Fig. 13. In this case, a ﬁve-element matching network can be used to allow the antenna to perform efﬁciently at each resonant frequency. The topology of a candidate matching network is shown in Fig. 17 along with the required component values. The candidate matching network was designed using a trial and error approach in combination with a random variable

PETKO AND WERNER: MINIATURE RECONFIGURABLE 3-D FRACTAL TREE ANTENNAS

1953

TABLE VIII LOAD COMPONENT VALUES FOR THE TRI-BAND FOUR-BRANCH, CENTER STUBBED FRACTAL TREE MONOPOLE ANTENNA SHOWN IN FIG. 16

Fig. 18. S versus frequency comparison of matched and unmatched tri-band four-branch, center stubbed fractal tree monopole antennas. The S with respect to 50

line.

was calculated

tion patterns at 300 MHz, 800 MHz, and 2220 MHz each only have a single lobe, indicating that all three cases represent primary resonant frequencies. B. Miniature Reconﬁgurable/Tunable Fractal Tree Antennas Using Electronic Switches Recently there has been a considerable amount of interest in design concepts for reconﬁgurable antennas capable of operating over a broad range of frequencies. In this section a reconﬁgurable antenna design approach will be introduced that exploits the branching structure of fractal tree dipoles. Because the current is distributed over the entire end-load, the effective removal or switching off of even relatively large sections of the fractal tree structure will reduce the resonant frequency only by small amounts. This property of reconﬁgurable fractal tree antennas can be exploited to allow the resonances of the antenna to be spaced close enough together to cover all of the frequencies between adjacent bands. Second, because the fractal tree structure has many parallel paths for the current to ﬂow, then only a relatively small number of switches may be required to achieve tunability over a desired range of frequencies. A design is presented here where RF switches are strategically placed on the third stage, six-branch 50 –30 fractal

Fig. 19. Radiation patterns for each band of the tri-band four-branch, center stubbed fractal tree monopole antenna shown in Fig. 16.

optimizer [15]. The plots for the matched and unmatched tri-band center-stubbed fractal tree monopoles are compared in Fig. 18. These plots demonstrate that by placing a reactive matching network at the feed of the antenna, all three resonances can be matched to operate at a VSWR under 2:1. Finally, Fig. 19 shows the radiation patterns at each of the resonances of the tri-band center stub fractal tree monopole antenna. The radia-

1954

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 21. Several states for the reconﬁgurable six-branch, 50 –30 fractal tree dipole antenna. For illustrative purposes, the branches that are switched off for each of the ﬁve different bands have been removed.

Fig. 20. Switch layout for the reconﬁgurable six-branch, 50 –30 fractal tree dipole antenna.

tree dipole antenna to make it reconﬁgurable (i.e., tunable) over a bandwidth of 68%. The antenna uses 204 separate switches placed throughout both tree structures (a total of 102 on each end) to produce 20 reconﬁgurable states. The locations of these RF switches are illustrated in Fig. 20. The switches are placed at every junction inside the tree structure with the exception of the junction joining the base to the ﬁrst stage. The junctions joining the ﬁrst and second stages all have six switches associated with each of the six branches. In addition the 36 junctions between the second and third stage of the fractal structure have switches associated with each of them. The six junctions between the second and third stages that are nearest to the center axis of the antenna (i.e., those that are closest to being vertical) also have six switches associated with each of the six branches. The remainder of the junctions between the second and third stage of the fractal tree have only one switch, which is placed near the end of the branch below the junction. For this particular design, there are 20 different combinations of switch settings which correspond to 20 different resonant frequencies at which the reconﬁgurable antenna is capable of operating. A rigorous moment method simulation is used to individually model each of the 20 different states of the reconﬁgurable fractal tree antenna. The switches in the fractal tree structure are modeled as ideal. Also, the MoM model for each state is generated from what remains of the fractal tree antenna after removing the appropriate sections that have been switched off. This is justiﬁed since coupling effects between these removed sections, which are signiﬁcantly less than a half-wavelength long, and the remainder of the antenna still being directly fed are expected to be minimal. In an effort to maintain the omni-directional radiation pattern of the fractal tree antenna in the azimuthal plane, the order in which the switches are turned off is performed in a manner that preserves the symmetry of the tree about its base as much as possible. Fig. 21 illustrates several different states of the reconﬁgurable fractal tree antenna, visualized by removing portions of the fractal tree above switches that are turned off. The resulting antenna can be reconﬁgurable from 770 to 1570 MHz for a bandwidth of 800 MHz with a VSWR under 3:1

Fig. 22. S versus frequency for the reconﬁgurable six-branch, 50 –30 fractal tree antenna. The light gray curves represent each of the 20 states the antenna can be conﬁgured to operate at. The dark gray curves represent reconﬁgured states which operate as stage 1, stage 2, and stage 3 fractal tree antennas. The black line represents the overall minimum S the antenna can operate over the entire band. The S was calculated with respect to 50

line.

and is reconﬁgurable from 970 to 1570 MHz for a bandwidth of 560 MHz with a VSWR below 2:1. In Fig. 22 each of the 20 reconﬁgurable states is represented by a separate curve (indicated by light gray lines) with the lowest resonant frequency representing the state with all the switches closed and the highest resonant frequency representing the state with all the switches open. The remaining states are achieved by opening the switches progressively from the top to the bottom. In addition, for three of the reconﬁgurable states the antenna effectively operate as a 50 –30 fractal tree dipole with fractal curves for these three special cases stages 3, 2, or 1. The are indicated on the graph by thick dark gray lines. Finally the the antenna solid black line represents the overall minimum can be conﬁgured to for a particular frequency over the entire operating range of the antenna. Fig. 23 presents an overlay of the co-polarized and cross-polarized radiation patterns plotted at the resonance for each state of the reconﬁgurable 50 –30 fractal tree dipole. Because of the asymmetry in the tree geometry for several of the reconﬁgurable states, the cross-polarized patterns corresponding to these particular states are somewhat higher 30 dB), but still remain within acceptable limits. The ( co-polarized radiation pattern plots are again seen to closely resemble those of a conventional linear half-wave dipole. Finally,

PETKO AND WERNER: MINIATURE RECONFIGURABLE 3-D FRACTAL TREE ANTENNAS

1955

Fig. 23.

Co-polarized and cross-polarized radiation patterns for each state of the reconﬁgurable six-branch, 50 –30 fractal tree dipole antenna.

we note that a monopole version of this antenna can also be created with half the number of switches. In this case a broadband or tunable matching network would be also required. V. CONCLUSION This paper begins by investigating the relationship between the geometrical structure of fractal tree antennas and their corresponding radiation characteristics. It was found, through a series of systematic MoM simulations, that the two most critical factors inﬂuencing the successful design of miniature 3-D fractal tree antennas appear to be the density and elevation angle of their branches. These observations subsequently led to the development of several new design conﬁgurations that employ fractal tree end-loads as a means of miniaturizing conventional dipole or monopole antennas. Hybrid fractal trees and centerstubbed fractal trees represent two types of end-load structures that have proven to be particularly effective in achieving a signiﬁcant amount of size reduction. Next, multiband and reconﬁgurable versions of these miniature antennas are introduced, where either reactive LC traps or RF switches are strategically placed throughout the branches and/or along the trunk of the

trees. A prime advantage of placing the traps or RF switches at critical locations on branches of the fractal tree end-loads is that the resonant frequencies of the antenna can be spaced much closer together than if they were applied only to the tree trunks. Among the designs considered was a miniature reactively loaded tri-band center stubbed fractal tree monopole with associated matching network. Also considered was a design for a miniature reconﬁgurable dipole antenna that achieves a 57% size reduction for the lowest intended band of operation and has a tunable bandwidth of nearly 70%. REFERENCES

[1] D. H. Werner, R. L. Haupt, and P. L. Werner, “Fractal antenna engineering: the theory and design of fractal antenna arrays,” IEEE Antennas Propagat. Mag., vol. 41, pp. 37–59, Oct. 1999. [2] D. H. Werner and R. Mittra, Frontiers in Electromagnetics. Piscataway, NJ: IEEE Press, 2000. [3] M. Sindou, G. Ablart, and C. Sourdois, “Multiband and wideband properties of printed fractal branched antennas,” IEE Electron. Lett., vol. 35, no. 3, pp. 181–182, Feb. 1999. [4] C. Puente, J. Claret, F. Sagues, J. Romeu, M. Q. Lopez-Salvans, and R. Pous, “Multiband properties of a fractal tree antenna generated by electrochemical deposition,” IEE Electron. Lett., vol. 32, no. 25, pp. 2298–2299, Dec. 1996.

1956

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[5] D. H. Werner, A. R. Bretones, and B. R. Long, “Radiation characteristics of thin-wire ternary fractal trees,” IEE Electron. Lett., vol. 35, no. 8, pp. 609–610, Apr. 1999. [6] J. P. Gianvittorio and Y. Rahmat-Samii, “Fractal element antennas: a compilation of conﬁgurations with novel characteristics,” in Proc. IEEE Antennas and Propagation Society Int. Symp., vol. 3, Salt Lake City, UT, July 2000, pp. 1688–1691. [7] J. P. Gianvittorio, “Fractal Antennas: Design, Characterization, and Applications,” M.S. thesis, Dept. Elect. Eng., University of California Los Angeles, 2000. [8] J. P. Gianvittorio and Y. Rahmat-Samii, “Fractal antennas: a novel antenna miniaturization technique, and applications,” IEEE Antennas Propagat. Mag., vol. 44, Feb. 2002. [9] D. H. Werner and S. Ganguly, “An overview of fractal antenna engineering research,” IEEE Antennas Propagat. Mag., vol. 45, pp. 38–57, Feb. 2003. [10] J. S. Petko and D. H. Werner, “Dense 3-D fractal tree structures as miniature end-loaded dipole antennas,” in Proc. IEEE Antennas and Propagation Society Int. Symp., vol. 4, San Antonio, TX, June 2002, pp. 94–97. [11] B. B. Mandelbrot, The Fractal Geometry of Nature. New York: Freeman, 1983. [12] H. O. Peitgen, H. Jurgens, and D. Saupe, Chaos and Fractals: New Frontiers of Science. New York: Springer-Verlag, 1992. [13] A. Boag, E. Michielssen, and R. Mittra, “Design of electrically loaded wire antennas using genetic algorithms,” IEEE Trans. Antennas Propagat., vol. 44, pp. 687–695, 1996. [14] Z. Altman, R. Mittra, P. L. Werner, and D. H. Werner, “Application of genetic algorithm to boradband antenna design,” in Electromagnetic Optimization by Genetic Algorithms, Y. Rahmat-Samii and E. Michielssen, Eds. New York: Wiley, 1999, pp. 181–210. [15] Microwave Ofﬁce, Applied Wave Research Inc., 2000.

Joshua S. Petko (S’02) was born in Brownsville, PA, in 1979. He received the B.S. degree in electrical engineering from The Pennsylvania State University, University Park, in 2002, where he is currently working toward the M.S. degree. Currently, he is a Research Assistant for the Communications and Space Sciences Laboratory and the Applied Research Laboratory, Pennsylvania State University. His research interests include antenna theory, computational electromagnetics, fractal electrodynamics, and evolutionary algorithms with a focus on fractal antenna elements and arrays. Mr. Petko is a Member of Eta Kappa Nu. He has been awarded the 2002 James A. Barnak Outstanding Senior Award from the Pennsylvania State University Eta Kappa Nu chapter and was a ﬁnalist and an honorable mention for the Alton B. Zerby and Carl T. Koerner National Outstanding Senior Award from Eta Kappa Nu. He also has been awarded second place in both the 2002 Penn State Undergraduate Poster Competition and the 2002 IEEE Region 2 Student Paper Competition.

Douglas H. Werner (S’81–M’89–SM’94) received the B.S., M.S., and Ph.D. degrees in electrical engineering and the M.A. degree in mathematics from The Pennsylvania State University (Penn State), University Park, in 1983, 1985, 1989, and 1986, respectively. He is an Associate Professor in the Department of Electrical Engineering, Penn State. He is a member of the Communications and Space Sciences Lab (CSSL) and is afﬁliated with the Electromagnetic Communication Research Lab. He is also a Senior Research Associate in the Electromagnetics and Environmental Effects Department of the Applied Research Laboratory at Penn State. He is a former Associate Editor of Radio Science. He has published numerous technical papers and proceedings articles and is the author of nine book chapters. He is an Editor of Frontiers in Electromagnetics (Piscataway, NJ: IEEE Press, 2000). He also contributed a chapter for Electromagnetic Optimization by Genetic Algorithms (New York: Wiley Interscience, 1999). His research interests include theoretical and computational electromagnetics with applications to antenna theory and design, microwaves, wireless and personal communication systems, electromagnetic wave interactions with complex media, meta-materials, fractal and knot electrodynamics, and genetic algorithms. Dr. Werner is a Member of the American Geophysical Union (AGU), International Scientiﬁc Radio Union (URSI) Commissions B and G, the Applied Computational Electromagnetics Society (ACES), Eta Kappa Nu, Tau Beta Pi, and Sigma Xi. He received the 1993 Applied Computational Electromagnetics Society (ACES) Best Paper Award and a 1993 URSI Young Scientist Award. In 1994, he received the Pennsylvania State University Applied Research Laboratory Outstanding Publication Award. He received a College of Engineering PSES Outstanding Research Award and Outstanding Teaching Award in March 2000 and March 2002, respectively. He recently received an IEEE Central Pennsylvania Section Millennium Medal. He has also received several Letters of Commendation from Penn State’s Department of Electrical Engineering for outstanding teaching and research. He is an Editor of IEEE ANTENNAS AND PROPAGATION MAGAZINE.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

1957

**Investigations on Miniaturized Endﬁre Vertically Polarized Quasi-Fractal Log-Periodic Zigzag Antenna
**

Satish K. Sharma, Member, IEEE, and Lotfollah Shafai, Fellow, IEEE

Abstract—This paper presents the investigations on a miniaturized vertically polarized traveling wave antenna for operation in the high frequency band (3–6 MHz), with a speciﬁc requirement of keeping its height near 1 8th of a wavelength. The antenna is desired to have a good endﬁre gain and front to back ratio, and small radiation levels in the vertical direction at broadside angle. A log-periodic Zigzag antenna (LPZA) has acceptable performance in both gain and polarization. Its height however is large, at about one wavelength (1 ). The concept of fractal antenna is employed 8, in this antenna to achieve the necessary height reduction to while keeping its radiation characteristics nearly constant. Both single and dual arm quasifractal log-periodic zigzag antenna (QFLPZA) conﬁgurations are investigated, with a maximum antenna height of only 1 8th of a wavelength, showing the desired radiation characteristics, and a wide impedance bandwidth of 67%. This type of antenna may ﬁnd applications in surveillance radar. Index Terms—Miniaturized wire antenna, vertically polarized, log-periodic (LP), zigzag (Z), backﬁre radiation, quasi-fractal (QF).

I. INTRODUCTION

T

HE antennas whose current and voltage distributions can be represented by one or more traveling waves, usually in the same direction, are referred to as traveling wave antennas. This antenna radiates from a continuous source. There are various examples of traveling wave antennas, such as dielectric rod, helix, and various log periodic antennas (LPA) [1], most of which are suitable for microwave frequencies. In the present study, different variations of traveling wave antennas were considered, i.e., traveling wave linear [2] and V-antennas [3], sandwich wire antennas [4]–[6], meander line planar array antennas [7], [8], and most importantly log-periodic zigzag antennas (LPZA) [9], [10]. Another antenna of interest was the electrically short umbrella top-loaded antenna [11]. However, among these options, the log periodic zigzag antenna seemed to provide the desired antenna radiation characteristics hence was selected for further study. Some investigation results on this study were presented by the authors in [12]. Zigzag antennas are classiﬁed as periodic structures [9]. The basic zigzag antenna is shown in Fig. 1(a), in which is the pitch angle. Assuming that the current along the wire travels with free space phase velocity, the near ﬁelds of the zigzag

Fig. 1. Conﬁgurations of (a) basic zigzag traveling wave antenna, (b) single arm log-periodic zigzag traveling wave antenna, and (c) balanced LPZA [9], [10].

Manuscript received July 16, 2003; revised November 5, 2003. The authors are with the Department of Electrical and Computer Engineering, The University of Manitoba, Winnipeg, MB R3T 5V6, Canada (e-mail: shafai@ee.umanitoba.ca). Digital Object Identiﬁer 10.1109/TAP.2004.832317

wire will have a fundamental-wave phase constant given by: , where is the free space wave number. This monoﬁlar zigzag antenna produces a backﬁre radiation below its resonance frequency, and a broadside beam at its resonance. A variation of this zigzag antenna as a single arm LPZA is shown in Fig. 1(b). It shows the small cells termed as transmission line region, which correspond to the low-frequency condition, and causes little radiation to occur as the fundamental and all space harmonics are slow waves. The region where the ﬁrst reverse traveling space harmonics approaches the backﬁre condition produces appreciable radiation and corresponds

0018-926X/04$20.00 © 2004 IEEE

1958

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE I

to the active region. If the radiation is sufﬁcient, the larger cells will not be excited, which is desirable to avoid radiation in directions away from backﬁre [9]. Fig. 1(c) shows the geometry of the double arm thin wire LPZA [10]. The angle refers to the angle from tip to tip, and the angle between the planes of the zigzag elements is called . By controlling the angles and , the E- and H-plane beamwidths and antenna directivity can be controlled. On the other hand, the control of input impedance, beamwidth and other radiation characteristics may be achieved , and , where using different log-periodic parameters like is the geometric factor deﬁned as [Fig. 1(c)]. Hence, for the present investigation, the LPZA is selected as a possible candidate, and will be studied next. The computational tool used for this study is NEC [16], which is a full wave analysis software based on the method of moments (MOM). The wire radius is 4.00 mm for the entire investigation. The antenna and design requirements are shown in Table I, where are the gains along vertical (broadside) and horizontal (endﬁre) directions. II. DOUBLE ARM LPZA In this section, a LPZA is investigated [Fig. 1(c)] based on [9], [10], which will be the basis of comparison for the miniaturized antenna presented next. The most important design parameters for this LPZA are the geometrical factor, which is the ratio of two adjacent similar dimensions on the antenna , and angles and . As shown in Fig. 1(c), determines the antenna lengths measured from the apex and is the property of a frequency independent log-periodic antenna. For this antenna, the selected parameters were: , its maximum element height , and length [Fig. 1(c)] along the principal axis of its single arm, where is the free space wavelength at the lower end of the frequency band, i.e., , where MHz. This antenna height of is much larger than the desired height of . It was assumed that the maximum length of the antenna is not a problem. Hence, the study was devoted to the reduction of to near , without sacriﬁcing the antenna height from signiﬁcantly the antenna’s radiation performance. This required an extensive study. Design details for this antenna are not given here, as it was known in literature [9], [10]. For the purpose of computations, the antenna is placed over an inﬁnite ground plane at a height of and is fed at the input end with a voltage source w.r.t. the ground plane. Its impedance characteristics and radiation gain patterns for 67% frequency bandwidth are shown in Fig. 2(a) and (b), respectively. An examination of Fig. 2(a) reveals that, with frequency variation the resistance and reactance oscillate around 700 and 0 , respectively. Their maximum span, for both resistance and reactance, is around 400 and shows a fairly constant trend. The radiation gain characteristic is shown in Fig. 2(b) remaining

Fig. 2. Characteristics of a double arm LPZA (Fig. 1(c)), (a) Resistance and reactance variation with frequency, and (b) Radiation patterns ( = 0 plane) at different frequencies within 67% bandwidth. The antenna is lifted from ground plane by 0:002 .

between 10.50 to 12.50 dBi. The gain level at broadside angle dBi. The front-to-back (F/B) ratio is is always less than varying between 6.50 dB and 9.00 dB. Since, the radiation patit is also referred as the backﬁre antenna. tern peak is near Next, height miniaturization of this antenna is presented, employing similar to the Koch type of fractal element [13] concept for a single arm LPZA. III. ANTENNA MINIATURIZATION TECHNIQUES FOR SINGLE ARM In this section, the height miniaturization of a single arm LPZA is discussed, while keeping its radiation characteristics almost unaltered. The techniques used to reduce the antenna height are as follows. 1) Fold the lower half of the antenna along its axis, i.e., axis, by 90 , keeping the upper half vertical and making the plane (Fig. 3, Step II). The lower half horizontal, i.e., antenna length along its axis remains constant and equal to . Its height reduces to . 2) Fold back the antenna triangular arms, when their height exceeds (Fig. 3, Step III). The antenna length along its axis remains constant. Its height decreases to . 3) Fold back the arms again, when their length exceeds (Fig. 3, Step IV). The antenna length along its axis remains . constant. The antenna height decreases to 4) The entire antenna is lifted (20 cm) above ground plane to prevent shorting the horizontal arms to ground plane during its operation.

SHARMA AND SHAFAI: INVESTIGATIONS ON MINIATURIZED ENDFIRE VERTICALLY POLARIZED QFLPZA

1959

Fig. 3. Step by step antenna miniaturization technique. (a) Step I: original antenna height of H , (b) Step II: reduced height, H=2, (c) Step III: reduced height, H=4, and (d) Step IV: ﬁnal reduced height, H=8. In all cases, the antenna is lifted from ground plane by 0:002 .

Fig. 5. Conﬁguration of the reduced length single arm quasifractal LPZA (QFLPZA), L = 1:96 , and H = =8. The antenna is lifted from ground plane by 0:002 .

Fig. 4. Effect of antenna height reduction on vertically polarized radiation pattern ( = 0 plane) at different frequencies for single arm QFLPZA. (a) Step I and (b) Step IV.

In the above process, the elements exceeding a speciﬁed height limit are reduced by the quasifractal elements. The quasifractal element concept employed is similar to the Ist iteration of the Koch curve [13]. The condition that geometrical factor, remains unaltered makes sure that after reducing the height the antenna is still log-periodic. The horizontal wires parallel to the ground plane will not add much to the radiation, though they will cause the necessary phase delay to occur [14], [15], behaving as a transmission lines. Contrary to this, the vertical wires will directly contribute to the vertically polarized radiation due to vertical component of current present in them. Furthermore, after braking the original height of wires, for both horizontal and vertical wires and converting them into

fractal form, the resultant path lengths for propagation remains almost the same as the original antenna. Breaking the wires, however, modiﬁes the direction of the radiating currents and thus, will inﬂuence the antenna radiation characteristics. This is investigated and the resulting radiation patterns are shown in Fig. 4(a) and (b), respectively, for the original antenna and the ﬁnal reduced height model. An examination of Fig. 4(a) reveals that for the original antenna in Step I of Fig. 3, the backﬁre gain varies between 9.50 to 11.40 dBi, and the F/B ratio changes from 11.20 to 14.60 dB. The radiating gain in the vertical direction is below 6 dBi, about 15 dB below the main beam. Its performance after the full height reduction to (Step IV of Fig. 3) as shown in Fig. 4(b), remains satisfactory. The backﬁre gain is varying between 6.00 to 9.00 dBi, and the radiation in the vertical direction is below 2.50 dBi. The back radiation, however, seems sensitive to the operating frequency. An interesting phenomenon is the null ﬁlling in the main beam and elimination of the ﬁrst sidelobe. As shown inc Fig. 4(b) the resulting gain patterns become smoother, and the gain drops gradually away from

1960

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Radiation pattern ( = 0 plane) for the reduced length (L = 1:96 ) single arm QFLPZA at different frequencies. Fig. 6.

Fig. 8. Effect of varying the arm-to-arm angle on double arm reduced length (L = 1:96 )QFLPZA. (a) Gain and (b) F/B ratio.

Fig. 7. Conﬁguration of double arm reduced length QFLPZA shown in xyz views. L = 1:96 ; H = =8; = 25 . The antenna is lifted from ground plane by 0:002 .

the main beam. This antenna is termed as the quasi-fractal log periodic zigzag antenna (QFLPZA), and studied further for its length reduction and fractal length variations. A. Length Reduction In this subsection, the Step IV antenna is further investigated for reduction in its length along the axis, which originally was . Extensive simulations were run, by reducing the wire elements near the transmission line region, which do not contribute signiﬁcantly to the radiation. For the sake of brevity, these are not presented here, instead the results for an antenna along the -coordinate axis is with length reduced to presented, which eliminates of length, and still performs similar to the Step IV antenna. The antenna conﬁguration is shown in Fig. 5. It is clear that the angle as deﬁned in Fig. 1 is now irrelevant, except at the antenna input end. The geometrical factor , still deﬁnes the conﬁguration. To prevent shorting of horizontal arms to ground, the antenna is placed over an inﬁand is fed at the input nite ground plane at a height of end with a voltage source w.r.t. the ground plane. The computed data for radiation patterns at is shown in Fig. 6. The gain varies within 3.00 dB range between 9.00 and 6.00 dBi for different frequencies within the band. B. Effect of Fractal Element Lengths It is known that the introduction of bends in a line reduces its effective electrical length due to the corner effects. The

amount of the length reduction depends on the corner angle and increases as the corner angle reduces. In the proposed height reduction method the number of fractal and thus corners increases progressively along the antennas. However, since each fractal has a single corner, this can easily be compensated for by increasing the length of each fractal element. This phenomenon was studied at different frequencies within 2 % length variation of the fractals. The results showed a rapid degradation of the performance, at all frequencies, with the length reduction. The improvement in the performance, due to length increase, however, was very slow. Consequently, and the fact that the antenna height reduction was desired, the length increase was not incorporated. Hence, the fractal element lengths were kept as in the original antenna (Section III: Step IV, and III-A).

IV. DOUBLE ARM QFLPZA The double arm antenna is a more practical one, in which the H-plane beamwidth can be controlled by the angle between the arms. Thus, in this part a double arm QFLPZA is investigated using the single arm design of QFLPZA presented in the previous Section III-A. The geometry of the antenna is shown in Fig. 7. A parametric study was carried out to ﬁnd the effect of the arm-to-arm angle as deﬁned in Figs. 1(c) and 7 on the gain and F/B ratio. The results are shown in Fig. 8(a) and (b). Fig. 8(a) reveals that by increasing the angle the gain drops slowly at the low frequency end of 0.75 , whereas for it is nearly constant. For 1.25 the gain increases with frequency. Similarly, Fig. 8(b) shows that by increasing the F/B ratio drops rapidly for 0.75 f , whereas for and 1.25 , it is nearly constant. Thus, based on Fig. 8(a) and (b), the arm-to-arm angle is an important parameter. Near a tradeoff between

SHARMA AND SHAFAI: INVESTIGATIONS ON MINIATURIZED ENDFIRE VERTICALLY POLARIZED QFLPZA

1961

From Fig. 9(a) the resistance undergoes signiﬁcant variation between 200 and 1000 in the frequency range, but the reac, except at 0.825 , where it tance varies only between is around 680 . The input impedance was also plotted on Smith Chart with 575 of normalization characteristic impedance, and is shown in Fig. 9(b). From Fig. 9(b) it is evident that impedance is balanced and can be matched to a 50 circuit using a commercial RF impedance transformer of impedance ratio 50 /575 . The impedance plot lies at the center of the Smith chart. Only a few points lie outside 2:1 circle. This is further clear from Fig. 9(c), which shows the return loss, (dB), computed using a reference impedance of 575 . It dB) over the reveals a very good impedance match ( complete frequency band, except at 0.775 and 1.00 where it becomes poor. Fig. 9(d) shows the radiation patterns within the 67% frequency band. The antenna shows good backﬁre gain at all frequencies, except at the lower end frequencies, which is attributed to the reduced effective antenna height at these frequencies. The gain variation is within 3.60 dB ranging between 11.60 and 8.00 dBi within the band. The F/B ratio is between 6.00 and 9.00 dB. The gain in the vertical direction is less than 5 dBi, except at the center frequency. Its performance is superior to the single arm QFLPZA, and its gain has improved by at least 2.00 dB throughout the frequency range. Further effort may be needed to improve the F/B ratio. The antenna was intended to be transportable and easy to install. The entire antenna was, therefore, to be mounted on a ﬂexible net like material, using ﬂexible wires. To install, nonconducting rigid poles of proper heights, were to be installed at location of the vertical arms, with predetermined mounting hooks. The deployed antenna was to be mounted on the poles, in a trellis like structure. For transportation, the antenna to be removed from the poles, folded along bends, and simply rolled for storage as a bundle. The mounting and removal of the antenna requires negligible time. Only the installation of poles needs careful planning, to insure its survival during operation. V. CONCLUSION Vertically polarized antennas are an important class of antennas that are used efﬁciently over conducting ground planes, and monopoles are the most popular conﬁgurations. Their , however, becomes large at low frequencies. height of They are also narrow band. In this study, we have investigated the possibility of using LPZA to achieve wide impedance bands . The and modiﬁed their geometry to reduce the height to resulting antenna is a quasifractal antenna. Steps leading to the height reduction are described and their effects on the antenna performance are studied. While the height reduction process has deteriorated the front-to-back ratio at some frequencies, the overall antenna performance has remained remarkably good. The gain pattern becomes smoother and decreases gradually with the angle away from the main beam. REFERENCES

[1] C. A. Balanis, Antenna Theory and Design, 2nd ed, New York: Wiley. [2] D. P. Nyquist and K. Chen, “The traveling wave linear antenna with nondissipative loading,” IEEE Trans. Antennas Propagat., vol. AP-16, pp. 21–31, Jan. 1968.

Fig. 9. Performance of reduced length (L = 1:96 ) double arm quasifractal LPZA (QFLPZA). (a) Resistance and reactance, (b) input impedance on Smith Chart with 575

normalization impedance, (c) return loss, and (d) vertically polarized radiation pattern at = 0 plane.

the gain and bandwidth can be achieved, and therefore is selected for the double arm QFLPZA. The results for its impedance, return loss, and radiation patterns are shown in Fig. 9(a)–(d), respectively.

1962

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[3] K. Iizuka, “The traveling wave V-antenna and related antennas,” IEEE Trans. Antennas Propagat., vol. AP-15, pp. 236–243, Mar. 1967. [4] W. Rotman and N. Karas, “Sandwich wire antenna: A new type of microwave line source radiator,” Proc. IRE Nat. Conf. Rec., pt. 1, pp. 166–172, 1957. [5] K. Chen, “Sandwich wire antenna,” IRE Trans. Antennas Propagat., vol. AP-10, pp. 159–164, Mar. 1962. [6] H. E. Green and J. L. Whitrow, “The new analysis of the sandwich wire antenna,” IEEE Trans. Antennas Propagat., vol. AP-19, pp. 600–605, Jan. 1971. [7] G. A. Hockham and R. I. Wolfson, “Broadband meander-line planar array antenna,” IEEE Trans. Antennas Propagat., vol. AP-27, pp. 645–648, 1979. [8] O. Aboul-Atta and L. Shafai, “Hemispherically radiating meander-line planar array antenna,” in Proc. Int. Conf. Antennas and Propagation, 1987, pp. 141–144. [9] Antenna Handbook, Theory, Applications, and Design, Y. T. Lo and S. W. Lee, Eds., Van Nostrand Reinhold, New York, 1988. [10] S. H. Lee and K. K. Mei, “Analysis of zigzag antennas,” IEEE Trans. Antennas Propagation, vol. AP-18, pp. 760–764, Nov. 1970. [11] A. F. Gangi, S. Sensiper, and G. R. Dunn, “The characteristics of electrically short, umbrella top-loaded antennas,” IEEE Trans. Antennas Propagat., vol. AP-13, pp. 864–871, Nov. 1965. [12] S. K. Sharma and L. Shafai, “Investigations of a compact vertically polarized backﬁre high frequency traveling wave antenna,” in Proc. IEEE Antennas Propagation Int. Symp., vol. 1, Columbus, OH, June 22–27, 2003, pp. 253–256. [13] J. P. Gianvittorio and Y. Rahmat-Samii, “Fractal antennas: A novel antenna miniaturization technique, and applications,” IEEE Antennas Propagat. Mag., vol. 44, pp. 20–36, Feb. 2002. [14] E. C. Jordan and K. G. Balmain, “Ch. 15,” in Electromagnetic Waves and Radiating Systems, 2nd ed. Englewood Cliffs, NJ: Prentice-Hall Inc., 1968, pp. 620–621. [15] V. H. Rumsey, Frequency Independent Antennas, New York: Academic, 1966, pp. 105–110. [16] Numerical Electromagnetic Code.

Satish Kumar Sharma (M’00) was born in Sultanpur (U.P.), India, in April 1970. He received the B. Tech. degree from Kamla Nehru Institute of Technology, Sultanpur afﬁliated with Avadh University, Faizabad, and the Ph.D. degree from Institute of Technology, Banaras Hindu University, Varanasi, both in India, in 1991 and 1997, respectively, both in electronics engineering. He was a Lecturer and Project Ofﬁcer with the Kamla Nehru Institute of Technology, Sultanpur, and Institute of Engineering and Rural Technology, Allahabad, from February 1992 to December 1993, respectively, both in Uttar Pradesh. There, he taught courses in electromagnetics, antennas and propagation, electronics instrumentation and electronic communication, etc. During December 1993 to February 1999, he was a Research Scholar, and Junior and Senior Research Fellow of the Council of Scientiﬁc and Industrial Research (CSIR), Government of India, in the Department of Electronics Engineering, Institute of Technology, Banaras Hindu University. He was a Postdoctoral Fellow with the Department of Electrical and Computer Engineering, The University of Manitoba, with Professor L. Shafai from March 1999 to April 2001. Since May 2001, he has been a Senior Antenna Researcher/Engineer with InfoMagnetics Technologies Corporation, Winnipeg, MB, Canada. Since June 2001, he has also been a part-time Research Associate with the Department of Electrical and Computer Engineering, The University of Manitoba. Here, he has been involved in the design and development of several antennas for wireless and satellite communications as feed for reﬂectors, polarizers, and MEMS phase shifters. His main research interests are in applied electromagnetics, antennas, and RF MEMS. He is a registered Professional Engineer (P. Eng.) of the Province of Manitoba, Canada. He is reviewer of IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION and Indian Journal of Radio and Space Physics (IJRSP).

Lotfollah Shafai (F’87) received the B.Sc. degree from the University of Tehran, Iran, in 1963 and the M.Sc. and Ph.D. degrees from the Faculty of Applied Sciences and Engineering, University of Toronto, ON, Canada, in 1966 and 1969, respectively, all in electrical engineering. In November 1969, he joined the Department of Electrical and Computer Engineering, University of Manitoba, Canada, as a Sessional Lecturer, then as an Assistant Professor in 1970, an Associate Professor in 1973, and as Professor in 1979. Since 1975, he has made a special effort to link the University research to the industrial development by assisting industries in the development of new products or establishing new technologies. To enhance the University of Manitoba’s contact with industry, in 1985 he assisted in establishing The Institute for Technology Development, and was its Director until 1987, when he became the Head of the Electrical Engineering Department. His assistance to industry was instrumental in establishing an Industrial Research Chair in Applied Electromagnetics at the University of Manitoba in 1989, which he held until July 1994. Dr. Shafai is a Member of the International Scientiﬁc Radio Union (URSI) Commission B, was its Chairman during 1985 to 1988, and is currently the Vice-Chairman. He was elected a Fellow of The Royal Society of Canada in 1998. He has been the recipient of numerous awards. In 1978, his contribution to the design of a small ground station for the Hermus satellite was selected as the 3rd Meritorious Industrial Design. In 1984, he received the Professional Engineers Merit Award and in 1985, “The Thinker” Award from Canadian Patents and Development Corporation. From the University of Manitoba, he received the “Research Awards” in 1983, 1987, and 1989, the Outreach Award in 1987 and the Sigma Xi, Senior Scientist Award in 1989. In 1990, he received the Maxwell Premium Award from IEE (London) and in 1993 and 1994, the Distinguished Achievement Awards from Corporate Higher Education Forum. In 1998, he received the Winnipeg RH Institute Foundation Medal for Excellence in Research. In 1999 and 2000, he received the University of Manitoba, Faculty Association Research Award. He was a recipient of the IEEE Third Millenium Medal in 2000 and in 2002 was elected a Fellow of The Canadian Academy of Engineering and Distinguished Professor at The University of Manitoba. He holds a Canada Research Chair in Applied Electromagnetics. He has been a participant in nearly all Antennas and Propagation symposia and participates in the review committees. In 1986, he established the symposium on Antenna Technology and Applied Electromagnetics, ANTEM, at the University of Manitoba, which is held every two years.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

1963

**Compact Wide-Band Multimode Antennas for MIMO and Diversity
**

Christian Waldschmidt, Student Member, IEEE, and Werner Wiesbeck, Fellow, IEEE

Abstract—This paper presents broadband multimode antennas for multiple-input multiple-output (MIMO) and diversity applications. The antenna system is not based on spatial diversity, as usual MIMO systems, but on a combination of pattern and polarization diversity. Different modes of self-complementary, thus extremely broadband, spiral and sinuous antennas are used to decorrelate the signals. It is shown that only one antenna is necessary to receive three uncorrelated signals, thus the space required to place the MIMO antenna is very small. Simulation results and measurements of a typical indoor scenario are given. Index Terms—Multimode diversity, multiple-input multipleoutput (MIMO), sinuous antenna, spiral antenna.

I. INTRODUCTION UTURE communication systems have to fulﬁll the requirements of high data rates and ﬂexible interfaces for different communication system standards. Multistandard radios, offering the demanded ﬂexibility to use different standards, require very broadband antennas. multiple-input multiple-output (MIMO) and diversity systems allow exploitation of the spatial channel properties. If the signals received by different antennas are uncorrelated, very high data rates may be reached as recent studies have shown, ﬁrst in [1] and later in [2], [3]. Usually uncorrelated signals are obtained by spatial diversity, which requires large antenna spacings. This paper presents new broadband antenna solutions, that are small enough to ﬁt into laptops or organizers, but that still yield uncorrelated signals for MIMO or diversity applications. The compactness of the broadband MIMO antenna system is not achieved by using different antennas, but by one antenna with different, independently fed, modes. This results in multimode diversity, a combination of pattern- and polarization diversity to obtain uncorrelated channel impulse responses for the MIMO or diversity system. As far as the authors are aware multimode diversity has ﬁrst been suggested in [4], where orthogonal azimuth patterns were used. In [5] a multimode patch antenna with different modes for diversity was presented. Multimode diversity for MIMO has been suggested in [6], but this paper presents a new and practical antenna concept, based on spiral and sinuous antennas. In [7] the ability of logarithmic spiral antennas to radiate in different polarizations is discussed and a possible application for diversity is mentioned, but not explicated. Besides uncorrelated signals at the antennas, which are obtained by orthogonal patterns the mean signal to noise ratio

Manuscript received February 4, 2003; revised August 25, 2003. The authors are with the Institut für Höchstfrequenztechnik und Elektronik (IHE), Universität Karlsruhe (TH), Karlsruhe D-76128 Germany (e-mail: Christian.Waldschmidt@ihe.uka.de). Digital Object Identiﬁer 10.1109/TAP.2004.832495

F

Fig. 1. Geometry of a spiral antenna with voltage sources between the single arms of the spiral.

(SNR) of all the signals has to be “similar,” see [8], to obtain a diversity gain or capable MIMO systems. “Similar” in this context means, e.g., less than 10 dB difference for two branch maximum ratio combining, [8]. In this paper it is shown, that the mean effective gain (MEG), which is linked to the SNR, of the single modes differs by only 1 to 2 dB, thus a high diversity gain is obtained. For MIMO the total received power or the mean SNR respectively is an important quality measure for an antenna array. By a comparison with a dipole array with large antenna spacings, which is generally considered as a capable array for MIMO, the ability of multimode antennas for MIMO is shown. This paper is organized as follows. In the ﬁrst section four-arm spiral and sinuous antennas and the different excitations for the modes are presented. Second, the correlation properties of signals received by different modes of the antenna and the mean effective gains are given as a function of the incident ﬁeld and its spatial distribution. In the last section MIMO capacity calculations and measurements with spiral antennas are given. II. SPIRAL AND SINUOUS ANTENNAS The self-complementary, archimedian, four-arm spiral antenna and sinuous antennas are well described in the literature, see, e.g., [9]–[11], thus only the properties crucial for multimode diversity are given here. The spiral antenna consists of four arms, that are rotated around the center of the antenna, see Fig. 1. The antenna can basically radiate three different modes depending on the excitation. For this application mode 1 and mode 2 are used. Mode 1 is characterized by a phase shift of 90 between adjacent sources at the single arms of the spiral, see Fig. 1. Mode 2 has a phase shift of 180 . Both modes are circularly polarized in the direction of the main radiation

0018-926X/04$20.00 © 2004 IEEE

1964

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Geometry of a four-arm sinuous antenna. describes the lengths of the teeth and is therefore a determining antenna parameter.

Fig. 4. Pattern of mode 2 of the spiral antenna with a radius of 10 cm at 2 GHz. The pattern hardly changes versus frequency for frequencies above 1.2 GHz.

Fig. 3. Pattern of mode 1 of the spiral antenna with a radius of 10 cm at 2 GHz separated into left (lhc) and right hand circular (rhc) polarization. If the spiral is fed at the outer end of the arms, the polarization is orthogonal to the one obtained by exciting at the center of the spiral.

Fig. 5. Phase of the pattern of mode 1 and mode 2, shown in Figs. 3 and 4. The phase of mode 1 changes 360 per circulation around the antenna, mode 2 changes 720 .

and elliptically polarized otherwise. Due to the self-complementarily the antennas are frequency-independent or, in other words, extremely broadband. Since the geometrical structure of the spiral antenna is ﬁnite, there exists a lower frequency bound. This bound is (1) is the speed of light, the outer radius of the spiral where the effective substrate permittivity. has to be deand termined by simulations of the spiral antenna or experimentally. is close to one for etched spirals, According to experience, also for a high of the substrate. For all simulations presented in this paper the antennas were simulated with FEKO [12], a standard software tool based on method of moments. Equation (1) is explicable by the current distribution within the active zones of the single modes [10]. The active zone is a circular area located around the center of the antenna. The energy is radiated from the antenna in the active zone. This zone is characterized by a certain ratio of its circumference to the wavelength. For

mode 1 the circumference is one wavelength, for mode 2 it is two wavelengths. Thus, the current distribution on the arms of the spiral in the active zone has two maxima for mode 1 and four for mode 2. Above this lower frequency bound all antenna properties are almost stable and change only slightly with freof mode 1 and 2 are given quency. The pattern in elevation in Fig. 3 and Fig. 4. The azimuth patterns are omnidirectional. The phase of the complex radiation pattern, which among other parameters determines the correlation among the receive signals, is shown in Fig. 5. The phase of mode 1 changes 360 and the one of mode 2 720 for each circulation around the antenna, which is explicable by the current distribution within the active zones. The modes can be excited in two ways: ﬁrst by feeding the spiral arms at the inner ends that is at the center of the spiral and second at the outer ends of the arms. Those modes are orthogonally polarized left-hand circular (lhc) and right-hand circular (rhc). The third mode of the spiral antenna (270 phase shift between adjacent arms at the excitation) has a pattern, whose amplitude is equal to mode 1, but the polarization changes from lhc to rhc. Thu,s mode 1 and mode 3 are orthogonally polarized. The unwrapped phase of the pattern of mode 3 changes 1080 per circulation around the antenna.

WALDSCHMIDT AND WIESBECK: COMPACT WIDE-BAND MULTIMODE ANTENNAS FOR MIMO AND DIVERSITY

1965

Fig. 6. Pattern of mode 2 of the sinuous antenna with a radius of 10 cm at 2 GHz. The pattern hardly changes versus frequency for frequencies above 1.7 GHz. Fig. 8. Gain of the rhc and lhc polarized ﬁeld of the sinuous antenna with = 50 and an outer radius of 0.1 m.

of the pattern of mode 3 is equivalent to mode 1, but both modes are orthogonally polarized. III. MULTIMODE-DIVERSITY MIMO transmission channels are characterized by the channel matrix , which contains the channel impulse responses or the channel coefﬁcient in the ﬂat fading case between the different sets of transmit and receive antenna ports. For broadband systems the spectrum can be divided into narrowband sections with ﬂat fading. The diversity gain or MIMO capacity depends on the correlation coefﬁcients among those channel coefﬁcients of , see [3], and the SNR. The correlation is inﬂuenced by the statistical properties of the wave propagation and the antenna properties, in this case the properties of the single modes. In the following the correlation coefﬁcient among two receive signals as a function of the incident ﬁeld is calculated. This is equivalent to the correlation for one transmit and two among the channel coefﬁcients of receive antennas in a MIMO system. The spatial wave propagation properties are describable by and the power azimuth and elevation proﬁle for both polarizations and . To allow for analytical calculations typical statistical functions to model the wave propagation are chosen. Measurements have shown, that the power azimuth is best modeled by a Laplacian function [13] spectrum a for both polarizations. For the power elevation proﬁle Gaussian function is assumed. The total power angle spectrum is given by the product of the Laplacian function for the azimuth and a Gaussian function for the elevation, normalized so that . With [14] (earlier shown in [8] in a similar way) it can easily be shown that the complex correlation coefﬁcient among two signals received by different antennas, in this case different modes, is given by (2)

Fig. 7. Pattern of mode 1 of the sinuous antenna with a radius of 10 cm at 2 GHz. is 50 . Mode 3 is orthogonally polarized, but has the same pattern.

The geometry of a four-arm sinuous antenna is given in Fig. 2 and described in detail in [11]. The antenna is self-complementary and used as a multimode antenna. The modes are excited the same way as for the spiral antenna. The lower frequency bounds of the modes are a function of different geometry parameters, thus are not as easy accessible as for the spiral. In general the lower frequency bounds are higher than the ones for the spiral antenna for a given outer radius of the antennas. They decrease with increasing (for see Fig. 2), since the antenna resembles in sections a spiral antenna for large . The patterns of mode 1 and 2 are given in Figs. 6 and 7. The shapes of the patterns change only slightly with frequency or , but the polarization changes. The pattern is alternately left and right hand elliptically polarized versus frequency, see Fig. 8. The axial ratio of the sinuous antenna depends on . For large the antenna acts in sections like a spiral, thus the axial ratio is almost 0 dB. For small the antenna is rather linearly polarized. Both modes may be excited at the center or at the outer ends of the arms, but in contrast to the spiral antenna, orthogonal polarizations are only , spiral-like behavior). The shape obtained for large (

1966

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 9. Correlation coefﬁcient among mode 1 (a) and mode 2 (c), excited at the center of the spiral, and mode 1 (b), excited at the outer edge of the spiral to generate orthogonal polarizations. The incident ﬁeld has an elevation spread of 5 and azimuth spread of 60 . The lower frequency bound of mode 2 is 1.2 GHz, thus the spiral does not work correctly for lower frequencies.

Fig. 10. Correlation coefﬁcient among the same modes as in Fig. 9, but the incident ﬁeld has an elevation spread of 5 and an azimuth spread of 20 . Due to slight changes in the pattern for different frequencies the correlation coefﬁcient changes. But it is over the whole frequency range low enough to obtain a diversity gain.

with the covariance

(3) where is constant and the variance

(4) is the ratio of the power in -polarization to the where is a funcpower in polarization at the receiver. Note, that tion of the polarimetric radiation pattern, thus disappears for orthogonally polarized antennas in this case lhc and rhc polarized modes. The power correlation coefﬁcient is obtained by , according to [15]. Basically it is possible to use spiral or sinuous antennas with any different modes and polarizations for multimode-diversity. In the following ﬁrst a spiral and second a sinuous antenna are used to calculate the correlation coefﬁcients among receive signals. The orientation of the both antenna planes is vertical. A spiral antenna with mode 1 and mode 2, excited at the center of the antenna, and a third mode (mode 1) with orthogonal polarization excited at the outer edge of the antenna is used. Figs. 9 between and 10 show the power correlation coefﬁcient different modes for a large azimuth angular spread of 60 and a small spread of 20 for a mean direction of 10 in azimuth and 0 in elevation of the incident waves. The third mode is orthogonally polarized to the other modes, thus the correlation is almost zero. The other modes are more strongly correlated as the pattern of mode 1 and 2 partly overlap. On the other hand,

Fig. 11. Correlation coefﬁcient among mode 1 (1), mode 2 (2) and mode 3 (3) of the sinuous antenna. The incident ﬁeld has an elevation spread of 5 and azimuth spread of 60 . The lower frequency bound of mode 2 is 1.7 GHz, thus the antenna does not work correctly for lower frequencies.

the different phases of the patterns of mode 1 and 2 (see Fig. 5) decorrelate the received signals, since the single plane waves from different directions superpose differently for each mode. The inﬂuence of the feed network on the pattern of the modes is neglected. is used with three modes, all The sinuous antenna excited at the center of the antenna. Figs. 11 and 12 show the correlation coefﬁcient versus frequency for the scenarios mentioned above. Mode 3 is orthogonally polarized to mode 1 and 2, thus the correlation is low. Mode 1 and 2 hardly overlap, thus different signals are received. In order to fulﬁll the requirement of an equal or “similar” SNR of the signals received by different modes to obtain a diversity gain the MEG may be used, see [16]. The MEG is deﬁned as the ratio of the mean received power of one antenna under test to the mean received power of a reference antenna, when both

WALDSCHMIDT AND WIESBECK: COMPACT WIDE-BAND MULTIMODE ANTENNAS FOR MIMO AND DIVERSITY

1967

A. Simulations of the Capacity The channel model used to calculate the capacity of MIMO systems consisting of one spiral antenna at the transmitter and one at the receiver is an extended version of the model described in [19]. This stochastic channel model is based on ray-tracing simulations and measurement campaigns in indoor scenarios. It is a three dimensional double-directional channel model, in other words provides the angle of departure and arrival of each path. The channel model takes only non line-of-sight (NLOS) connections into account. The power azimuth spectrum is modeled by multiple Laplacian functions, each modeling a cluster of scatterers. The elevation proﬁle is modeled by a sine function. The cross polarization coupling is 8 dB. The antennas used for the simulation are one spiral antenna at the transmitter and one at the receiver. Mode 1 and 2 are excited at the center of the spiral, and mode 1 with orthogonal polarization is excited at the outer edge of the spiral. Thus, the same modes as in Section III at both transmitter and receiver are used. The orientation of the spiral plane is again vertical. The result of the simulations with this channel model are channel matrices (obtained the same way as in [20]). Therefore, the capacity of a MIMO system with no channel state information at the transmitter can be calculated [2] SNR (6)

Fig. 12. Correlation coefﬁcient among the same modes as in Fig. 11, but the incident ﬁeld has an elevation spread of 5 and an azimuth spread of 20 . With decreasing angular spread the correlation increases. TABLE I MEG OF DIFFERENT MODES IN DECIBELS (ELEVATION ANGULAR SPREAD 5 )

antennas are used in the same channel with the same transmit antenna, see [17], [18]. For the assumptions on the wave propagation made above the MEG can be calculated analytically for an isotropic reference antenna.

conjugate complex transpose and where SNR denotes SNR is the number of transmit antennas, in this case the number of different modes. The channel matrices in (6) are normalized with (7)

(5) are the gain patterns for both polarizations. Table I where shows the MEGs for different antennas and modes for a crosspolarization coupling of 8 dB. The MEGs of mode 3 of both antennas are equal to the one of mode 1. Since the requirement of orthogonal patterns, i.e., uncorrelated signals, and similar MEGs are fulﬁlled, a diversity gain over a large bandwidth with both antenna types, used as multimode antennas, is obtained. IV. MIMO SYSTEMS BASED ON MULTIMODE-DIVERSITY In order to show the potential of multimode antennas in MIMO systems, simulations and measurements of the capacity of a MIMO system with one multimode spiral antenna on each side of the link, were performed. Additionally a comparison with dipole antennas, arranged in parallel, is drawn. For the simulations a sophisticated channel model is used. This model does not allow for analytical calculations like in Section III, but it allows to assess the MIMO performance in very realistic environments.

to obtain a constant mean gain of each channel matrix, see [21]. The SNR in (7) is the average SNR at the receiver. This normalization allows to show the inﬂuence on the capacity of the correlation properties and the distribution of the mean gains of the channel coefﬁcients. This distribution inﬂuences the capacity. The channel coefﬁcients between co-polarized modes have a larger mean gain than those between cross-polarized modes. Thus, the mean gains of the channel coefﬁcients are not equal. Equality is considered to be optimal, [2]. Fig. 14 shows the capacity distribution for a constant mean SNR at the receiver of 10 dB for 1000 channel realizations at 2 GHz. The 10% outage capacity is approximately 7.3 bit/s/Hz. B. Power Considerations When the normalization in (7) is used in other words when the gain of each channel matrix is normalized, the information about it is lost. But to assess arrays for MIMO completely, this information needs to be taken into account to assure a high efﬁciency of the complete MIMO channel. Fig. 15 shows the cumulative distribution function of the gain of the channel matrices of the simulations. The comparison with a MIMO system with arrays consisting of three vertical half-wavelength

1968

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 13. Scenarios for the measurements. For the LOS scenarios transmitter and receiver are placed in the same room. For the NLOS scenarios the transmitter is placed in the corridor.

Fig. 15. Transmission gain of different MIMO systems for the path based channel model. The SISO system has one transmit and one receive dipole.

Fig. 14. Measured cumulative distribution functions of the capacity for different antenna scenarios at 2 GHz. The three dipoles have spacings of =4. The capacity of the NLOS measurements reaches higher capacities as in the LOS scenarios for a constant mean SNR of 10 dB.

dipoles (also simulated with FEKO) with spacings on each side of the link shows, that the gain of the channel matrices of the multimode MIMO system is not worse than with the dipole arrays. Additionally this distribution function is given for dipole spacings on each side of the link. arrays with C. Measurements The measurements were performed with two spiral antennas. The antenna were designed for a frequency range from 1.2 GHz up to 2.5 GHz, limited by the feeding network. Mode 1 and 2 are excited with the feeding network given in [9]. At the outer ends of the arms a hybrid mode with orthogonal polarization compared to the other modes is excited. The coupling between the single modes is below 20 dB. The measurement system consists of a two channel network analyzer, ampliﬁers and coaxial switches. The channel coefﬁcients were measured one by one. All measurements were done during night, in order to reduce the time variance of the channel. The measurements were performed in an ofﬁce building, with concrete ceilings and concrete and wood covered walls. The average ofﬁce size is 4 5 m, see Fig. 13. The receive antennas were placed at the same position

for all measurements. The transmitter was moved along two different routes, shown in Fig. 13. During the ﬁrst route a strong LOS component is present, whereas the other route is always NLOS. Along each route measurements at 801 discrete frequencies in the frequency range from 1.5 to 2.5 GHz at 210 different positions were performed. The measured data are normalized, according to (7), to obtain a constant mean SNR of each channel matrix of 10 dB. Fig. 14 shows the cumulative capacity distribution for both routes at 2 GHz. The capacity distribution changes negligibly with frequency. Due to the higher multipath richness of the NLOS route, it outperforms the LOS route. For comparison two dipole arrays, consisting of three dipoles each, were used, one at each side of the link. The dipoles were and vertical polarizaarranged in parallel with spacings of tion. The array covers approximately the same area as the spiral antenna with dimensions, so that the resonance frequency of the dipoles equals the lower frequency bound of the spiral. Fig. 14 shows, that the dipoles perform worse than the spiral, since neither polarization nor pattern diversity is exploited. The space diversity is very limited due to the small antenna spacings. V. CONCLUSION This paper shows that four-arm spiral and sinuous antennas allow to exploit multimode diversity, which is a combination of pattern and polarization diversity. The antennas are extremely broadband, thus allow applications for multistandard radios. The space required for the antennas is relatively small. If placing dipoles on the same space required by the spiral, the dipoles do not reach the capacity of multimode-based MIMO-systems. REFERENCES

[1] J. H. Winters, “On the capacity of radio communication systems with diversity in a rayleigh fading environment,” IEEE J. Select. Areas Commun., vol. 5, pp. 871–877, May 1987. [2] G. J. Foschini and M. J. Gans, “On limits of wireless communications in a fading environment when using multiple antennas,” Wireless Personal Commun., vol. 6, pp. 311–335, 1998.

WALDSCHMIDT AND WIESBECK: COMPACT WIDE-BAND MULTIMODE ANTENNAS FOR MIMO AND DIVERSITY

1969

[3] C. Chuah, D. N. C. Tse, J. M. Kahn, and R. A. Valenzuela, “Capacity scaling in MIMO wireless systems under correlated fading,” IEEE Trans. Inform. Theory, vol. 48, pp. 637–650, 2002. [4] E. N. Gilbert, “Energy reception for mobile radio,” Bell Syst. Tech. J., vol. 44, pp. 1779–1803, 1965. [5] R. G. Vaughan and J. B. Andersen, “A multiport patch antenna for mobile communications,” in Proc. 14th Eur. Microwave Conf., 1984, pp. 607–612. [6] T. Svantesson, “An antenna solution for mimo channels: the multimode antenna,” in Conf. Record 34th Asilomar Conf., vol. 2, 2000, pp. 1617–1621. [7] O. K. Kim and J. D. Dyson, “A log-spiral antenna with selectable polarization,” IEEE Trans. Antennas Propagat., vol. AP-19, pp. 675–677, Apr. 1971. [8] R. G. Vaughan and J. B. Andersen, “Antenna diversity in mobile communications,” IEEE Trans. Veh. Technol., vol. 36, no. 4, pp. 149–172, July 1987. [9] E. Gschwendtner and W. Wiesbeck, “Multi-service dual-mode spiral antenna for conformal integration into vehicle roofs,” in Proc. IEEE Int. Symp. Antennas and Propagation AP-S, vol. 3, Davos, Switzerland, 2000, pp. 1532–1535. [10] R. G. Corzine and J. A. Mosko, Four-Arm Spiral Antennas. Norwood, MA: Artech House, 1990. [11] T. T. Chu and H. G. Oltman, “The sinuous antenna,” Microwave Syst., News and Commun. Technol., vol. 18, pp. 40–48, 1988. [12] www.emss.de [Online] [13] K. I. Pedersen, P. M. Mogensen, and B. H. Fleury, “Spatial channel characteristics in outdoor environments and their impact on BS antenna system performance,” in Proc. IEEE Vehicular Technology Conf., 1998, pp. 719–724. [14] K. Fujimoto and J. R. James, Mobile Antenna Systems Handbook. Norwood, MA: Artech House, 1994. [15] J. R. Pierce and S. Stein, “Multiple diversity with nonindependent fading,” in Proce. IRE, vol. 48, 1960, pp. 89–104. [16] M. G. Douglas, M. Okoniewski, and M. A. Stuchly, “Performance of pcs handset antennas in mobile environments,” in Proc. IEEE MTT-S Int. Microwave Symp. Dig., vol. 3, 1997, pp. 1759–1762. [17] J. B. Andersen and F. Hansen, “Antennas for VHF/UHF personal radio: a theoretical and experimental study of characteristics and performance,” IEEE Trans. Veh. Technol., vol. AP-26, pp. 349–357, 1977. [18] T. Taga, “Analysis for mean effective gain in mobile antennas in land mobile radio environments,” IEEE Trans. Veh. Technol., vol. 39, pp. 117–131, 1990. [19] T. Zwick, C. Fischer, and W. Wiesbeck, “A stochastic multipath channel model including path directions for indoor environments,” IEEE J. Select. Areas Commun., vol. 20, pp. 1178–1192, 2002. [20] C. Waldschmidt, T. Fügen, and W. Wiesbeck, “Spiral and dipole antennas for indoor MIMO-systems,” Antennas Wireless Propagat. Lett., vol. 1, no. 9, pp. 176–178, 2002. [21] J. W. Wallace and M. A. Jensen, “Characteristics of measured 4 4 and 10 10 MIMO wireless channel data at 2.4 GHz,” in Proc. IEEE Symp. Antennas and Propagation, vol. 3, 2001, pp. 96–99.

Christian Waldschmidt (S’01) was born in Basel, Switzerland, in 1976. He received the Dipl.-Ing. (M.S.E.E.) degree in electrical engineering from the Universität Karlsruhe, Karlsruhe, Germany, in 2000, where he is currently working toward the Ph.D. degree. From 2001 to 2003, he was with the Institut für Höchstfrequenztechnik und Elektronik (IHE), Universität Karlsruhe (TH), as a Research Assistant. He serves as a Lecturer for smart antennas and radar antenna systems for the Carl Cranz Series for scientiﬁc education. His research activities mainly focus on multiple input multiple output systems, smart antennas, small antennas, integration of antennas and vehicular antennas for radar and mobile communications applications.

2

2

Werner Wiesbeck (SM’87–F’94) received the Dipl.Ing. (M.S.E.E.) and the Dr.-Ing. (Ph.D.E.E.) degrees from the Technical University of Munich, Munich, Germany, in 1969 and 1972, respectively. From 1972 to 1983, he was with AEG-Telefunken in various positions including the Head of Research and Development, Microwave Division, Flensburg, Germany, and Marketing Director in the Receiver and Direction Finder Division, Ulm. During this period he had product responsibility for millimeter-wave radars, receivers, direction ﬁnders and electronic warfare systems. Since 1983, he has been Director of the Institut für Höchstfrequenztechnik und Elektronik (IHE), University of Karlsruhe, Karlsruhe, Germany, where he is presently Dean of the Faculty of Electrical Engineering. In 1989 and 1994, respectively, he spent a six month sabbatical at the Jet Propulsion Laboratory, Pasadena. He serves as a Permanent Lecturer for radar system engineering and for wave propagation For the Carl Cranz Series for Scientiﬁc Education. He is a Member of an Advisory Committee of the EU-Joint Research Centre (Ispra/Italy), and he is an advisor to the German Research Council (DFG), to the Federal German Ministry for Research and to industry in Germany. His research topics include radar, remote sensing, wave propagation and antennas. Dr. Wiesbeck has received a number of awards including the IEEE Millennium Medal. Since 2002, he has been a Member of the “Heidelberger Akademie der Wissenschaften.” He was a Member of the IEEE GRS-S AdCom from 1992–2000, Chairman of the GRS-S Awards Committee from 1994 to 1998, Executive Vice President IEEE GRS-S from 1998 to 1999, President IEEE GRS-S from 2000 to 2002, Associate Editor IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION from 1996 to 1999, past Treasurer of the IEEE German Section. He has been General Chairman of the 1988 Heinrich Hertz Centennial Symposium, the 1993 Conference on Microwaves and Optics (MIOP ’93) and he has been a member of scientiﬁc committees of many conferences.

1970

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Ground Inﬂuence on the Input Impedance of Transient Dipole and Bow-Tie Antennas
**

Andrian Andaya Lestari, Alexander G. Yarovoy, Member, IEEE, and Leo P. Ligthart, Fellow, IEEE

Abstract—In this paper, the inﬂuence of a lossy ground on the input impedance of dipole and bow-tie antennas excited by a short pulse is investigated. It is shown that the ground inﬂuence on the input impedance of transient dipole and bow-tie antennas is significant only for elevations smaller than 1 5 of the wavelength that corresponds to the central frequency of the exciting pulse. Furthermore, a principal difference between the input impedance due to traveling-wave and standing-wave current distributions is pointed out. Index Terms—Bow-tie antenna, impedance, transient antenna. dipole antenna, input

I. INTRODUCTION

D

IPOLE and bow-tie antennas are employed in many transient applications such as impulse ground penetrating radar (GPR) for transmitting short transient pulses. The large antenna bandwidth required to transmit such pulses with minimal distortion (e.g., antenna ringing) is usually obtained by the application of resistive loading [1], [2]. As resistive loading substantially reduces radiation efﬁciency [1], it is essential to achieve maximum power transfer from the generator to the antenna, for which the input impedance of the antenna should be known. The input impedances of time-harmonic and transient antennas are principally different since the former is due to standing-wave current distribution, while the latter is due to traveling-wave current distribution. Publications with regard to the input impedance of time-harmonic dipole and bow-tie antennas near the ground are abundantly available in the literature. On the contrary, not much of the input impedance of transient dipole and bow-tie antennas near the ground has been reported. In the free-space case, signiﬁcant contributions were given by Wu [3] and Carrel [4] who presented analytical expressions of the input impedance of transient dipole and bow-tie antennas, respectively. In this paper we analyze the input impedance of transient dipole and bow-tie antennas near a lossy ground. A numerical method to predict the input impedance of arbitrary metallic transient antennas in free space using the time-domain integral equation (TDIE) method has been demonstrated by Booker, et al. [5]. In their work the TDIE is

Manuscript received March 7, 2003; revised August 20, 2003. This work was supported by the Dutch Technology Foundation (STW) under the projects “Improved Ground Penetrating Radar Technology” (1999–2000) and “Advanced Re-Locatable Multisensor System for Buried Landmine Detection” (2001–2002). The authors are with the International Research Centre for Telecommunications-Transmission and Radar (IRCTR), Delft University of Technology, Delft, The Netherlands (e-mail: a.lestari@irctr.tudelft.nl). Digital Object Identiﬁer 10.1109/TAP.2004.832371

numerically solved by the method of moments (MoM) using the marching-on-in time approach. By neglecting end reﬂections in the time-domain solution, the input impedance is given by the high-frequency limit of the frequency-domain solution obtained by Fourier transforming the mentioned time-domain solution. Unfortunately, when it comes to layered-medium problems, well-suited Green’s functions in space-time domain are not yet well documented. One of few developments of such Green’s functions has recently been reported for analyzing the response of a transient dipole in stratiﬁed media [6]. However, layered-medium Green’s function formulations in space-time domain which are directly applicable to surface-patch MoM methodologies, are not yet widely reported. The numerical analysis carried out in this work is based on the frequency-domain integral equation (FDIE) method, as the methods of solution for problems with layered media in frequency domain are already well established. The FDIE incorporates a layered-medium Green’s function and is numerically solved by a surface-patch MoM scheme for metallic nonwire structures, whereas wire structures are approximated by narrow strips. The input impedance of the transient antennas is obtained by the Fourier transformation and a time-window technique for excluding end reﬂections. An experimental analysis is performed to verify the computed results. II. NUMERICAL METHOD The work reported in this paper is based on a frequencydomain mixed-potential integral equation (MPIE) formulation. To account for the presence of the ground the dyadic Green’s function formulation C for layered-medium problems by Michalski and Zheng [7] is incorporated into the MPIE, which is numerically solved by MoM according to the triangular surface-patch methodology introduced by Rao, et al. [8]. Computation time is minimized by employing the efﬁcient numerical implementation introduced in [9]. In this work the antennas are excited by a monocycle with 0.8-ns duration shown in Fig. 1(a). In Fig. 1(b) its normalized spectrum is given. The central frequency of this pulse is about dB levels are found at frequencies 424 MHz 1 GHz and the and 1.670 GHz. It can be seen in Fig. 1(b) that the spectrum of the exciting pulse is essentially contained within the 0–5 GHz range. In this paper, we develop a nonstraightforward numerical method to predict the input impedance of a transient antenna in four steps as follows. is computed in the frequency 1) Antenna feed current domain with 1 Volt input voltage by means of the MoM

0018-926X/04$20.00 © 2004 IEEE

LESTARI et al.: GROUND INFLUENCE ON THE INPUT IMPEDANCE OF TRANSIENT DIPOLE AND BOW-TIE ANTENNAS

1971

Fig. 2.

Setup for input impedance measurements.

in which is the value of the real or imaginary part of the argument, and is the decay rate of the truncation process . We have found that assumes values in the range that typically the performance of (2) is optimal with . Note that when (2) reduces to a rectangular time window. 4) Finally, the input impedance of the transient antenna is obtained as (3)

III. MEASUREMENT TECHNIQUE We perform input impedance measurements in frequency domain using a standard network analyzer. The antennas are situated horizontally over a lossy ground, which in this case is dry sand. The measurements are carried out without an anechoic chamber, and consequently the results are heavily disturbed by reﬂections from surrounding objects. To deal with this, the measurement results are inverse-Fourier transformed into time domain, after which use of time gating is made to remove those unwanted reﬂections. The actual impedance of the antenna can be extracted afterwards by performing the Fourier transformation of the results back to frequency domain. Furthermore, to properly measure balanced loads such as the antennas discussed here, a balun is required. However, since ultra-wideband baluns are difﬁcult to produce and commercially available ones are expensive, baluns are not used in the measurements. A technique to measure the input impedance of balanced antennas without baluns introduced in [10] is here simpliﬁed. The antenna under test is fed using two identical 50-Ohm semi-rigid cables soldered together over their length, except for a small part near the ends where SMA connectors are attached. Each of the inner conductors of the other ends is soldered to one of the antenna terminals in the way shown in Fig. 2. The semi-rigid cables are connected by 50-Ohm Sucoﬂex coaxial cables to the ports of the network analyzer, which has been previously calibrated at the SMA connectors. Hence, the reﬂection coefﬁcient at the SMA connectors is given by (4) where and are the -parameters measured by the network analyzer. Disturbances caused by unwanted reﬂections can

Fig. 1. Exciting pulse used in this work: (a) a monocycle with 0.8-ns duration and (b) its normalized spectrum.

scheme mentioned above. For obtaining time-domain solutions, the computations are performed at 100 frequency points from 50 MHz to 5 GHz with 50-MHz steps. is synthesized directly from 2) Exciting pulse function measurement of the 0.8-ns monocycle in Fig. 1(a), and used as the excitation model in the computations. 3) The feed current of the same antenna with inﬁnite length (thus, no end reﬂections) is computed by means of the Fourier transformation and a time-window technique for removal of end reﬂections, which can concisely be written as F W F (1)

where is the normalized F is the discrete Fourier transformation operator, F is the discrete inverse Fourier transformation operator, and W is the time-window operator with smooth truncation process given by for the left end (2a) for the right end (2b)

1972

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

be removed from (4) by means of a time-gating operation, which can be described by F W F (5)

in which the time-window operator W implements the truncaand the tion process in (2). Using the dielectric constant insertion loss of the semi-rigid cables provided by the manufacturer, the reﬂection coefﬁcient at the antenna terminal can then be written as (6) is two where is the phase constant which depends on times the length of the semi-rigid cables and is given in dB/m. Several available values of for different frequencies are interpolated to obtain the values of over the whole frequency range. Finally, the input impedance of the antenna is given by (7) where is the characteristic impedance of the semi-rigid cables, which has a value of 100 Ohms due to the double-line conﬁguration shown in Fig. 2. It is worth noting that antenna length is actually not important for the input impedance of transient antennas. However, in the above technique antenna length is important for computation and measurement procedures, i.e., — to ensure separation in time between feed-point and end reﬂections for time window application, — to obtain adequate duration of the time window since a ﬁnite time window limits the lowest frequency at which input impedance can be determined. In this work we use antennas with 50-cm length, which allows application of a time window with 2-ns duration for excitation with the 0.8-ns monocycle. IV. ANALYSIS As the input impedance of a transient dipole is frequency dependent [3], the input impedance is determined with respect to the exciting pulse using the averaging given by [11]

Fig. 3. Experimental dipole over dry sand. Length 2 mm.

=

= 50 cm and wire diameter

(8)

and are the frequencies which correspond to the in which lower and upper limits of the pulse spectrum, respectively, and is the antenna input impedance. It has been indicated in over the range [11] that it is adequate to assume , and this is followed here with and cordB limits of the exciting pulse in Fig. 1. responding to the Hence, in (8) can be interpreted as the input impedance with respect to the exciting pulse (the 0.8-ns monocycle). In this in (8) is replaced with in (3). In this way, work (8) is improved because prior to using the averaging in (8) we

perform time windowing in (1), which greatly reduces the oscillation of the impedance curve. In effect, this approach improves the averaging process. The computed and measured input impedances of the horizontal transient dipole in Fig. 3 with respect to the 0.8-ns monocycle as function of elevation above the sand are presented in Fig. 4(a). The wire is modeled as a thin strip using the equivalent radius formula [12]. The result computed by the Numerical Electromagnetics Code (NEC-2) is included for comparison. It can be seen that in general the computations agree with the measurement. At the highest elevation the computed and measured values of the reactance are about 200 , in accordance with the result obtained using the expression given in [3]. For very small distances from the interface we observe that the result computed by the triangular-patch MoM suffers from a slight discrepancy with the measurement. We notice that this discrepancy might be caused by the variation of the electrical length of the feed gap of the experimental antenna when it approaches the ground. Such a phenomenon is however not experienced by the delta-function generator assumed in the triangular-patch MoM. The delta-gap model used by NEC, on the other hand, accommodates this phenomenon as it uses a feed segment with the same length as the feed gap width of the experimental antenna. This leads to better agreement with the measurement for small distances from the interface as shown in the ﬁgure. The result computed using the commercial MoM code FEKO, which is based on the same triangular-patch MoM methodology, is also included. It can be seen that generally agreement between our code and FEKO is achieved. To further test the accuracy of our results, computations are carried out using our code and FEKO for a 1-mm wire diameter and the results are presented in Fig. 4(b), in which the results for a 2-mm diameter in Fig. 4(a) are also shown. It is demonstrated in Fig. 4(b) that agreement is generally achieved. The observed slight discrepancy in reactance may be attributed to different densities of the mesh generated by the codes, especially in the feed region of the antenna. Moreover, we note that agreement between the results for small elevations indicates the accuracy of numerical evaluation of the layered-medium Green’s function.

LESTARI et al.: GROUND INFLUENCE ON THE INPUT IMPEDANCE OF TRANSIENT DIPOLE AND BOW-TIE ANTENNAS

1973

Fig. 5. Input impedance of transient dipoles in free space for two different wire diameters: analytical against numerical results. Analytical results are obtained using [3]; numerical results are computed by our code.

Fig. 4. (a) Input impedance of the transient dipole in Fig. 3 with respect to the 0.8-ns monocycle as function of elevation above the sand. (b) Comparison between this work and FEKO for two different wire diameters (1 and 2 mm).

Additional proof of the accuracy of the numerical results is given in Fig. 5 by comparison with analytical results, obtained using the theory in [3]. The input impedance of a transient dipole in free space is plotted for two different wire diameters, 1 and 2 mm. It can be seen that for sufﬁciently high frequencies generally agreement between numerical and analytical results is achieved. The observed slight discrepancy at high frequencies may be explained by difference in the excitation models of the antenna. The theory in [3] assumes excitation from a coaxial feed system, while for excitation we use a delta-function generator. The large discrepancy at low frequencies is due to the ﬁnite length of the time window, which imposes limitation on the lowest frequency at which accuracy of the result is ensured. To obtain improved accuracy at lower frequencies one should use a longer antenna for allowing a longer time window. It is advisable to mention the advantages and disadvantages of the used codes. The advantage of NEC-2 is the efﬁciency for handling wire structures as it employs thin-wire approximation, which reduces formulation of the problem into a

one-dimensional integral equation. However, to this work the main drawback of NEC-2 is that it renders inaccurate when modeling antennas very close to the ground. In Fig. 4(a) NEC-2 computation is interrupted at 5-mm elevation because for lower elevations the results become inaccurate. One of the advantages of our triangular-patch MoM code is its capability of modeling antennas touching the interface. In addition, it offers ﬂexibility for modeling metallic antennas of arbitrary shape. In comparison with NEC-2 the obvious disadvantage of the code when treating wire structures is its lower computation efﬁciency since it employs a surface integral formulation. To analyze the inﬂuence of a lossy ground on the input impedance of the transient dipole we compute the input impedance with respect to the 0.8-ns monocycle as functions of antenna elevation for different ground types. In particular, we assume the ground to be sandy soil S/m), dry clay ( S/m), ( S/m), and muddy soil wet clay ( S/m). Furthermore, two different wire ( diameters of 1 and 2 mm are assumed to investigate the inﬂuence of the wire thickness on the results. The computed input resistance and reactance with respect to the 0.8-ns monocycle are plotted in Fig. 6, where it is evident that the presence of the ground signiﬁcantly affects the impedance only for very small distances from the interface. In Fig. 6 it is shown that, the ground inﬂuence is already very small at elevations higher than 6 cm for a wide range of ground types. Noting that the central frequency of the 0.8-ns monocycle is about 1 GHz (corresponding to a wavelength of 30 cm), as a generalization of the results one may state that the ground essentially affects the input impedance of a transient dipole only when the antenna of the wavelength that corresponds elevation is smaller than to the central frequency of the exciting pulse. Evidently, wire diameter exhibits a considerable inﬂuence on the resistance as demonstrated in Fig. 6, which indicates that doubling the wire diameter from 1 to 2 mm reduces the resistance by about 19%. It is also worth noting that for a very

1974

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 7. Computed input impedance of a transient bow tie as functions of elevation above sandy soil for different ﬂare angles.

of the wavelength that corresponds to the is smaller than central frequency of the exciting pulse. V. CONCLUSION The inﬂuence of a lossy ground on the input impedance of dipole and bow-tie antennas excited by a short pulse is investigated. It is shown that the ground inﬂuence on the input impedance of a transient dipole and bow-tie antennas is signifof the wavelength icant only for elevations smaller than that corresponds to the central frequency of the exciting pulse. Furthermore, it is shown that for a very close proximity to the interface, the input resistance of a transient dipole decreases as the dipole approaches the interface. This behavior is the opposite of the time-harmonic case, in which for very small distances from the interface the input resistance increases as the dipole is lowered. ACKNOWLEDGMENT The authors thank P. Hakkaart for his assistance in the construction of the experimental antenna and J. Zijderveld for his assistance in the measurements. REFERENCES

[1] T. P. Montoya and G. S. Smith, “A study of pulse radiation from several broad-band loaded monopoles,” IEEE Trans. Antennas Propagat., vol. 44, pp. 1172–1182, Aug. 1996. [2] K. L. Shlager, G. S. Smith, and J. G. Maloney, “Optimization of bow-tie antennas for pulse radiation,” IEEE Trans. Antennas Propagat., vol. 42, pp. 975–982, July 1994. [3] T. T. Wu, “Input admittance of inﬁnitely long dipole antennas driven from coaxial lines,” J. Math. Phys., vol. 3, pp. 1298–1301, 1962. [4] R. L. Carrel, “The characteristic impedance of two inﬁnite cones of arbitrary cross section,” IRE Trans. Antennas Propagat., vol. AP-6, pp. 197–201, Apr. 1958. [5] S. M. Booker, A. P. Lambert, and P. D. Smith, “A numerical calculation of transient antenna impedance,” in Proc. 2nd Int. Conf. Computation in Electromagnetics, 1994, pp. 359–362. [6] A. G. Tijhuis and A. Rubio Bretones, “Transient excitation of a layered dielectric medium by a pulsed electric dipole,” IEEE Trans. Antennas Propagat., vol. 48, pp. 1673–1684, Oct. 2000.

Fig. 6. Computed input impedance of a transient dipole with respect to the 0.8-ns monocycle as functions of elevation for different ground types.

close proximity to the interface, the input resistance decreases as the dipole approaches the interface. We note that this behavior is the opposite of the time-harmonic case, in which for very small distances from the interface the input resistance increases as the dipole is lowered [13], [14]. This result indicates a principal difference between traveling-wave and standing-wave current distributions of transient and time-harmonic antennas, respectively. In this paper the input impedance of a transient bow-tie antenna is computed as functions of elevation above the ground for ﬂare angles of 30 , 50 , and 70 . The computed input and impedance of a transient bow tie above sandy soil ( S/m) with respect to the 0.8-ns monocycle is presented in Fig. 7. We notice in the ﬁgure that for the three ﬂare angles the value of the reactance at a 6-cm elevation is already close to zero, which is the free-space value of the characteristic reactance of a transient bow tie [4]. Moreover, by inspection of the result given in [4] it is found that that at a 6-cm elevation the resistance nearly assumes its free-space value. Hence, similar to the case of the transient dipole discussed previously it is shown that the ground essentially affects the input impedance of a transient bow-tie antenna only when the antenna elevation

LESTARI et al.: GROUND INFLUENCE ON THE INPUT IMPEDANCE OF TRANSIENT DIPOLE AND BOW-TIE ANTENNAS

1975

[7] K. A. Michalski and D. Zheng, “Electromagnetic scattering and radiation by surfaces of arbitrary shape in layered media, part I: Theory, part II: Implementation and results for contiguous half-spaces,” IEEE Trans. Antennas Propagat., vol. 38, pp. 335–352, Mar. 1990. [8] S. M. Rao, D. R. Wilton, and A. W. Glisson, “Electromagnetic scattering by surfaces of arbitrary shape,” IEEE Trans. Antennas Propagat., vol. AP-30, pp. 409–418, May 1982. [9] A. A. Lestari, A. G. Yarovoy, and L. P. Ligthart, “Numerical and experimental analysis of circular-end wire bow-tie antennas over a lossy ground,” IEEE Trans. Antennas Propagat., vol. 52, pp. 26–35, Jan. 2004. [10] K. D. Palmer and M. W. van Rooyen, “Simple broadband measurement of balanced loads using a network analyzer,” in CD-ROM Proc. Millenium Conf. Antennas Propagat. (AP-2000), Davos, Switzerland, Apr. 2000. [11] R. W. P. King and H. J. Schmitt, “The transient response of linear antennas and loops,” IRE Trans. Antennas Propagat., vol. 10, pp. 222–228, May 1962. [12] C. M. Butler, “The equivalent radius of a narrow conducting strip,” IEEE Trans. Antennas Propagat., vol. AP-30, pp. 755–758, July 1982. [13] G. Turner, “The Inﬂuence of subsurface properties on ground penetrating radar pulses,” Ph.D. dissertation, Macquarie University, Sydney, NSW, Australia, 1993. [14] C. J. Leat, N. V. Shuley, and G. F. Stickley, “Complex image model for ground-penetrating radar antennas,” IEEE Trans. Antennas Propagat., vol. 46, pp. 1483–1488, Oct. 1998.

Alexander G. Yarovoy (M’96) received the Diploma (with honors) in radiophysics and electronics and the Cand. Phys. & Math. Sci. and Dr. Phys. & Math. Sci. degrees in radiophysics, from Kharkov State University, Kharkov, Ukraine, in 1984, 1987, and 1994, respectively. In 1987, he joined the Department of Radiophysics, Kharkov State University, as a Researcher and became a Professor in 1997. From September 1994 through 1996, he was with the Technical University of Ilmenau, Germany, as a Visiting Researcher. Since 1999, he has been with the International Research Centre for Telecommunications-Transmission and Radar (IRCTR), Delft University of Technology, Delft, The Netherlands, where he coordinates all GPR-related projects. His main research interests are in ultrawide-band electromagnetics, wave scattering from statistically rough surfaces and penetrable obstacles and computational methods in electromagnetics.

Andrian Andaya Lestari was born in Bogor, Indonesia. He received the Ingenieur and Ph.D. degrees in electrical engineering from Delft University of Technology, Delft, The Netherlands, in 1993 and 2003, respectively. From 1993 to 1998, he was with a government research agency in Jakarta, Indonesia. He joined the International Research Centre for Telecommunications-transmission and Radar (IRCTR), Delft University of Technology, as a Researcher in 1998. His work at IRCTR has resulted in over 20 publications, which include national and international patents, journal and conference papers, and scientiﬁc reports. Currently he works on ultrawide-band antennas and numerical tools for transient antenna analysis.

Leo P. Ligthart (M’94–SM’95–F’02) was born in Rotterdam, the Netherlands, on September 15, 1946. He received the Engineer’s degree (cum laude) and the Doctor of Technology degree from Delft University of Technology, Delft, The Netherlands, in 1969 and 1985, respectively, the Doctorates (honoris causa) from Moscow State Technical University of Civil Aviation, Moscow, Russia, in 1999, and the Doctorates (honoris causa) from Tomsk State University of Control Systems and Radioelectronics, Tomsk, Russia, in 2001. Since 1992, he has held the Chair of Microwave Transmission, Radar and Remote Sensing in the Department of Information Technology and Systems, Delft University of Technology, where in 1994, he became Director of the International Research Centre for Telecommunications-Transmission and Radar. His principal areas of specialization include antennas and propagation, radar and remote sensing, but he has also been active in satellite, mobile, and radio communications.

1976

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Adaptive Crossed Dipole Antennas Using a Genetic Algorithm
**

Randy L. Haupt, Fellow, IEEE

Abstract—Antenna misalignment in a mobile wireless communications system results in a signal loss due to a decrease in antenna directivity and a polarization mismatch. A genetic algorithm (GA) is used to adaptively alter the polarization and directivity of a crossed dipole receive antenna in order to increase the link budget. The three orthogonal dipole conﬁguration works better than only two crossed dipoles, but both improved the link loss as the angular pointing errors increased. A GA with a high mutation rate works best for a noiseless open loop adaptation, while a GA with a low mutation rate works best for noisy fully adaptive system. Index Terms—Adaptive antenna, crossed dipole, genetic algorithm (GA), polarization, smart antenna.

I. INTRODUCTION

I

F ONE OR MORE of the antennas in a wireless communication system is mobile, then as the antennas move, the direction of the peak gains and the polarization of the antennas change. As a result, the power received goes down. For instance, a wireless system that transmits vertical polarization has some of its power converted to horizontal polarization as the signal reﬂects from the environment. Unless the receive antenna can detect both polarizations, the received power decreases. Another example is when a spacecraft orbits the earth; the antennas in the communications system no longer align for optimum power transfer. Antenna engineers design the antennas for maximum directivity and polarization match when the antennas point at each other. Both the directivity and polarization of an antenna change with angle. If the antennas do not point at each other, then the product of the receive and transmit antenna directivities goes down. The directivity loss coupled with the polarization mismatch reduces the received power. An obvious solution to this problem is to keep the antennas pointing at each other. Constantly maneuvering a spacecraft requires an unacceptable expenditure of precious fuel, though. Steering the ground antenna is another option but only solves half of the problem, since the spacecraft might still be out of alignment. If the antenna is a phased array, then steering the main beam maximizes the directivity but does not improve the polarization mismatch. One way to improve the link budget is to adaptively change the antenna polarization and directivity to maximize the power transfer. In order to improve the link budget, the polarization

Manuscript received February 10, 2003; revised May 27, 2003. The author was with the Utah State University, Electrical and Computer Engineering, Logan, UT 84322-4120 USA. He is now with the Applied Research Laboratory, Pennsylvania State University, State College, PA 16804 USA (e-mail: haupt@ieee.org). Digital Object Identiﬁer 10.1109/TAP.2004.832493

and directivity of the receive and/or transmit antennas must be adaptively changed if the positions of the two antennas change. Adaptive antennas usually place a null in the antenna pattern to reject interference or steer a beam toward a desired signal. Phased arrays have more than one antenna, so they are perfect for adapting their patterns. Adapting the polarization, however, requires an antenna that can modify the major and minor axes of its polarization ellipse. Crossed dipoles are perfect for this application. The crossed dipole antenna has found use before in systems requiring antennas that change their polarization. A wireless communications system in a high multipath environment can improve the link better through polarization diversity in the form of crossed dipoles than through spatial diversity (antenna separation) [1]. A crossed dipole consists of two or three orthogonal dipoles. When linearly polarized signals become depolarized due to reﬂections, each dipole can receive the electric ﬁeld component parallel to it. Using three orthogonal crossed dipoles has been experimentally shown to signiﬁcantly increase channel capacity of a wireless communication system inside a building [2]. The polarization and directivity of a crossed dipole antenna are easy to control. One dipole controls the electric ﬁeld parallel to it and the orthogonal dipoles control the electric ﬁelds parallel to them. Each dipole has an independent complex weight. Controlling the amplitude and phase of the signal at each dipole modiﬁes the electric ﬁeld amplitude and phase in orthogonal directions resulting in any polarization from linear through elliptical to circular. Modifying the amplitude and phase of the signal at the dipoles also modiﬁes the directivity of the antenna as well. Adaptive crossed dipoles alter their polarization based upon environmental conditions. When a transmitted circularly polarized millimeter wave passes through rain, it becomes elliptically polarized. The depolarization can be calculated if the rainfall rate is known. Reference [3] proposed an open loop adaptive transmit antenna that adjusted its polarization based upon the measured rainfall in the propagation path. In [4], a least mean square (LMS) algorithm adapted the polarization and pattern of a two element array of crossed dipoles to improve the signal to interference plus noise ratio (SINR). As long as the desired and interference signals are not at the same angles and have the same polarizations, the SINR was improved. In another paper, the LMS algorithm was used to ﬁnd amplitude and phase weights for three orthogonal dipole antennas in order to improve the SINR. This arrangement provided some rejection for interference signals for most angles of arrival and polarizations [5]. A previous paper presented results from adaptively adjusting the

0018-926X/04$20.00 © 2004 IEEE

HAUPT: ADAPTIVE CROSSED DIPOLE ANTENNAS USING A GA

1977

the crossed dipole current is the sum of the constant currents on each short dipole (1) Substituting this current into the equation for the magnetic vector potential for a short dipole yields (2) where distance from the origin to the ﬁeld point at ; dipole length in the and directions; radial frequency; wave number; permeability; constant current in or direction. In the far ﬁeld, the electric ﬁeld in rectangular coordinates is found from the magnetic vector potential by

Fig. 1. Coordinate system for the crossed dipole transmit and receive antennas.

(3) amplitude and phase of the current fed to one dipole while the other dipole had an amplitude of one and phase of zero. The Numerical Electromagnetic Code generated the electromagnetic response of the dipole antennas and a local optimizer performed the optimization [6]. It was found that optimizing only for circular polarization produces losses in radiated power that offset the polarization correction. An improvement in the power trans. Using ferred increased up to a maximum of 2.0 dB at adaptive crossed dipoles at the transmitter and receiver was also considered and further improved the model. This paper expands upon a recent presentation that introduced the application of a genetic algorithm (GA) to adaptively change the current fed to crossed dipole antennas in order to improve the link budget [7]. The dipole model consists of three orthogonal short dipoles with variable control of the phase and amplitude fed to each element. A GA is used to maximize the received signal by improving the directivity and polarization match through weighting the currents at each dipole. Improvements in the link budget of up to 6 dB are possible. II. CROSSED DIPOLE MODEL Satellite communications systems use circularly polarized antennas for the satellite and the ground antennas. In this paper, the circularly polarized antennas are modeled as crossed dipoles. Consequently, controlling the amplitude and phase of the signals at the dipoles of the transmit and receive antennas, modiﬁes the directivity and polarization of both antennas. In this case, the crossed dipole has three orthogonal dipoles. from The receive antenna is located at an angle of the transmit antenna (Fig. 1). Similarly, the transmit antenna from the receive antenna. is located at an angle of and . Maximum power transfer occurs when In order to determine the directivity and polarization of the antennas, the electric ﬁelds can be found from the currents on the dipoles. If the dipoles are assumed to be short , The transmitted electric ﬁeld is given by

Converting this rectangular form of the electric ﬁeld into spherical coordinates produces the far ﬁeld components

(4) (5) The directivity is given by

(6) and the polarization loss factor is

(7)

with a perfect match. The where and subscripts represent transmit and receive, respectively. Equations (10) and (11) are key ingredients to the link budget. The examples presented here assume the earth station conplane sists of a pair of orthogonal crossed dipoles in the transmitting a circularly polarized ﬁeld in the -direction. Increasing toward the horizon transitions from circular polarization through elliptical until linear polarization results at the horizon. In this paper, the transmit antenna has the following

1978

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Circularly polarized crossed dipoles along the x and y axes. This is a hemispherical plot of the inverse axial ratio. Light color indicates high with a value of 1 in the z -direction and a value of 0 in the x y plane.

0

Fig. 4. Circularly polarized crossed dipoles along the x and y axes. This is a hemispherical plot of the directivity. Light color indicates high with a value in the y -direction and a low value in the x y plane.

0

Fig. 3.

Plot of the inverse axial ratio versus the elevation angle.

Fig. 5. Plot of the directivity versus the elevation angle.

currents: , , and . Fig. 2 shows a hemispherical plot and Fig. 3 a linear plot of the inverse axial ratio as a function of . White (at the poles) represents an inverse axial ratio of one (circular polarization), while black (at the equator) is an inverse axial ratio of zero (linear polarization). Increasing also changes the antenna directivity as shown in Figs. 4 and 5 from a maximum at 0 (white) to a minimum at 90 (black). Compensating for the loss in directivity and polarization match can improve the link budget by several decibels. III. GENETIC ALGORITHM OPTIMIZATION In a wireless communications system, the goal is to maximize the power transfer between the transmit and receive antennas. The following ﬁtness function calculates the portion of the link budget relating to polarization and directivity from the amplitude and phase of the currents for each dipole: (8) where is the directivity of the transmit crossed dipoles in the is the directivity direction of the receive crossed dipoles, and of the receive crossed dipoles in the direction of the transmit

crossed dipoles. Since the crossed dipoles have a maximum directivity close to one, the directivities are not normalized in the objective function. The maximum possible value of this ﬁtness function is approximately 2.25 or 3.5 dB. There are several approaches to performing the adaptation. With current technology, the most likely approach is to do an open loop adaptation. The satellite uses various sensors to make it aware of its orientation. Once its orientation is known, then the dipole currents can be found from a lookup table of optimized values, or the optimization could be done at that time. Noise and orientation errors limit the improvement possible. The other approach is a fully adaptive system capable of correcting for noise and system inaccuracies. This approach would be necessary for using the adaptive dipoles on another mobile system, such as an airplane, where the orientation and environment of the dipoles quickly changes and cannot be predicted ahead of time. A continuous parameter GA was used to ﬁnd the values of the amplitude and phase of the receive dipole currents that maximize (12). The GA has a population size of 8, mutation rate of .2, single point crossover, and 50% replacement. This small population size and high mutation rate results in a very fast convergence as will be shown in the following section. The goal of

HAUPT: ADAPTIVE CROSSED DIPOLE ANTENNAS USING A GA

1979

Fig. 6. Average number of function calls needed to get the ﬁtness above 3 for various population sizes and mutation rates.

Fig. 8. In this case, varies with time and = 0. The receive antenna consists of two crossed dipoles. The solid line results from adaptation and the dashed line has no adaptation.

Fig. 7. Link is optimal when the two antennas face each other or = = 0.

the optimization process is to quickly improve the communications link, not necessarily ﬁnd the global minimum. Fig. 6 shows the results of optimizing (8) using a GA for population sizes between 8 and 32 and mutation rate between 0.1 and 0.2. No noise was used in these runs. The plot is of the mean number of function calls to get (8) above 3 dB averaged over 50 independent runs when the GA begins with a random population. A small population size and large mutation rate produce the fastest convergence on average for the open loop adaptation. IV. RESULTS In all the examples presented here, the orientation of the ground and satellite antennas are assumed to change with time unless otherwise speciﬁed Fig. 7. Even though the distances between the antennas would also change, this variation is ignored. As the orientations of the antennas vary with time, so do their directivity and polarizations in the directions of each other. Assume that the transmit antenna (ground station) tracks the and the receive antenna (satellite) points at satellite the ground ( varies). The transmit antenna continues to deliver a circularly polarized signal at maximum directivity to the moving receive antenna. If the receive antenna consists of two crossed dipoles, then the maximum receive power transfer occurs when the receive antenna is directly overhead of the transmit antenna. If the receive antenna remains circularly polarized as it moves, then the power received drops off at the rate shown by the dashed line in Fig. 8. The loss in power transfer

Fig. 9. In this case, varies with time and = 0. The receive antenna consists of three crossed dipoles. The solid line results from adaptation and the dashed line has no adaptation.

is due to the reduction in the directivity and the PLF. If the currents at each dipole are optimally weighted using the GA, then the power loss follows the solid line in Fig. 8. This curve results from running the GA to ﬁnd the optimum weights for a range of angles. The difference between the two curves is the link improvement. The link improvement is as much as 3 dB at . In this case, all the improvement is due to increasing the directivity of the receive antenna. Adding a third orthogonal dipole to the receive antenna provides another degree of freedom. Now, adapting the receive antenna to the tracking transmit antenna results in no change in the (solid line in Fig. 9). The three link budget as a function of orthogonal dipoles can compensate for the change in directivity and polarization of the receive antenna as it moves. This scenario produces up to 6 dB improvement in the link budget at . Another scenario has both two dipole antennas pointing straight ahead (no tracking) while the satellite moves ( and change with time). Fig. 10 shows the improvement (solid line) possible through adaptation compared to the link loss with no adaptation (dashed line). The link improvement is as much . Adding a third dipole produces even better as 3 dB at

1980

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 10. In this case, no adaptation.

=

vary with time. The receive antenna consists of two crossed dipoles. The solid line results from adaptation and the dashed line has

Fig. 11. In this case, no adaptation.

=

vary with time. The receive antenna consists of three crossed dipoles. The solid line results from adaptation and the dashed line has

link improvement than the two dipole case, particularly at smaller angles (Fig. 11). The maximum improvement is 3 dB . at How fast can a GA adapt? If the receive antenna continuously adapts as it moves, then only small perturbations are necessary at each angle and the adaptation is very fast. As an example, consider maximizing the link budget of the receive antenna with . If the adaptation starts three orthogonal dipoles at , then the solid curve with the optimal weightings at in Fig. 12 results. In order to reach the steady state solution at

generation 21, the number of ﬁtness function evaluations made by this GA run is

(9) A ﬁtness function evaluation equates to a power measurement in a real system. If the adaptation starts with the optimal weightings at , then the dashed curve in results. Continuously

HAUPT: ADAPTIVE CROSSED DIPOLE ANTENNAS USING A GA

1981

Fig. 12. Typical link improvement versus generation for a three orthogonal dipole receive antenna at an angle of 50 . If the adaptation process is started with the receive antenna at circular polarization (0 ), then the GA ﬁnds an optimum in 21 generations or about 100 power measurements (solid line). If the process starts with the optimum dipole weights at 45 , then minimal adaptation is necessary (dashed line). TABLE I AVERAGE MAXIMUM AND MEAN OF THE POPULATION OVER 100 GENERATIONS FOR VARIOUS MUTATION RATES AND NOISE VARIANCES

Fig. 13. Performance of the GA when normally distributed noise is added to the amplitude and phase of the currents fed to the dipoles ( 0 and = 0:1). The GA has a population size of 8 and mutation rate of 0.2.

=

adapting the signal results in constant incremental improvement of the link. Even if the antenna must be adapted from circular polarization, the GA quickly ﬁnds an acceptable solution. The results are not dependent upon . Noise was added to the currents of the dipoles of the transmit antennas to see how well the GA performs in a noisy environment. Using a population size of 8 and mutation rate of 0.2, a plot of the link improvement versus generation shows ups and downs due to the random variations (Fig. 13). The transmit dipole current amplitude and phase errors are normally distributed with a mean and standard deviation given by and . Note that the mean of the population has high variations due to the large mutation rate. Using a small mutation rate of 0.02, results in a much lower variations in the mean of the population (Fig. 14). Even though the higher mutation rate ﬁnds an optimal solution faster than the lower mutation rate, the average power measurements associated with the high

Fig. 14. Performance of the GA when normally distributed noise is added to the amplitude and phase of the currents fed to the dipoles ( = 0 and = 0:1). The GA has a population size of 8 and mutation rate of 0.02.

mutation rate are lower. Consequently, when noise is included, a lower mutation rate is more desirable. Table I shows that the average for the population mean over 100 generations is better when the mutation rate is smaller.

1982

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

V. CONCLUSION The received signal in a mobile communications system loses strength due to a decrease in antenna directivity and polarization mismatch. The current fed to a set of crossed dipoles can be modiﬁed to increase the directivity and polarization match between the transmit and receive antennas. Two orthogonal dipoles can compensate for the loss in gain but not polarization. Three adaptive orthogonal dipoles can fully restore the loss due to loss in directivity and polarization mismatch if tracked by the transmit antenna. A GA quickly adapts the receive antenna to the transmitted signal. Three orthogonal dipoles provide more improvement than just two orthogonal dipoles. REFERENCES

Fig. 15. Performance of the GA when normally distributed noise is added to the amplitude and phase of the currents fed to the dipoles ( 0 and = 0:1) and the satellite is moving 1 per generation. The GA has a population size of 8 and mutation rate of 0.2.

=

[1] A. Singer, “Space versus polarization diversity,” Wireless Review, pp. 164–168, Feb. 15, 1998. [2] M. R. Andrews, P. P. Mitra, and R. deCarvalho, “Tripling the capacity of wireless communications using electromagnetic polarization,” Nature, vol. 409, pp. 316–318, Jan 18, 2001. [3] R. E. Marshall and C. W. Bostian, “An adaptive polarization correction scheme using circular polarization,” in Proc. IEEE Int. Antennas and Propagation Society Symp., Atlanta, GA, June 1974, pp. 395–397. [4] R. T. Compton, “On the performance of a polarization sensitive adaptive array,” IEEE Trans. Antennas Propagat., vol. AP-29, pp. 718–725, Sept. 1981. [5] , “The tripole antenna: an adaptive array with full polarization ﬂexibility,” IEEE Trans. Antennas Propagat., vol. AP-29, pp. 944–952, Nov. 1981. [6] B. D. Grifﬁn, R. Haupt, and Y. C. Chung, “Adaptive polarization for spacecraft communications system,” presented at the Proc. IEEE Aerospace Conf., Big Sky, MT, Mar. 2002. [7] R. Haupt, “Adaptive crossed dipole antennas,” in URSI General Assembly, Maastricht, Netherlands, Aug. 2002.

Fig. 16. Performance of the GA when normally distributed noise is added to the amplitude and phase of the currents fed to the dipoles ( = 0 and = 0:1) and the satellite is moving 1 per generation. The GA has a population size of 8 and mutation rate of 0.02.

In a fully adaptive system, the GA would also have to adapt while the satellite moves. The next set of simulations used the same error statistics as before and had the satellite move 1 per generation. Fig. 15 shows the convergence curve when the mutation rate is 0.2. Again, the mean of the population has very high variations. Fig. 16 shows the convergence curve when the mutation rate is 0.02. The smaller mutation rate is more desirable, because the variations in the population mean are small. A high mutation rate works best for a no noise environment, and a low mutation rate works best in the presence of noise.

Randy L. Haupt (M’82–SM’90–F’00) received the B.S. degree in electrical engineering from the U.S. Air Force Academy, U.S. Academy, CO, the M.S. degree in engineering management from Western New England College, Springﬁeld, MA, in 1981, the M.S. degree in electrical engineering from Northeastern University, Boston, MA, in 1983, and the Ph.D. degree in electrical engineering from the University of Michigan, Ann Arbor, in 1987. He was a Professor of electrical engineering at the U.S. Air Force Academy and Professor and Chair of Electrical Engineering at the University of Nevada - Reno. In 1997, he retired as a Lt. Col. in the U.S. Air Force. He was a Project Engineer for the OTH-B radar and a Research Antenna Engineer for Rome Air Development Center. From 1999 to 2003, he was Professor and Department Head of Electrical and Computer Engineering at Utah State University, Logan. He is currently a Senior Scientist at the Applied Research Laboratory, Pennsylvania State University, State College. He has many journal articles, conference publications, and book chapters on antennas, radar cross section and numerical methods and is coauthor of the book Practical Genetic Algorithms, 2nd edition (New York: Wiley, May 2004). He has eight patents in antenna technology. Dr. Haupt is a Member of Tau Beta Pi, Eta Kappa Nu, International Scientiﬁc Radio Union (URSI) Commission B, and the Electromagnetics Academy. He was the Federal Engineer of the Year in 1993.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

1983

**Modeling and Investigation of a Geometrically Complex UWB GPR Antenna Using FDTD
**

Kwan-Ho Lee, Student Member, IEEE, Chi-Chih Chen, Member, IEEE, Fernando L. Teixeira, Member, IEEE, and Robert Lee, Member, IEEE

Abstract—A detailed analysis of ultrawide-band (UWB), dual-polarized, dielectric-loaded horn-fed bow-tie (HFB) antennas is carried out using the ﬁnite-difference time-domain (FDTD) method. The FDTD model includes realistic features of the antenna structure such as the feeding cables, wave launchers, dielectric loading, and resistive-ﬁlm loading. Important antenna characteristics that are usually difﬁcult to obtain via measurements can be obtained more directly from this FDTD model. Since the HFB antennas under consideration are intended for ground penetrating radar (GPR) applications, the effects of the half-space medium are also investigated. The simulated results serve to verify the performance of the HFB antenna design, and to optimize various antenna parameters. Index Terms—Bow-tie antenna, coaxial cable, dielectric loading, ﬁnite-difference time-domain (FDTD), ground penetrating radar (GPR), impedance, resistive, ultrawide-band (UWB).

I. INTRODUCTION

G

ROUND PENETRATING RADAR (GPR) ﬁnd applications in many areas such as geophysical prospecting, archeology, civil engineering, environmental engineering, and defense technologies as a noninvasive sensing tool [1], [2]. One key component in any GPR system is the receiver/transmitter antenna(s). Desirable features for GPR antennas include broadband operation, good impedance matching, and small size. The frequency range of a GPR antenna is determined by the particular application and its relation to the nature of the target, soil constitution, desired depth of penetration, and inversion/classiﬁcation method being used. For example, the frequency of operation for detection and classiﬁcation of anti-tank and anti-personnel landmines is usually from 0.1 to 1 GHz [3] and from 1 to 6 GHz [4], [5], respectively. A good frequency range for detecting 6-inch drainage pipes is found to be from 100 to 400 MHz. For unexploded ordnance (UXO) detection, the 10 800 MHz frequency range is often used [6], [7]. For the detection of shallow objects where high sensitivity is not an issue, elevated antennas are often used for easier scanning and better antenna calibration. In particular, many antennas used for detection of shallow landmines [5], [8], evaluation of integrity of concrete [9] and soil hardness [10] are all elevated

Manuscript received December 2, 2002; revised November 3, 2003. This work was supported in part by the Department of Defense (DoD) Strategic Environmental Research and Development Program (SERDP) Project 1122 by the National Science Foundation (NSF) under Grant ECS-0347502. The authors are with ElectroScience Laboratory, Department of Electrical Engineering, The Ohio State University, Columbus OH, 43210 USA (e-mail: lee.1333@osu.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832501

systems that exhibits low antenna-ground interaction. On the other hand, most GPR antennas used for the detection of deep targets are operated very close to the ground so that most of the energy is radiated into the ground to improve sensitivity. This conﬁguration also minimizes radiation into the air to comply with the FCC regulations. The characteristics of such GPR antennas while in ﬁeld operation are usually difﬁcult to determine a priori because of the large coupling with the environment. For instance, the input impedance of the commonly used dipoles or ﬂat bowtie dipoles are directly affected by the electrical property of the particular ground for antennas operated close to the surface. Moreover, the amount of energy coupled into the ground changes as the permittivity increases and hence the radiation patterns also depend of the soil permittivity [11]–[13]. Hence, one major disadvantage is that the antenna characteristics in the ﬁeld become dependent on the electrical properties of the ground and surroundings. This also makes calibration more difﬁcult. In order to make antenna characteristics less susceptible to ground characteristics, a new dielectric-loaded horn-fed bowtie (HFB) antenna design was introduced in [7]. The HFB antenna was designed to minimize the antenna ringing by: 1) employing a stable and well matched surge impedance and 2) using specially designed tapered resistive loadings. Unlike most conventional antennas, the surge impedance was designed to be less dependent on the ground property because the feed point is elevated off the ground. Low loss dielectric material was then used to ﬁll the space between the feed front and the ground surface to reduce ground-surface reﬂections and increase the electrical height of the feed. Both single-polarized and dual-polarized HFB antenna prototypes have been built and employed in actual applications. Due to its ﬂexibility, the ﬁnite-difference time-domain (FDTD) method has been widely used in recent years for the numerical simulation of GPR systems [14]–[17]. Some of the previous studies have modeled GPR antennas as a series of point sources or short dipoles with or without the presence of conducting shields [18], [19]. In order to better characterize HFB antennas and to provide a more convenient tool for their design and optimization, a full-scale detailed three-dimensional (3-D) FDTD model of a dual-polarized HFB prototype was developed in this work and simulated for GPR applications. To reduce the computational cost, a special partition scheme [20] is adopted for the 3-D FDTD domain. This scheme divides the whole inhomogeneous region into several small homogeneous regions. In each homogeneous region, volumetric material property matrices are replaced by constants to save the

0018-926X/04$20.00 © 2004 IEEE

1984

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 1.

Prototype of HFB antenna.

memory. This partition scheme for modeling the electrically large HFB antenna in the presence of ground also allows for faster simulations on a personal computer. An anisotropic perfectly matched layer (APML) especially formulated for the dielectric or lossy half spaces [16], [21], [22] is implemented. This paper is organized as follows. The HFB antenna design is discussed in Section II. Section III describes the construction of the FDTD model for the dual-polarized HFB design and the performance of the resistive-ﬁlm loading which is optimized for a given length. Section IV presents various HFB antenna characteristics obtained from the FDTD simulations. II. BASIC DUAL-POLARIZED HFB ANTENNA DESIGN Fig. 1 illustrates the basic structure of the dual-polarized UWB HFB antenna design. This is somewhat similar to a planar bowtie dipole with the feed point being raised off the ground. The feed section resembles that of a small transverse electromagnetic (TEM) horn except that it is ﬁlled with low loss dielectric material. Each antenna arm is smoothly curved in the transition from the horn section to the planar bowtie dipole section. The ends of the dipoles are terminated with tapered resistive cards (R-card) to reduce antenna ringing. A. Resistive Taper Section Tapered R-cards have many useful applications for radiation and scattering control [23]–[25], but commercial tapered R-cards are often expensive and have very limited choices of tapering proﬁle and taper length. The R-card used in the

HFB prototype was constructed in-house using multiple layers of commercial window ﬁlms [23]. These have various sheet resistance for different percentage of light transmission. When multiple ﬁlms are overlaid properly together, one can obtain a desired resistivity proﬁle with desired taper length. Fig. 1 illustrates how the tapered R-card was constructed for the HFB prototype. The objective of the resistive card is to reduce reﬂections by gradually dissipating the currents propagating toward the end of each antenna arm. This requires the resistivity on the R-card to be tapered from a small value to a large value along the antenna arm. An exponential taper of the resistivity was adopted in the HFB prototype with a tapering shown as follows: (1) where is the initial sheet resistance of the R-card at the perfect electric conducting (PEC)/R-card interis the sheet resistance at the far face, and is the length of the R-card, and end of the R-card, is the distance along the R-card from the PEC arm. B. Feed Section The feed section of HFB resembles a dual polarized TEM horn except that the end of each antenna arm is curved outward gradually to be connected to the ﬂat bowtie section, and the internal space of the horn was ﬁlled with low loss dielectric material. The geometry of the horn and the antenna arms was

LEE et al.: MODELING AND INVESTIGATION OF A GEOMETRICALLY COMPLEX UWB GPR ANTENNA USING FDTD

1985

Fig. 2.

Dimension and computation domain partitioning of the fully polarimetric dielectric ﬁlled HFB antenna.

chosen based upon the tradeoff among the dielectric constant, size, weight, and cost. The objective was to obtain a surge impedance of 100 to match to the characteristic impedance of the feeding twin-coaxial cables shown at the bottom of Fig. 1 (each cable has a characteristic impedance of 50 ohms). Although tabulated characteristic (or surge) impedances for an inﬁnite TEM horn with arbitrary geometry are available [26], [27], the exact impedance of a dual-polarization TEM horn with dielectric ﬁlling is complicated to obtain analytically. The experimental data obtained from [28] was used during the construction of HFB prototype. Note that the center of each coaxial cable was connected to one antenna arm and coaxial cable feed one polarization. each pair of the 50 A 0–180 broadband hybrid was used as a balun for each pair of cables. Accurate FDTD models recently constructed to calculate the surge impedance for such an antenna geometry are employed here [29]. The prototype to be analyzed here has a dielectric constant of 5. The plate angle of each antenna arm is 11.5 . The horn angle itself is approximately 150 . III. FDTD MODEL DESCRIPTION A full scale model of the UWB HFB antenna prototype respace. A spatial quires a minimum of cell size of 6.3 mm was chosen to accurately model the geometrical details of the antenna and cable structure [30]. This yields approximately 96 million unknowns. The FDTD grid is shown in Fig. 2. All dimensions in the model were chosen to be as close to the actual prototype as possible. The four antenna arms were modeled as PEC plates, and the curved edges and surfaces were approximated by staircases. Each tapered R-card attached to the

end of the PEC arm is 63 cm in length and is implemented via a conductive sheet. The ground was assumed to be a lossless half space with relative permittivity of 5. A. Heterogeneous FDTD Domain Partition The antenna geometry under study is very complicated and resides in a complex environment. A traditional FDTD approach to represent the geometry would require either the storage of the material properties for each cell or else a data organization similar to what is used in the ﬁnite element method, which would also require a signiﬁcant amount of memory overhead. To minimize the memory usage, we have adopted a partitioning scheme [20]. The FDTD domain is divided into blocks. The size and number of the blocks are judiciously chosen, so that the material properties within most of the blocks are homogeneous. Within the code, the FDTD algorithm is computed in different ways, and based on the properties of the block, the appropriate FDTD algorithm will be chosen. If the block is a perfect conductor, the FDTD code will recognize this, and not perform any computations for that block. Thus, there is no need to store either the ﬁelds or the material properties for that particular block. If the block is an homogeneous dielectric, the material properties are not treated as a function of the grid points within the block but instead represented just as a constant parameter. Thus only the ﬁeld values need to be stored for each cell within that block. If the block is an inhomogeneous dielectric, then the FDTD algorithm used will assume a constant permeability and no conductivity. Thus, only the ﬁelds and permittivity must be stored for each cell. In our case, we divide the geometry into 196 blocks with only ﬁve of the blocks being heterogeneous as demonstrated in Fig. 2.

1986

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 3. Coaxial cable modeling in rectangular FDTD grid and the TEM current excitation scheme. (a) Top view, (b) side view, and (c) J (t).

B. Feed Cable Modeling In the discretized FDTD model, each coaxial cable has a square cross sectional area with a single-cell PEC wire surrounded by four PEC walls. As shown in Fig. 3(a), a relative dielectric constant of 1.5 is speciﬁed between the center wire and the PEC walls. Each cable is terminated with perfectly matched layer at one end and connected to the tip of an antenna arm at the other end. A balanced excitation is introduced to the opposite pair of cables to excite one antenna polarization as shown in Fig. 3(a) and (b). The time history of the response is also recorded at the excitation position to obtain reﬂection and transis obtained with the exmission data. The reﬂection data citation and observation points co-located in the same cable and is obtained with the observation point cross-coupling data located at the second cable. A differential Gaussian pulse is chosen as the time-domain excitation current (2) , and . where These parameters for the Gaussian pulse are determined so as to provide signiﬁcant spectral energy in the frequency range of 10 to 800 MHz. Fig. 3(c) illustrates the pulse. C. Resistive Card Modeling In the FDTD model, the R-card is modeled as a single-cell corresponding to the delayer with a tapered conductivity . Conductivity along the direction sired sheet resistance where is the thickness of is calculated by

**Fig. 4. Resistive card overlay conﬁgurations for the PEC launcher section = 300
**

= , R = 3

= ). (R

single layer. This assumption is valid when is much greater than the penetration depth but much smaller than the free space wavelength [31]. In addition to the exponential taper described in Section II-A, a linear taper with the following taper function was also investigated using the FDTD model as a comparison (3) where is the initial sheet resistance of the R-card is the end sheet at the PEC/R-card junction, resistance. The taper length is equal to that of the previous

LEE et al.: MODELING AND INVESTIGATION OF A GEOMETRICALLY COMPLEX UWB GPR ANTENNA USING FDTD

1987

Fig. 5.

S

, S , and surge impedance of the HFB antenna. (a) Reﬂection coefﬁcient S

and S

and (b) antenna surge impedance.

exponential taper, i.e., 0.63 m. As it will be shown shortly, a linear taper provides a better performance, i.e., lower reﬂection at low frequency end, due to relatively short taper length with respect to wavelength. A more detailed analysis on this aspect can be found in [29]. Fig. 4 plots the linear resistive taper as well as its position relationship with respect to the antenna arm. The lateral edges of the R-card were kept aligned to the edges of the PEC arms to avoid undesired diffractions (see Fig. 1). IV. CHARACTERISTICS OF DUAL-POLARIZED HFB ANTENNA DESIGN A. & and Input Impedance The simulated and measured reﬂection and transmission coand , of the HFB design are compared in efﬁcients, Fig. 5(a). Note that the antenna is located on the surface of a half space with a dielectric of 5, corresponding to the dry sand in reis similar to since the both antenna arms have ality. The

and provide the co-polarized backscatthe same design. provides the cross-polarized kscattering data. tering data. A calibration procedure was carried out in a similar manner as done in real measurement using “short” and “matched” (PML) reference loads at the end of the feed cables

(4) is the response obtained with the coaxial In the above, cables connected to the antenna, is the response obtained is the with the coaxial cables connected to matched load, response obtained when the coaxial cables are shorted at the end is the response obtained at coaxial with a conducting wire, cable 2 with antenna connected when the excitation is applied is the incident wave. to cable 1, and It is observed that the both linear and exponential taper have similar performance at frequency above 0.3 GHz where the

1988

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 6. Comparison of reﬂected electric ﬁeld difference with various ground proﬁles.

Fig. 7. Reﬂected (E ) ﬁeld in time domain for HFB antennas with different resistive card overlay conﬁgurations and using same conductivity proﬁle.

taper length becomes comparable or longer than one wavelength (considering dielectric constant of 5). It is also observed that the linear taper produces lower reﬂection level than the exponential taper at frequencies below 0.1 GHz. Overall, the reﬂection level is less than 10 dB above 0.05 GHz. This veriﬁes broadband characteristic of the HFB design. The measured data is found to be on average 10 dB higher than that predicted from the simulation. This difference is most likely caused by the asymmetry of the construction of prototype antenna arms and the feed structure. Good agreement between the measurement and simulation is the result of the geometrical ﬁdelity between of the FDTD numerical model and the prototype, including the R-card geometry, conductive plates, dielectric ﬁlling and coaxial cable feed modeling. However, the prototype measurement introduces additional environmental variables more difﬁcult to control such as ground loss, slight asymmetry of the antenna arm design due to hand-made fabrication, and discrepancies between the equivalent conductive single layer R-card used in the FDTD model and the thin ﬁlm R-card conductivity value.bac as The surge impedance can also be calculated from the shown in (5), often applying a time gate to keep only the ﬁrst peak associated with the feed point near 0 ns position as shown in Fig. 6 (5) is the characteristic impedance of the twin-coaxial where cable. The resultant surge impedance is shown in Fig. 5(b). For , the surge most of the band range, as desired. impedance is found to be within B. Ground Effect In order to see how the ground properties affect the surge impedance of the HFB design, four different ground dielectric constants: 5, 7, 9, and 11 are simulated. Fig. 6 shows the reﬂected ﬁeld from 0 to 3 ns. The height of the antenna feed above the ground is equal to 0.1 m. This causes the reﬂection from the ground surface to be delayed by approximately 1.5 ns since the antenna dielectric ﬁller has a relative permittivity of 5. This

Fig. 8. Comparison of co-polarized (E ) reﬂected ﬁeld in time domain from HFB antenna with the different resistive cards (R ).

agrees with the signiﬁcant variations shown in the data near 1.5 ns position. Most importantly, the ﬁrst reﬂection peak arising from the feed point remain unaffected by the ground property, as desired. C. R-Card Performance Investigation We investigate two parameters that play an important role in minimizing reﬂections from the truncated antenna arms. The ﬁrst parameter is the overlay distance between the PEC and R-card. In the actual HFB prototype, a 5 cm overlay was used to allow the electromagnetic energy to be coupled into the R-card section because the R-card was coated with a protective insulator and could not have a direct electrical contact with the antenna arm. The second parameter is the far-end resistance value that affects the tapering rate of the R-card. If the taper is done too rapidly, undesired diffractions would be produced by the R-card. On the other hand, if the taper is too slow, the far-end reﬂection may still be too strong. To investigate the effect of PEC and R-card overlay distances, the following three cases were simulated as shown in Fig. 4. In case 1 through 3, the overlay distances are 11.3, 5, and 0 cm, respectively. The simulated reﬂection responses are plotted in

LEE et al.: MODELING AND INVESTIGATION OF A GEOMETRICALLY COMPLEX UWB GPR ANTENNA USING FDTD

1989

**Fig. 9. Snap shots from FDTD simulation for E ﬁeld strength in dB scale where R = 300
**

= . (a) t = 7:4539 ns with R-card attached-(dB) scale; (b) t = 7:4539 ns without R-card-(dB) scale; (c) t = 13:0954 ns with R-card attached-(dB) scale; (d) t = 13:0954 ns without R-card-(dB) scale.

Fig. 7. As expected, the overlay distance of the tapered R-card affects the reﬂection at the PEC end. Note that the R-card in the overlay section is shorted out by the PEC, this section would have an effective resistance of zero regardless of the R-card value. The larger geometric discontinuity in Case 3 provides the stronger junction reﬂection observed in the ﬁgure. Case 1 and 2 provide a smoother transition and result in a 35 dB reﬂected ﬁeld at the end of the R-card. Based on the simulations, we concluded that a linear-tapered R-card with either a 11.3 or 5 cm overlay at the PEC/R-card does the best job of suppressing the reﬂections. , values of 100, 200, 300, and To optimize the choice of were implemented and simulated separately. From the 400 reﬂected ﬁeld observed at the feed point, the amount of endreﬂection suppression was compared as shown in Fig. 8, where late time (after 20 ns) antenna reﬂections can be observed. These provides the maximal results indicate that suppression of the arm end reﬂections. D. Antenna Ringing Fig. 9 compares snapshots of the instantaneous ﬁeld dis) plane with and without the tribution in the vertical (or

R-card attached to the HFB antenna arms. Without the R-card, signiﬁcant diffraction and reﬂection at the end of the PEC arms are observed. The reﬂected ﬁelds later propagate back to the observation point inside the cables as shown in Fig. 9(b). On the other hand, the R-card extension signiﬁcantly reduces the diffraction and reﬂection at the ends as depicted in Fig. 9(a) and lowers the antenna ringing by approximately 20 dB. Note that the signals that propagate back to the feed point are partially reﬂected due to the imperfect matching. This reﬂected ﬁelds generate the secondary reﬂection. This process repeats and becomes the well known “antenna ringing” effect, a major clutter source in GPR measurements. E. Radiated Field Distribution & Polarization The near-ﬁeld radiation characteristics are investigated next. Fig. 10 depicts the simulated horizontal co-polarized and crosspolarized ﬁeld distributions at a plane 40 cm below the antenna aperture, (corresponding to the ground surface plane), at the center frequency of 400 MHz. The cases with and without the R-card are also plotted for comparison. The ﬁelds are nearly linearly polarized in the principal planes. The results with the R-card clearly show a more uniform distribution, because the

1990

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Fig. 10. Comparison of co- and cross-polarized aperture ﬁeld distributions at f = 400 MHz, depth z = 40 cm or 0.53 ,R = 300
**

= in R-card. (a) Co-polarized ﬁeld with R-card, (b) cross-polarized ﬁeld with R-card, (c) co-polarized ﬁeld without R-card, and (d) cross-polarized ﬁeld without R-card.

diffracted ﬁelds from the antenna arm ends modify the radiated ﬁelds that otherwise would have been close to simple spherical wavefronts. The more uniform ﬁeld distributions simplify the subsequent signal processing and inverse problem and improve the overall detection/classiﬁcation capabilities of a GPR system. As the observation point moves away from the principal planes, the level of depolarization increases and reaches a maximum of approximately 12 dB between the two antenna polarizations. This is, of course, due to the spherical nature of the wavefront. We note that the cross-polarized ﬁeld levels with the R-card present are a little bit higher. This again may be caused by distributed diffractions along the resistive cards. V. CONCLUSION In this work, a detailed FDTD model was used to incorporate realistic features of UWB HFB antennas such as feeding cables, dielectric loading and tapered resistive terminations. The FDTD model is ﬂexible enough to model different geometries, structures, and materials for both the antenna and the ground medium. Fully-polarimetric simulations were performed to obtain the radiation characteristics of HFB antennas over a broad frequency range. A parametric study on the effect of the resistive taper of the R-card termination was also performed. It was found that a linear taper performs better than the commonly used exponential taper for short taper length. It was also found that a proper overlapping between the PEC and R-card improves the transition and reduces the diffraction at the end of PEC. The R-card termination also signiﬁcantly reduces the undesired antenna ringing. The surge impedance of the

HFB antenna was calculated from the reﬂection coefﬁcients and was found to be approximately 100 ohms over the entire frequency band of interest. This result conﬁrms the broadband characteristic of the HFB design. The FDTD model also provided useful visualization of dynamic ﬁeld distributions that can help identify undesired radiations and reﬂections sources. The near-ﬁeld distributions of the co-polarized and cross-polarized ﬁelds were examined. This information is particularly useful in GPR applications where the depth of the target is unknown. Overall, the simulated results conﬁrm that the optimized HFB antenna design is a very attractive choice for broadband, fully polarimetric GPR applications. ACKNOWLEDGMENT The authors acknowledge the reviewers for their helpful comments. REFERENCES

[1] L. Peter Jr, J. D. Young, and J. Daniels, “Ground penetration radar as a subsurface environmental sensing tool,” Proc. IEEE, vol. 82, pp. 1802–1822, Dec. 1994. [2] S. Agosti, G. G. Gentili, and S. Spagnolini, “Electromagnetic inversion of monostatic GPR: application to pavement proﬁle,” in Proc. Int. Conf. Electromag. Adv. Applicat. (ICEAA’97), 1997, pp. 491–494. [3] L. C. Chan, D. L. Moffatt, and L. Peters Jr., “A characterization of subsurface radar targets,” Proc. IEEE, vol. 67, pp. 91–110, July 1979. [4] C.-C. Chen, S. Nag, W. Burnside, J. Halman, K. Shubert, and L. Peters Jr, “A stand-off, focused-beam land mine radar,” IEEE Trans. Geosci. Remote Sensing, vol. 38, pp. 507–514, Jan. 1998. [5] C.-C. Chen, K. R. Rao, and R. Lee, “A new ultra-wide bandwidth dielectric rod antenna for ground penetrating radar applications,” IEEE Trans. Antennas Propagat., vol. 51, pp. 371–377, Mar. 2003.

LEE et al.: MODELING AND INVESTIGATION OF A GEOMETRICALLY COMPLEX UWB GPR ANTENNA USING FDTD

1991

[6] C.-C. Chen and L. Peters Jr, “Buried unexploded ordnance identiﬁcation via complex natural resonances,” IEEE Trans. Antennas Propagat., vol. 45, pp. 1645–1654, Nov. 1997. [7] C. C. Chen, B. Higgins, K. O’Neil, and R. Detsch, “Ultrawide-bandwidth fully-polarimetric ground penetrating radar classiﬁcation of subsurface unexploded ordnance,” IEEE Trans. Geosci. Remote Sensing, vol. 39, pp. 1259–1270, June 2001. [8] T. P. Montoya and G. S. Smith, “Land mine detection using a groundpenetrating radar based on resistively loaded vee dipoles,” IEEE Trans. Antennas Propagat., vol. 47, pp. 1795–1806, Dec. 1999. [9] J. Hugenshmidt, “A one-to-one comparison between radar results and reality on a concrete bridge,” in Proc. 9th Int. GPR Conf., vol. SPIE4758, May 2001, pp. 427–432. [10] M. Higgins and C.-C. Chen, “Nondestructive evaluation of soil hardness using elevated focused-beam radar,” in Proc. 9th Int. GPR Conf., vol. SPIE-4758, May 2001, pp. 54–57. [11] M. Moghaddam, W. C. Chew, B. Anderson, E. Yannakis, and Q. H. Liu, “Computation of transient electromagnetic waves in inhomogeneous media,” Radio Sci., vol. 26, pp. 265–273, Jan. 1991. [12] S. J. Radzevicius, J. J. Pariels, and C.-C. Chen, “GPR H-Plane Antenna Patterns for a horizontal dipole on a half space interface,” in Proc. 8th Int. GPR Conf., vol. SPIE-4084, Gold Coast, Australia, June 2000. [13] C. C. Chen and J. D. Young, “Unfurlable folded-dipole UWB antenna for mars explorer subsurface sensing,” in Proc. 8th Int. GPR Conf., vol. SPIE-4084, Gold Coast, Australia, Jun. 2000. [14] M. Moghaddam, E. J. Yannakakis, W. C. Chew, and C. Randoll, “Modeling of the subsurface interface radar,” J. Electromag. Waves Applicat., vol. 5, pp. 17–39, 1991. [15] J. M. Bourgeois and G. S. Smith, “A fully three-dimensional simulation of a ground-penetrating radar: FDTD theory compared with experiment,” IEEE Trans. Geosci. Remote Sensing, vol. 34, pp. 36–44, 1996. [16] K. R. Rao, K. H. Lee, C. C. Chen, and R. Lee, “Application of fullpolarmetric ground penetration radar for buried UXO Classiﬁcation,” The Ohio State Univ., ElectroSci. Lab., Tech. Rep. 738 520-1, Feb. 2001. [17] K.-H. Lee, N. V. Venkatarayalu, and C.-C. Chen, “Numerical modeling development for characterizing complex gpr problems,” in Proc. Int. GPR Conf., vol. SPIE-4758, May 2002, pp. 625–652. [18] F. L. Teixeira, W. C. Chew, M. Straka, M. L. Orstaglio, and T. Wang, “Finite-difference time-domain simulation of ground penetrating radar on dispersive inhomogeneous, and conductive soils,” IEEE Trans. Geosci. Remote Sensing, vol. 36, pp. 1928–1937, Nov. 1998. [19] L. Gurel and U. Oguz, “Three-dimensional FDTD modeling of a ground penetrating radar,” IEEE Trans. Geosci. Remote Sensing, vol. 38, pp. 1513–1521, Jul. 2000. [20] J. Nehrbass, “Physics based partitioning,” in Proc. 26th General Assembly for URSI, Ontario, Canada, Aug. 2000. [21] F. L. Teixeira and W. C. Chew, “Finite-difference simulation of transient electromagnetics ﬁelds for cylindrical geometries in complex media,” IEEE Trans. Geosci. Remote Sensing, vol. 38, pp. 1530–1543, July 2000. [22] K.-H. Lee, N. Venkalayalu, C.-C. Chen, F. L. Teixeira, and R. Lee, “Application of full-polarmetric ground penetration radar for buried UXO Classiﬁcation (II),” The Ohio State Univ., ElectroSci. Lab., Tech. Rep. 778 520, May 2002. [23] C. Handel, I. J. Gupta, and W. D. Burnside, “Low frequency modiﬁcation of a dual chamber compact range,” The Ohio State Univ., ElectroSci. Lab., Tech. Rep. 732 264, Sep. 1997. [24] L. Chaung, T. Chang, and W. D. Burnside, “An ultrawide-bandwidth tapered resistive TEM horn antenna,” IEEE Trans. Antennas Propagat., vol. 48, pp. 1848–1857, Dec. 2000. [25] M. S. A. Mahmoud, T.-H. Lee, and W. D. Burnside, “Enhanced compactrange reﬂector concept using an R-card fence: two-dimensional case,” IEEE Trans. Antennas Propagat., vol. 49, pp. 419–428, Mar. 2001. [26] F. C. Yang and K. S. H. Lee, “Impedance of a Two-Conical-Plate Transmission Line,” Tech. Rep., Sensor and Simulation Company, Nov. 1976. [27] H. M. Shen, R. W. P. King, and T. T. Wu, “V-conical antenna,” IEEE Trans. Antennas Propagat., vol. 36, pp. 1519–1525, Nov. 1988. [28] C. C. Chen, “A new ground penetrating radar antenna design—the horn-fed bowtie (HFB),” in Proc. Antenna Measurement Techniques Association (AMTA) Symp., Nov. 1997, pp. 67–74. [29] N. Venkatarayalu, C.-C. Chen, F. L. Teixeira, and R. Lee, “Modeling of ultrawide-band dielectric horn antennas using FDTD,” IEEE Trans. Antennas Propagat., vol. 52, pp. 1318–1323, May 2004. [30] A. Taﬂove, Computational Electrodynamics. Norwood, MA: Artech House, 1995. [31] T. B. A. Senior, “Approximate boundary conditions,” IEEE Trans. Antennas Propagat., vol. 29, pp. 826–829, Sept. 1981.

Kwan-Ho Lee (M’02) received the B.S. degree from the Department of Radio Science and Engineering, Kwangwoon University, Seoul, Korea, in 1997 and the M.S. degree in electrical engineering from The Ohio State University, Columbus, in 1999, where he is currently working toward the Ph.D. degree. Since 1997, he has been a Graduate Research Associate at the ElectroScience Laboratory, Department of Electrical Engineering, The Ohio State University. His research interests include computational electromagnetics, ultrawide-bandwidth antenna development, subsurface target detections and classiﬁcations, RF circuits and object oriented programming.

Chi-Chih Chen (S’92–M’97) was born in Taiwan, R.O.C., in 1966. He received the B.S.E.E. degree from the National Taiwan University, Taiwan, R.O.C., in 1988 and the M.S.E.E. and Ph.D. degrees from The Ohio State University, Columbus, in 1993 and 1997, respectively. He joined the ElectroScience Laboratory, The Ohio State University, as a Postdoctoral Researcher in 1997 and became a Senior Research Associate in 1999. His main research interests include the ground penetrating radar, UWB antenna designs, radar target detection and classiﬁcation methods, automobile radar systems. In recent years, his research activities have been focused on the detection and classiﬁcation of buried landmines, unexploded ordnance and underground pipes. Dr. Chen is a Member of Sigma Xi and Phi Kappa Phi.

Fernando L. Teixeira (S’89–M’93) received the B.S. and M.S. degree in electrical engineering from the Pontiﬁcal Catholic University of Rio de Janeiro (PUC-Rio), Brazil, in 1991 and 1995, respectively, and the Ph.D. degree in electrical engineering from the University of Illinois at Urbana-Champaign, in 1999. From 1999 to 2000, he was a Postdoctoral Research Associate with the Research Laboratory of Electronics, Massachusetts Institute of Technology (MIT), Cambridge. Since 2000, he has been an Assistant Professor at the ElectroScience Laboratory (ESL) and the Department of Electrical Engineering, The Ohio State University, Columbus. His current research interests include analytical and numerical techniques for wave propagation and scattering problems in communication, sensing, and devices applications. He has edited one book Geometric Methods for Computational Electromagnetics (PIER 32, EMW: Cambridge, MA, 2001), and has published over 30 journal articles and 50 conference papers in those areas. Dr. Teixeira is a Member of Phi Kappa Phi. He was awarded the Raj Mittra Outstanding Research Award from the University of Illinois, and a 1998 MTT-S Graduate Fellowship Award. He received paper awards at 1999 USNC/URSI National Symposium (Orlando, FL), and received a Young Scientist Award at the 2002 URSI General Assembly. He was the Technical Program Coordinator of the Progress in Electromagnetics Research Symposium (PIERS), Cambridge, MA, in 2000.

Robert Lee (M’92) received the B.S.E.E. degree in 1983 from Lehigh University, Bethlehem, PA, and the M.S.E.E. and Ph.D. degree in 1988 and 1990, respectively, from the University of Arizona, Tucson. From 1983 to 1984, he worked for Microwave Semiconductor Corporation, Somerset, NJ, as a Microwave Engineer. From 1984 to 1986, he was a Member of the Technical Staff, Hughes Aircraft Company, Tucson, AZ. From 1986 to 1990, he was a Research Assistant at the University of Arizona. During summer 1987 through 1989, he worked at Sandia National Laboratories, Albuquerque, NM. Since 1990, he has been at The Ohio State University, where he is currently a Professor. His major research interests are in the development and application of numerical methods for electromagnetics.

1992

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Radiation Properties of an Arbitrarily Flanged Parallel-Plate Waveguide
**

Dao Ngoc Chien, Student Member, IEEE, Kazuo Tanaka, Member, IEEE, and Masahiro Tanaka, Member, IEEE

Abstract—The radiation properties of an arbitrarily ﬂanged parallel-plate waveguide are investigated by means of the boundary integral equations that are called guided-mode extracted integral equations. The boundary integral equations derived in this paper can be solved by the conventional boundary-element method. Numerical results are presented for a number of cases of ﬂanged parallel-plate waveguide. Reﬂection coefﬁcient, reﬂected and radiated powers as well as radiation patterns are numerically calculated for the incidence of transverse electric guided-mode wave. Index Terms—Boundary-element method (BEM), boundary integral equations (BIE), electromagnetic radiation, numerical analysis, parallel-plate waveguides (PPW).

(GMEIEs). We derive GMEIEs for the problems of dielectric ﬁlled and unﬁlled PPW having an arbitrarily ﬂanged surface. By treating these problems, we can easily understand the advantages of GMEIEs compared with other techniques proposed before. Since the method in this paper does not employ any approximation, the results are accurate in principle. The numerical results of computer simulations are presented, in which, the reﬂection coefﬁcient, the reﬂected and radiated powers as well as the radiation pattern are calculated numerically for the incidence of TE guided-mode wave. The results are compared with those reported in the literature, and are conﬁrmed by the law of energy conservation. II. PPW WITH AN ARBITRARILY FLANGED SURFACE A. Formulation of the Reﬂection Coefﬁcient in Terms of GMEIEs Consider a dielectric ﬁlled PPW of width having a tilted ﬂange surface radiating into a free space as shown in Fig. 1(a). The dielectric is with refractive index of . The waveguide is assumed to be satisﬁed the single-mode condition. Referring to Fig. 1(d), we denote the actual boundaries (solid – . The boundary (dotted lines) of the waveguide by line) does not express an actual boundary, but rather express a virtual boundary. The whole space is assumed to be magnetically homogeneous with a magnetic permeability H/m. In the following analysis, a harmonic time depenis supposed and suppressed for the electromagdence netic ﬁeld quantities, the free-space wave number is denoted by , where is the velocity of light in a vacuum. Since the waveguide is assumed to be inﬁnite-extended in the -direction, all ﬁeld quantities are independent of (i.e., ) and thus the electromagnetic ﬁeld can be decomposed in terms of TE mode. guidedTo derive GMEIEs, we assume that a dominant mode wave is incident upon the aperture in the tilted ﬂange surface from inside of the waveguide. Since the electric ﬁelds have only a -component under the above-mentioned condition, we denote the electric ﬁelds of the -component by (1) in the coordinate systems and , as shown in , the reﬂected Fig. 1(a). The incident guided-mode wave , and the radiated wave are guided-mode wave used to express electric ﬁeld quantities. We ﬁrst consider the case in which an observation point is in the region surrounded by the boundary . From Maxwell’s equations and Green’s theorem,

I. INTRODUCTION

T

HERE has been remarkable progress in the development of communication systems over the last decade. Significant improvements in noise ﬁgure, gain, output power, and efﬁciency have been achieved at millimeter-wave frequencies. However, the demand of the wireless broadband communication at millimeter-wave frequency recently increases with activities of digital multimedia-contents circulation. One of the problems of millimeter-wave communication is the large transmission loss in free space. For instance, the transmission loss of the signal at 60 GHz frequency for 5 m distance between transmitter and receiver is about 82 dB [1]. Therefore, the antenna with high output radiation power is required to compensate the large transmission loss. The ﬂange-shaped parallel-plate waveguide (PPW) is known well as a fundamental structure extensively used for electromagnetic wave radiation (as, e.g., in feed horns, ﬂush-mounted antennas, etc.). So far, even though a closed-form solution to the problem of the ﬂanged PPW radiation is unavailable, the waveguide-radiation behavior has been well understood using a number of numerical techniques and approximate theories [2]–[13]. However, since most of the studies have based on the approximation solution, the presented results have restricted to the problem of perpendicularly ﬂanged PPW. As far as we know, no one has reported to the problem of an arbitrarily ﬂanged PPW that expects to give high output radiation power. In this paper, the radiation properties of an arbitrarily ﬂanged PPW are investigated by the boundary-element method (BEM) based on the guided-mode extracted integral equations

Manuscript received July 9, 2003. The work of D. N. Chien was supported by the Rotary Yoneyama Memorial Foundation, Inc., Japan, under a Yoneyama Scholarship. The authors are with the Department of Electronics and Computer Engineering, Gifu University, Gifu 501-1193, Japan (e-mail: chien@tnk.info.gifuu.ac.jp). Digital Object Identiﬁer 10.1109/TAP.2004.832323

0018-926X/04$20.00 © 2004 IEEE

CHIEN et al.: RADIATION PROPERTIES OF AN ARBITRARILY FLANGED PPW

1993

denotes the derivative with respect to the unit where to boundary C as shown in Fig. 1(d). The normal vector on boundary condition of perfect electric conductor, , is enforced in the process of deriving represents Green’s function in free space, (2). In (2), whose refractive index is given by , and it is expressed as (3) denotes the zeroth-order Hankel function of with the second kind. As can be seen, it is difﬁcult to solve the boundary integral (2) by use of the conventional BEM or method of moments (MoM) directly because of that the (2) . To overcome has an inﬁnite-length integral boundary this difﬁculty, we use the previously proposed idea [14]–[18] that: Even though the total electric ﬁelds near the aperture are very complicated, only the reﬂected guided-mode wave can survive at points far away from the aperture. Therefore we decompose the total electric ﬁelds on the boundary into the ﬁeld components as (4) the disturbed ﬁeld. In (4), is the and we call the ﬁeld reﬂection coefﬁcient. We also express the total electric ﬁelds by the same notation with the on the boundary disturbed ﬁeld as follows: (5) In (4), it is possible to consider that the disturbed ﬁeld will vanish at points far away from the aperture. Substituting (4) and (5) into (2), we obtain an integral equation that includes the semi-inﬁnite line integrals of the guidedas follows: mode wave along the boundary

(6) with

(7)

Fig. 1. (a)–(c) Models of arbitrarily ﬂanged PPW. (b) Location of the boundaries on integral equations.

Here the Green’s theorem for the guided-mode waves in is applied the region surrounded by the boundary as

the well-known boundary integral equation (BIE) for the total is given by electric ﬁeld

(8) (2)

1994

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

to the process of deriving (6). Since the boundary is a virtual boundary, theoretically, we can obtain the (8) with arbitrary position of the boundary . To derive the expression of the reﬂection coefﬁcient in terms of GMEIE, we put the observation point to far away from the aperture. Under this condition, it is possible to approximate Green’s function by the asymptotic expression as (9) with (10) (11) Substituting (9) into (6) and dividing both sides of the resultant , we can obtain the relation equation by

where (17) (18) Since will vanish at points far away from the aperture, , which has inﬁnite length, can the integral boundary be regarded as ﬁnite length in (16). When the observation point is in the free space region that , as shown in surrounded by the boundary Fig. 1(d), the well-known BIE for the total electric ﬁelds is given by

(19)

(12)

with

(13) Since it is impossible for a reﬂected radiation ﬁeld to exist at points in the waveguide far away from the aperture, we can set (14) So if we use (14) in (12), we can ﬁnd that the reﬂection coefﬁcient can be expressed in terms of GMEIE as

It can be seen that the (19) has the integral boundary also with semi-inﬁnite length. However, it is easy to truncate the boundary in the numerical solution procedure at a ﬁnite length where the total electric ﬁelds are enough small to be regarded as vanished. The BIEs (16) and (19) are equations to be solved numerically by using the conventional BEM or MM for the problem of an arbitrarily ﬂanged PPW as shown in Fig. 1(a)–(c). Once the ﬁelds on all the boundaries have been obtained. The reﬂection coefﬁcient can be obtained by the use of (15). And ﬁelds at any point can also be calculated by the boundary integral representations similar to (16) and (19). B. Radiation Fields in the Free Space The radiation ﬁeld in the free-space region can be expressed by using the asymptotic form of Green’s function in as follows: free-space with the refractive index of

(20) (15) with

Physically, the reﬂection coefﬁcient is an invariable value for a speciﬁc structure of the waveguide and thus we can use (15) to verify the independence of the numerical results on the location of the virtual boundary . Substitution of (15) into (6) yields

(21)

(16)

So far, we have discussed to the case in which a dielectric with is ﬁlled inside the waveguide. For the the refractive index of case of dielectric unﬁlled PPW, only one GMEIE is required. Because it is easy to derive by using the same procedure as that used in the above derivation of (16), it is not necessary to show here for saving space.

CHIEN et al.: RADIATION PROPERTIES OF AN ARBITRARILY FLANGED PPW

1995

TABLE I COMPARISON BETWEEN THE VARIOUS METHODS USED TO CALCULATE THE REFLECTION COEFFICIENT R OF A DIELECTRIC UNFILLED PPW HAVING A PERPENDICULAR FLANGE SURFACE FOR d= = 0:5001

TABLE II REFLECTED POWER 0 , RADIATED POWER 0 , AND THEIR TOTAL 0 OF A DIELECTRIC FILLED PPW HAVING A TILTED FLANGE SURFACE FOR d= = 0:5001, AND n = 1:6

Fig. 3. (a) Distribution of the disturbed ﬁeld @E =@n on the boundary C . (b) Distribution of the total ﬁeld @E =@n on the boundary C . The parameters used in calculations are the same as for Fig. 2, the virtual boundary C is located at k a = 2.

j

j

j

j

0

Fig. 2. Reﬂection coefﬁcient R of a dielectric ﬁlled PPW having a tilted ﬂange surface as a function of location of the virtual boundary C for d= = 0:5001; n = 1:6, and ' = 10 .

III. NUMERICAL RESULTS AND DISCUSSION The BIEs derived in this paper were solved with using the conventional BEM. The quadratic functions are used as basic functions, and the delta functions are used as testing functions. A. Accuracy and Convergence Tests We ﬁrst consider the problem of a dielectric unﬁlled PPW having a perpendicular ﬂange surface. Many papers have reported to this problem before and thus we can compare our results with those obtained by the methods appearing in previously published papers. In Table I the results of comparison for reﬂection coefﬁcient , including amplitude and phase, of the incident guided-mode wave are presented. It can be seen that our results are in good agreement with the results reported in the literature. Notice that owing to the different time convention used, there is a minus sign difference in the phase of reﬂection coefﬁcient in the literature.

Fig. 4. Radiated power 0 as a function of refractive index n of a dielectric ﬁlled PPW having a perpendicular ﬂange surface of width d= = 0:5001.

These numerical results show the validity of the method in this paper.

1996

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 5. Numerical results of a dielectric ﬁlled PPW having a tilted ﬂange surface for d= = 0:5001 and n = 1:6. (a) Dependence of radiated power 0 on tilting angle '. (b) Typical radiation patterns.

Fig. 7. Numerical results of a dielectric ﬁlled PPW having a tapered ﬂange surface for d= = 0:5001 and n = 1:6. (a) Dependence of radiated power 0 on tapering angle '. (b) Typical radiation patterns.

Fig. 6. Relationship between the angle of beam center and the tilting angle of a dielectric ﬁlled PPW having a tilted ﬂange surface for d= = 0:5001 and n = 1 :6 .

To verify the feasibility of the method in this paper, we next apply the method to the problem of a dielectric ﬁlled PPW having a tilted ﬂange surface as shown in Fig. 1(a). Because it seems to be difﬁcult to solve using the methods based on approximate theories, to our knowledge, no one has reported to

this kind of problem before. In Table II the results of reﬂected power , radiated power , and their total are presented for the case of , and . As can be seen, the results satisfy the energy conservation law within an accuracy of 1% well. In Section II, mathematically and physically, it has been shown that the reﬂection coefﬁcient is independent of location of the virtual boundary . For numerical demonstration, the reﬂection coefﬁcient of a dielectric ﬁlled PPW having a tilted ﬂange surface as a function of location of the virtual boundary is plotted in Fig. 2 for , and . It is observed that the reﬂection coefﬁcient is independent of location of the virtual boundary , except at . This error is caused by the numerical method used, because when the virtual boundary approaches the aperture the segments of discretized boundary approach zero. The validity of truncation of the inﬁnite-length boundaries in the numerical solution procedure is proved numerically in Fig. 3(a) and (b). Where Fig. 3(a) shows distribution of the disturbed ﬁeld on boundary , and Fig. 3(b) shows distribution of the total ﬁeld on boundary . The parameters used in calculations are the same as for Fig. 2, the virtual boundary is located at .

CHIEN et al.: RADIATION PROPERTIES OF AN ARBITRARILY FLANGED PPW

1997

Fig. 8. Numerical results of a dielectric ﬁlled PPW having an up-tapered aperture for d= = 0:5001 and n = 1:6. (a) Dependence of radiated power 0 on tapering width w=d for h=d = 0:5. (b) Typical radiation patterns of (a)-case. (c) Dependence of radiated power 0 on tapering height h=d for w=d = 2. (d) Typical radiation patterns of (c)-case.

From Fig. 3, it is found that the use of BEM based on GMEIEs is possible to treat waveguide discontinuity problems as an isolated object of ﬁnite size. So that it is suitable for the basic theory of computer-aided design (CAD) software for waveguide circuits. B. Examples In the ﬁrst sequence of examples we consider the conventional problem of a dielectric ﬁlled PPW having a perpendicular ﬂange surface. The result of radiated power as a function of refractive index is shown in Fig. 4 for . From Fig. 4, it is evident improvement of radiated power of a dielectric ﬁlled PPW compared with a dielectric unﬁlled PPW. Since most of the solid dielectrics have the index larger than approximate 1.4, we choose the dielectric with index of 1.6 for the next investigations. In subsequent examples we apply the method to a number of cases of arbitrarily ﬂanged PPW as shown in Fig. 1(a)–(c). The results of computer simulations are shown below. 1) PPW Having a Tilted Flange Surface [Fig. 1(a)]: For a dielectric ﬁlled PPW of width having a tilted ﬂange surface, the dependence of radiated power on the tilting angle is shown in Fig. 5(a), and the typical radiation patterns are plotted in Fig. 5(b).

The results in Fig. 5 show that the radiated power of a dielectric ﬁlled PPW can be improved by the use of a tilted ﬂange surface, and the symmetry of radiation pattern is maintained even though changing the tilting angle. In particular, from Fig. 5(b) it is found that the angle of beam center (i.e., the angle of center of radiation pattern) with respect to the -axis is changed with changing the tilting angle. In order to see the relationship between the angle of beam center and the tilting angle , we numerically plot that relationship in Fig. 6. It is observed that for the tilting angle less than 15 the above-mentioned relationship is linear, i.e., the center axis of beam is perpendicular to the ﬂange surface, but for the tilting angle larger than 15 that relationship is nonlinear. This effect may be important in the prediction of radiation properties from antennas. 2) PPW Having a Tapered Flange Surface [Fig. 1(b)]: For the case of a dielectric ﬁlled PPW having a tapered ﬂange surface with and , the dependence of radiated power on the tapering angle is shown in Fig. 7(a), and the typical radiation patterns are plotted in Fig. 7(b). In Fig. 7(a), the weak effect of change of tapering angle on radiated power is observed. But on the contrary, the strong effect of that on radiation pattern is found from Fig. 7(b). It seems obvious that the beam width decreases and the far-ﬁeld intensity increases with up-tapering the ﬂange surface. This is an

1998

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

interesting and important result for millimeter-wave free-space communication systems. 3) PPW Having an Up-Tapered Aperture [Fig. 1(c)]: The radiated power as functions of tapering width and tapering height , as shown in Fig. 1(c), are respectively shown . Typically, the in Fig. 8(a) and (c) for corresponding radiation patterns are shown in Fig. 8(b) and (d). is Notice that in Fig. 8(a) and (b) the tapering height given by , and in Fig. 8(c) and (d) the tapering width is given by . From Fig. 8, it is found that the radiated power of a dielectric ﬁlled PPW can be improved signiﬁcantly by using an up-tapered aperture. In particular, as shown in Fig. 8(c), the radiated power is improved to approximate 0.99 , i.e., only 1% of by a tapering height power is reﬂected. The strong effect on radiation patterns is also found by changing the tapering parameters. However, it is observed from Fig. 8(d) that the number of lobe of radiation pattern is more than one, and the far-ﬁeld intensity ﬂuctuates with increasing the size of aperture. Let us note that these results, which are very interesting and potentially important in design of antennas, have not shown by any researcher so far. IV. CONCLUSION The radiation properties of a dielectric ﬁlled and unﬁlled PPW having an arbitrarily ﬂanged surface have been studied by the BEM based on GMEIEs. Based on the theory developed in Section II, the typical numerical evaluations have been perguided-mode wave. The formed for the case of incident numerical results were conﬁrmed by using the law of energy conservation. It has been found that the numerical results are in good agreement with previous results and physical consideration. It is apparent that the method in this paper is suitable for the basic theory of CAD)software for the antennas systems. Since we do not employ any approximations, such as simple end-shape, in the formulation of GMEIEs used in this paper, so that it is easy to extend the GMEIEs to more complicated waveguide circuits that have more than one port, etc. REFERENCES

[1] H. Shiomi and S. Yamamoto, “Numerical simulation of fat dielectric loaded waveguide antenna using FDTD method,” in IEICE Proc. Int. Symp. Antennas and Propagation ISAP i-02, Nov. 2002, pp. 520–523. [2] R. C. Rudduck and D. C. F. Wu, “Slope diffraction analysis of TEM parallel-plate guide radiation patterns,” IEEE Trans. Antennas Propagat., vol. AP-17, pp. 797–799, Nov. 1969. [3] D. C. F. Wu, R. C. Rudduck, and E. L. Pelton, “Application of a surface integration technique to parallel-plate waveguide radiation-pattern analysis,” IEEE Trans. Antennas a Propagat., vol. AP-17, pp. 280–285, May 1969. [4] S. W. Lee, “Ray theory of diffraction by open-ended waveguide, I, ﬁeld in waveguides,” J. Math. Phys., vol. 11, pp. 2830–2850, 1970. [5] K. Hongo, “Diffraction by a ﬂanged parallel-plate waveguide,” Radio Sci., vol. 7, pp. 955–963, Oct. 1972. [6] T. Itoh and R. Mittra, “TEM reﬂection from a ﬂanged and dielectricﬁlled parallel-plate waveguide,” Radio Sci., vol. 9, pp. 849–855, Oct. 1974. [7] K. Hongo, Y. Ogawa, T. Itoh, and K. Ogusu, “Field distribution in a ﬂanged parallel-plate waveguide,” IEEE Trans. Antennas Propagat., vol. AP-23, pp. 558–560, July 1975. [8] S. Lee and L. Grun, “Radiation from ﬂanged waveguide: Comparison of solutions,” IEEE Trans. Antennas Propagat., vol. AP-30, pp. 147–148, Jan. 1982.

[9] M. S. Leong, P. S. Kooi, and XQXQXQ Chandra, “Radiation from a ﬂanged parallel-plate waveguide: Solution by moment method with inclusion of edge condition,” in Proc. Inst. Elect. Eng. Microwaves, Antenna and Propagation, vol. 135, Aug. 1988, pp. 249–255. [10] C. M. Butler, C. C. Courtney, P. D. Mannikko, and J. W. Silvestro, “Flanged parallel-plate waveguide coupled to a conducting cylinder,” in Proc. Inst. Elect. Eng. Microwaves, Antenna and Propagation, vol. 138, Dec. 1991, pp. 549–558. [11] C. H. Kim, H. J. Eom, and T. J. Park, “A series solution for TM-mode radiation from a ﬂanged parallel-plate waveguide,” IEEE Trans. Antennas Propagat., vol. 41, pp. 1469–1471, Oct. 1993. [12] T. J. Park and H. J. Eom, “Analytic solution for TE-mode radiation from a ﬂanged parallel-plate waveguide,” in Proc. Inst. Elect. Eng. Microwaves, Antenna and Propagation, vol. 140, Oct. 1993, pp. 387–389. [13] J. W. Lee, H. J. Eom, and J. H. Lee, “TM-wave radiation from ﬂanged parallel plate into dielectric slab,” in Proc. Inst. Elect. Eng. Microwaves, Antenna and Propagation, vol. 143, June 1996, pp. 207–210. [14] M. Tanaka and K. Tanaka, “Computer simulation for two-dimensional near-ﬁeld optics with use of a metal-coated dielectric probe,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 18, pp. 919–925, Apr. 2001. [15] D. N. Chien, M. Tanaka, and K. Tanaka, “Numerical simulation of an arbitrarily ended asymmetrical slab waveguide by guided-mode extracted integral equations,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 19, pp. 1649–1657, Aug. 2002. [16] D. N. Chien, K. Tanaka, and M. Tanaka, “Accurate analysis of power coupling between two arbitrarily ended dielectric slab waveguides by boundary-element method,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 20, pp. 1608–1616, Aug. 2003. , “Optimum design of power coupling between two dielectric slab [17] waveguides by the boundary-element method based on guided-mode extracted integration equation,” IEICE Trans. Electron., vol. E86-C, Nov. 2003. [18] , “Guided wave equivalents of Snell’s and Brewster’s Laws,” Opt. Commun., vol. 225, pp. 319–329, Oct. 2003.

Dao Ngoc Chien (S’03) received the B.E. degree from the Department of Telecommunication Systems, Faculty of Electronics and Telecommunications, Hanoi University of Technology, Hanoi, Vietnam, in 1997 and the M.S. degree from the Department of Electronics and Computer Engineering, Gifu University, Gifu, Japan, in 2002, where he is currently working toward the Ph.D. degree. In 1997, he became a Teaching Assistant in the Department of Telecommunication Systems, Faculty of Electronics and Telecommunications, Hanoi University of Technology. He is currently on leave from Hanoi University of Technology and is a Visiting Researcher in the Department of Electronics and Computer Engineering, Gifu University. His current research interests are the CAD of optical waveguide circuits, and waveguide technology for antennas and feeds. Mr. Chien is a Student Member of the Optical Society of America (OSA), Washington, DC, and the Institute of Electrical, Information and Communication Engineers (IEICE), Japan. He was awarded a Yoneyama Scholarship by the Rotary Yoneyama Memorial Foundation, Inc., Japan.

Kazuo Tanaka (M’75) received the B.E., M.S., and Ph.D., degrees from the Department of Communications Engineering, Osaka University, Osaka, Japan, in 1970, 1972, and 1975, respectively. In 1975, he became a Research Associate in the Department of Electrical Engineering, Gifu University, Gifu, Japan, where he became an Associate Professor in 1985 and a Professor in 1990. His research since 1970 has been a general-relativistic electromagnetic theory and application, radiographic image processing and computational electromagnetic and he is currently interested in the CAD of integral optical circuits, near-ﬁeld optical circuits and simulation of Anderson localization hypothesis of ball-lightning. He was a Visiting Professor at the University of Toronto, ON, Canada, in 1994. Dr. Tanaka was awarded the Uchida Paper Award by the Japan Society of Medical Imaging and Information Science. He was a Chair of the Technical Group of Electromagnetic Theory of the Institute of Electrical, Information and Communication Engineers (IEICE), Japan.

CHIEN et al.: RADIATION PROPERTIES OF AN ARBITRARILY FLANGED PPW

1999

Masahiro Tanaka (M’00) received the B.E. and M.S. degrees from the Department of Electrical and Computer Engineering, Gifu University, Gifu, Japan, in 1992 and 1994, respectively, and the Ph.D. degree from the Department of Communication Engineering, Osaka University, Osaka, Japan, in 2002. He was a Research Associate at Tokoha-Gakuen Hamamatsu University, Japan, from 1994 to 1996. He joined the Department of Electrical and Computer Engineering, Gifu University, as a Research Assistant in 1996. He was a Visiting Researcher at the Department of Electrical and Computer Engineering, The University of Arizona, Tempe, from 1997 to 1998. His research interests are the CAD of optical waveguide circuits and near-ﬁeld optical circuits.

2000

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Scan Blindness Free Phased Array Design Using PBG Materials
**

Lijun Zhang, Jesus A. Castaneda, and Nicolaos G. Alexopoulos, Fellow, IEEE

Abstract—Scan blindness occurs for phased arrays when propagation constants of Floquet modes (space harmonics) coincide with those of surface waves supported by the array structure. In this paper, we studied the possibility of using photonic band-gap (PBG) substrate to eliminate scan blindness. A specially designed printed PBG substrate can suppress surface wave propagation inside its bandgap range, therefore it can be used to eliminate scan blindness. In this paper, we presented a method of moments (MoM) analysis of the scan properties of dipole arrays on PBG substrates with ominidirectional bandgap(s). We found that scan blindness is completely eliminated. The elimination of scan blindness makes PBG materials very attractive in phased array design. Index Terms—Phased array, photonic band-gap (PBG) substrate, scan blindness.

I. INTRODUCTION

S

CAN blindness for phased arrays can be traced to the forced surface waves by phase matching with those of the Floquet modes (space harmonics). This is common for printed arrays on dielectric substrates, phased arrays with radomes, etc. [1]–[3]. Scan blindness limits the scan range and lowers the antenna efﬁciency, therefore it must be considered in phased array design. Many efforts have been devoted to eliminate scan blindness, for example, the subarray technique is used to suppress scan blindness but at the expense of a larger unit cell size, which causes an increase in power loss to the grating lobes [3]. But the idea of using subarray to perturb the phase progression of surface waves in the substrate is quite an inspring idea. In [4] the authors studied the scan properties of antennas on perturbed inhomogeneous substrates, and they found that the inhomogeneous substrate can mitigate the scan blindness for proper designed substrate media. Scan properties of phased arrays on ferrite substrates have been investigated in [5], the scan properties of dipole arrays in a two-layer structure have been studied in [6], [7]. It is found that in a two layer structure it is possible to select the parameters to prevent the excitation of any surface waves, therefore to eliminate scan blindness. This however may be a very narrow band operation.

In this paper photonic band-gap (PBG) materials are used as antenna substrates to treat the scan blindness problem. PBG materials are essentially periodic structures whose dispersive properties may be controlled by the periodicity and the electromagnetic properties of lattice elements [8]. For PBG materials realized on dielectric substrates, surface waves can be controlled. For example, they can be suppressed along certain direction for PBG materials with partial bandgaps or along any direction for those with ominidirectional (or complete) bandgaps [9]–[12] inside a certain frequency range. Such properties of PBG could be used to eliminate scan blindness. In [13] the authors studied the active array pattern of phased arrays on a PBG substrate which is composed of air holes inside a dielectric substrate, but since that kind of material with ﬁnite thickness does not have an ominidirectional bandgap, the scan blindness cannot be eliminated completely. Recently, a novel printed PBG substrate with ominidirectional bandgap was presented in [9], [10]. In this PBG material, periodic metallic patches are printed on a substrate and each patch is connected to the ground plane through a via. It is both experimentally and numerically veriﬁed that this PBG material has a complete surface wave bandgap. In [11] a planar PBG without any via was also fabricated and an ominidirectional bandgap was reported. In this paper our analysis is based on the ﬁrst kind of PBG substrate in [9], [10], however we believe that any PBG substrate with complete bandgap can eliminate scan blindness. In this paper, we ﬁrst present the theorectical method of moments (MoM) formulation for the array analysis, then we discuss the surface-wave bandgap properties of the PBG materials. After that we present detailed theoretical case studies for phased arrays on PBG substrates, from low permittivity to high permittivity substrates, from thin to thick substrates. Finally, waveguide simulator experimental results are presented to validate the theory and the code we developed. II. NUMERICAL ANALYSIS: THE METHOD OF MOMENTS In Fig. 1, we show the PBG material [9] and a dipole phased . array printed on it. The unit cell of the PBG is around For phased arrays the radiating dipole is printed in every by PBG unit cell, here in the ﬁgure the dipole is printed in every 2-by-2 PBG unit cells as an example. Printed arrays on uniform substrate and their scan blindness phenomenon have been extensively analyzed using a MoM in [1], [2]. The MoM is a very fast and accurate full-wave analysis method for the analysis of phased array and therefore is used in

Manuscript received May 4, 2001; revised March 26, 2003. This work was supported in part by MURI. L. Zhang and J. A. Castaneda were with the Department of Electrical Engineering, University of California, Los Angeles, CA 90095-1544 USA. They are now with Broadcom Corporation, Irvine, CA 92619-7013 USA (e-mail: lzhang@broadcom.com; jcastan@broadcom.com). N. G. Alexopoulos is with the Department of Electrical and Computer Engineering, University of California, Irvine, CA 92695-2625 USA (e-mail: alﬁos@uci.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832516

0018-926X/04$20.00 © 2004 IEEE

ZHANG et al.: SCAN BLINDNESS FREE PHASED ARRAY DESIGN USING PBG MATERIALS

2001

A. Current Expansion From Attachment Modes For the speciﬁc PBG structure, it is more advantageous to use the entire domain basis (EDB) function and the attachment mode expansion technique in the MoM solution to achieve fast convergence [16]. Since in this paper we are also interested in thick substrates, we used more than one attachment mode. The current expansion for each attachment mode is obtained through the solving of a Sturm-Liouville problem with a line current source excitation inside an equivalent cavity. Current distributions for attachment modes on each patch-via in the unit cell of PBG is given in the following equations, (4) where

Fig. 1. Structure of the phased dipole array on a PBG substrate. The top ﬁgure shows 3-by-3 unit cells of the PBG substrate. The bottom ﬁgure gives the top view of a unit cell of the inﬁnite phased dipole array printed on the PBG susbstrate. Some key parameters are: patch size L by W , gap between patches g , via size x by y , dipole size D (length) by w (width), substrate thickness h and substrate permittivity .

(5) (6) (7) stands for the number of attachment modes, where and are the center coordinates of each patch

4

4

4

our analysis. The Dyadic Green’s function, which is the electric ﬁeld caused by an inﬁnite array of such dipoles is given by (1) where and are the Floquet’s propagation constants determined by the scan angle, and are the array period, is the free space impedance. The dyadic quantity is deﬁned in the following [15]

(8)

(9) (10) (also see (11) shown at the bottom of the page), and similarly for . B. Entire Domain Current Expansion on the Dipole and Patch Current expansions on each patch are simply entire domain basis functions given as follows:

(2) On metallic surfaces tangential electric ﬁelds should be zero

(3) where is the tangential dyadic and is the current expansion which will be discussed in following parts.

(12)

for (11) for .

2002

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

(13) (14) Current expansion on the dipole follows:

(15) is the number of entire domain basis functions where used in the dipole current expansion. C. Final Matrix Equation After weighting on (3), the following matrix equation is obtained: (16) The total matrix size is , with the number of PBG patches in each unite cell of the array. Each submatrix entry results from the current basis function from structure type and the weighting function from structure type , with , from , , standing for attachment mode, entire domain basis (EDB) function on patch, or EDB on dipole, respectively. The matrix entries are given as (17) (18) (19) (20) (21) (22) (23) (24) (25) Notice that , are local indexes for submatrices. For detailed formulation of the matrix entries please refer to [18], they are omitted here for brevity. The input impedance of the dipole is calculated as (26) (27) where is the width of voltage gap source, and width of the dipole. is the

Fig. 3. Convergence of the input impedance versus the number of Floquet modes and the number of EDB functions. Impedances are for broadside scan.

leny

Fig. 2.

Surface wave bandgap of the PBG substrate. h = 2:3855 mm, lenx = = 2:88 mm, W = L = 2:061 mm, g = 0:819 mm, = 4:4.

III. SURFACE WAVE BANDGAP OF PRINTED PBG SUBSTRATES The MoM procedure discussed in the previous section can also be used to calculate the bandgap properties of printed PBG

materials. Here we present some simulation data of the printed PBG material calculated using three different methods which are the MoM, the ﬁnite-element-integral-equation method (FEIEM) [17] and the FE perfect-matched-layer method (FE-PML) [14]. Fig. 2, shows the eigenmodes of the PBG material with paare rameters given in the caption of the Figure. , , and vertices of a reduced Brillouin zone. From to , the waves varies from 0 to . are propagating along the direction, From to , the waves are propagating at an angle between and 0 and 45 with respect to the direction, where varies from 0 to . From to , waves are propagating and varies from or to 0. along the 45 and both The MoM and FE based codes produce consistent results. The MoM code only searches for bounded modes, therefore, it stops when it intercepts the light line in the air (straight line in this ﬁgure) for the second mode. Between the ﬁrst and second mode is the surface wave bandgap. The higher band edge is determined by the interception point of the second mode and the light line. Simulation suggests that the second mode is

ZHANG et al.: SCAN BLINDNESS FREE PHASED ARRAY DESIGN USING PBG MATERIALS

2003

Fig. 4. Comparison of the scanned impedances for the dipole array on PBG and uniform substrates. FREQ = 13:0 GHz. = 2:2. For PBG case dipole size is 8.982 mm 0.06 mm, broadside impedance is Z = 71:083 j 0:032

. Array unit cell is 0:4992 . For uniform case the dipole size is tuned to have a resonance at broadside.

2

0

radiating inside the bandgap. In the FE-PML calculation, the eigen frequencies are complex inside the bandgap region for given real propagation constants, which implies that modes inside this bandgap region are damping in the lateral direction. The leakage properties of these modes need further investigation, and some similar work has been done in [12] on dielectric PBGs. The good agreement between the MoM and FEM based codes veriﬁes the validity of the MoM analysis. IV. NUMERICAL RESULTS ON PHASED ARRAY In the ﬁrst example of phased array, one unit cell is made of one dipole in the middle of two by two PBG unit cells. Dipoles mm in the rectangular array are separated by in and directions, respectively. The substrate thickness is mm, , mm, mm, mm. The surface wave bandgap is between 9.55 and 14.10 GHz. Fig. 3 shows convergence plot of the input impedance versus the number of Floquet modes and the number of entire domain axis, is basis (EDB) functions for patch current. In the the highest index of the Floquet mode, which means that the . It is seen that the total number of Floquet modes is impedance converges as the number of the Floquet modes and the EDB functions are increased. In our following calculation, for the Floquet modes and 16 EDB functions. we use The reﬂection of the array at two different frequencies are plotted in Figs. 4 and 5. In Fig. 4, the array operates at 13.0 GHz. For the case with uniform substrate scan blindness occurs at 69.0 in the -plane, there is no blindness in the -plane because of polarization mismatch. For the PBG case, there is no blindness spot. In Fig. 5 the operating frequency is at 13.5 GHz. Since the array unit cell is greater than , a grating lobe

occurs for both the uniform and PBG substrate. For the array on the uniform substrate, the scan blindness spot moves toward broadside comparing to the 13.0 GHz case. There is no blindness spot for the array on the PBG substrate. In the next example, a substrate with higher dielectric constant is picked. The PBG lattice sizes are , patch width mm, mm, . One array unit is made of one dipole in the middle of four-by-four PBG units. The surface wave bandgap is from 9.25 to 12.85 GHz. Fig. 6 shows the scanned impedance at 12.8 GHz. For the case on uniform substrate, the blindness spot moves toward the broadside when the permittivity is higher. Again, no blindness spot is observed for the PBG case. Compact arrays can be realized on high permittivity substrates, however strong surface waves excited in the substrate cause problems such as small scan range, strong mutual coupling between elements, low efﬁciency, etc. Use of PBG substrate can avoid the formentioned drawbacks. In the third example, an array on a thick substrate is analyzed. Thicker substrate provides wider bandwidth. A unit of the array is made of one dipole in the middle of four by four PBG unit mm, , cells. The substrate thickness is mm, with the PBG unit period of mm and mm. The bandgap is from 9.7 to 15.1 GHz. Fig. 7 shows the scanned impedance. Again no scan blindness exists for the array on the PBG substrate. V. PBG SUBSTRATE DESIGN PROCEDURES In practical phased array design, the most important issue is how to design the PBG substrate. For example, given the substrate thickness ( ), permittivity ( ), array unit cell size ( ),

2004

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 5.

j :

0 085

. Array unit cell is 0:5184

Scan impedance of the dipole array on PBG substrate. FREQ .

= 13:5 GHz, dipole 8.668 mm 2 0.06 mm, broadside impedance is Z = 62:635 +

**Fig. 6. Scan impedance of the dipole array on PBG and uniform substrates. impedance is Z = 50:30 j 0:08
**

. Array unit cell is 0:49152 .

0

**= 4:4. FREQ = 12:8 GHz. For PBG case, dipole length is 7.14 mm, broadside
**

In the array design, one also needs to consider the effects of the feeding and the radiating element loading.

how to design the PBG substrate to meet the bandgap and band). The suggested steps are listed in width requirement ( the following:

(1) Design a reference PBG substrate, with given h, r , a; b, the bandgap is denoted as fref ; fref , (2) Scale the unit cell size a; b to make the ~ new bandgap to be f ; ~f , which is close to the f, aiming frequency f; (3) Fine tune the gap between the patches, (4) Check the loading effect. Go back to step 3 if necessary.

VI. EXPERIMENTS A large phased array is always an extremely costly piece of hardware, and unfortunately some phenomena are related to size speciﬁcally. A commonly used method to examine the scan properties of phased arrays is called waveguide simulator, which provides a compact and inexpensive test piece for phased arrays. For references on the theory of waveguide simulator, please refer to [19], [20].

4

4 4

ZHANG et al.: SCAN BLINDNESS FREE PHASED ARRAY DESIGN USING PBG MATERIALS

2005

Fig. 7.

j :

0 15

.

Scanned impedance of a dipole array on thick PBG substrate. FREQ = 13:0 GHz, Dipole length is 9.766 mm, broadside impedance is Z

= 82 43 0

:

The experimental setup is explained in Fig. 8[18]. An S-band WR284 rectangular waveguide is used, the inner dimensions are ) band is 2.84 inches by 1.34 inches. The single mode ( between 2.60 and 3.95 GHz. The PBG substrate is fabricated and then press ﬁtted into the waveguide, a shorting plate is added at the bottom of the substrate ground plane. The coax is fed through two holes drilled from the waveguide broadside wall. The substrate is put into a shallow waveguide section so that we can access the monopole to solder it to the inner conductor of the coax excitation. Another long waveguide section is put above the shallow section to allow the excited high order modes be attenuated, and ﬁnally a matched load is put at the end to simulate the wave propagation in free space. The waveguide simulator models the phased array with an -plane scan angle governed by the following equation: (28)

Fig. 8. Waveguide simulator experimental setup.

with which is the broadside dimension of the wavethe free space wavelength. For frequencies from guide and 2.6 to 3.9 GHz, the simulated scan range is from 51.22 to 31.32 . In the experiment, a power divider was used and the two outputs are connected to two coax cables used to excite the monopoles. The return loss is measured at the input of the power divider, and then de-embedded to obtain the reﬂection coefﬁcient at the feeding point of the monopole. The substrate is a , the thickness is RT 5870 Duriod board, with 390 mil. In the ﬁrst case we fabricated a PBG substrate using one layer of this board, the metal patch size is 8.9 mm by 10.5 mm, with a period of 17.018 mm by 18.034 mm. By mirroring, the equivalent dipole is between every two by four PBG unit cells. The monopole length is 16.32 mm. The experimental and MoM

simulation data are plotted in Fig. 9, where consistency can be observed between the two. In the second case, we fabricated a PBG material with double substrate thickness, the surface wave bandgap is between 1.9 and 3.9 GHz. The patch size is 6.5 mm by 7.5 mm. The monopole length is 21.0 mm. Results are shown in Fig. 10. For the uniform substrate, scan blindness occurs at 3.7 GHz, which is corresponding to an -plane scan angle of 34 from the broadside. This agrees with the scan angle predicted by the following [1]: (29) where is the surface wave propagation constant, mm is the period of the PBG in broadside. The nonunit

2006

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

ACKNOWLEDGMENT The authors would like to acknowledge Prof. F. De Flaviis and Mr. R. Ramirez from the University of California at Irvine for their kind help in the waveguide experiments. The authors would like to thank Prof. D. Yang from the University of Illinois at Chicago for helpful suggestions on the use of attachment modes in the MoM solution. REFERENCES

[1] D. M. Pozar and D. H. Schaubert, “Scan blindness in inﬁnite phased arrays of printed dipoles,” IEEE Trans. Antennas Propagat., vol. 32, pp. 602–610, June 1984. [2] , “Analysis of an inﬁnite array of rectangular microstrip patches with idealized probe feeds,” IEEE Trans. Antennas Propagat., vol. 32, pp. 1101–1107, Oct. 1984. [3] D. M. Pozar, “Scanning characteristics of inﬁnite arrays of printed antenna subarrays,” IEEE Trans. Antennas Propagat., vol. 40, pp. 666–674, June 1992. [4] W. J. Tsay and D. M. Pozar, “Radiation and scattering from inﬁnite periodic printed antennas with inhomogeneous media,” IEEE Trans. Antennas Propagat., vol. 46, pp. 1641–1650, Nov. 1998. [5] H. Y. Yang and J. A. Castaneda, “Inﬁnite phased arrays of microstrip antennas on generalized anisotropic substrates,” Electromagn., vol. 11, no. 1, pp. 107–124, Jan.-Mar. 1991. [6] J. A. Castaneda, “Inﬁnite phased array of microstrip dipoles in two layers,” Ph.D. dissertation, University of California, Los Angeles, 1988. [7] J. Castaneda and N. G. Alexopoulos, “Inﬁnite arrays of microstrip dipoles with a superstrate (cover) layer,” in Proc. Antennas and Propagation Int. Symp., vol. 2, 1985, pp. 713–717. [8] E. Yablonovitch, “Photonic band-gap structures,” J. Opt. Soc. Amer. B, vol. 10, no. 2, pp. 283–295, Feb. 1993. [9] D. Sievenpiper, L. Zhang, R. F. J. Broas, N. G. Alexopoulos, and E. Yablonovitch, “High-impedance electromagnetic surfaces with a forbidden frequency band,” IEEE Trans. Microwave Theory and Techniques, vol. 47, pp. 2059–2074, Nov. 1999. [10] D. Sievenpiper, “High-impedance electromagnetic surfaces,” Ph.D. dissertation, Univ. California, Los Angeles, 1999. [11] F. R. Yang, K. P. Ma, Y. Qian, and T. Itoh, “A novel TEM waveguide using uniplanar compact photonic-bandgap (UC-PBG) structure,” IEEE Trans. Microwave Theory and Techniques, vol. 47, pp. 2092–2098, Nov. 1999. [12] H. Y. Yang, “Characteristics of guided and leaky waves on multilayer thin-ﬁlm structures with planar material gratings,” IEEE Trans. Microwave Theory and Techniques, vol. 45, pp. 428–435, Mar. 1997. [13] P. K. Kelly, L. Diaz, M. Piket-May, and I. Rumsey, “Investigation of scan blindness mitigation using photonic bandgap structure in phased arrays,” in Proc. SPIE, vol. 3464, July 1999, pp. 239–248. [14] L. Zhang, N. G. Alexopoulos, D. Sievenpiper, and E. Yablonovitch, “An efﬁcient ﬁnite-element method for the analysis of photonic band-gap materials,” in Proc. IEEE Int. Microwave Symp. Dig., vol. 4, 1999, pp. 1703–1706. [15] I. E. Rana and N. G. Alexopoulos, “Current distribution and input impedance of printed dipoles,” IEEE Trans. Antennas Propagat., vol. 29, pp. 99–105, 1981. [16] J. T. Aberle and D. M. Pozar, “Analysis of inﬁnite arrays of probe-fed rectangular microstrip patches using a rigorous feed model,” IEE Proc., pt. H, vol. 136, no. 2, pp. 109–119, Apr. 1989. [17] L. Zhang and N. G. Alexopoulos, “Finite-element based techniques for the modeling of PBG materials,” Electromagn., Special Issue on Theory and Applications of Photonic Band-Gap Materials, vol. 19, pp. 225–239, May-June 1999. [18] L. Zhang, “Numerical characterization of electromagnetic band-gap materials and applications in printed antennas and arrays,” Ph.D. dissertation, Univ. Calif., Los Angeles, 2000. [19] P. W. Hannan and M. A. Balfour, “Simulation of a phase-array antenna in waveguide,” IEEE Trans. Antennas Propagat., vol. 13, pp. 342–353, May 1965. [20] N. Amitay, V. Galindo, and C. P. Wu, Theory and Analysis of Phased Array Antennas. New York: Wiley-Interscience, 1972, pp. 59–63.

Fig. 9.

Waveguide simulator results for dipole array on PBG substrate.

Fig. 10. Waveguide simulator results for dipole array on PBG and corresponding uniform substrate with substrate thickness twice those in Fig. 9.

reﬂection at scan blindness point is due to the loss associated with the waveguide simulator set up and possibly loss in the susbtrate. Mismatch between the simulation and measurement for the PBG case may be due to the nonlongitudinal currents excited on the patches and vias which affects the waveguide simulator accuracy. For a resonance frequency at 3.25 GHz, the . The good agreement veriﬁes that substrate thickness is the MoM analysis is also valid for thick substrates. VII. CONCLUSION Scan properties of phased arrays on PBG substrates have been investigated. Through the example of printed dipoles on PBG substrates, it is found in both simulation and experiment that scan blindness can be completely eliminated, due to the suppression of surface wave propagation inside PBG substrates. PBG substrates, especially those with complete surface wave bandgaps, will ﬁnd extensive applications in printed antennas and arrays.

ZHANG et al.: SCAN BLINDNESS FREE PHASED ARRAY DESIGN USING PBG MATERIALS

2007

Lijun Zhang received the B.S. and M.S. degrees in electrical engineering from University of Science and Technology of China, in 1993 and 1996, respectively, and the Ph.D. degree in electrical engineering from the Department of Electrical Engineering, University of California, Los Angeles, in 2000. From June 2000 to December 2000, he was with Agilent Technology in West Lake Village, CA, working on RF-CMOS CAD. Since December 2000, he has been working for Broadcom Corporation, Irvine, CA, in the areas of on chip passives modeling and wireless radio transceiver design.

Jesus A. Castaneda received the B.S. degree in physics from Saint Mary’s College, Moraga, CA, in 1970, and the M.S.E.E. and Ph.D.E.E. degrees from the Department of Electrical Engineering, University of California, Los Angeles (UCLA), in 1978, 1981, and 1988, respectively. From 1978 to 1985, he was with the Antenna Department, Radar Systems Group, Hughes Aircraft Company, working in the area of microwave antenna design and analysis, including electronically scanned antenna arrays. Other work areas included planar arrays, frequency selective surfaces, radomes, and adaptive arrays. From 1986 to 1997, he was with Phraxos Research and Development, Inc. as Senior Research Engineer and Engineering Manager with responsibility for the technical management of projects in the area of electromagnetic modeling for microwave and millimeter wave applications. From 1995 to 2000, he was a Senior Lecturer at the School of Engineering and Applied Science, UCLA. Since 2000, he has been with Broadcom Corporation, Irvine, CA, as a Senior Principal Scientist working in the areas of antennas for wireless systems and on-chip passives design.

Nicoloas G. Alexopoulos (S’68–M’69–SM’82–F’87) received the B.S.E.E., M.S.E.E., and Ph.D.E.E. degrees from the University of Michigan, Ann Arbor, in 1965, 1967, and 1968, respectively. He was a member of the faculty in the Department of Electrical Engineering, University of California, Los Angeles (UCLA), from 1969 to 1996. While at UCLA, he served as Associate Dean for Faculty Affairs from 1986 to 1987, and Chair of the Electrical Engineering Department from 1987 to 1992. Under his leadership and tenure as Chair the department doubled in size, created a highly successful Corporate Afﬁliates Program, raised more than $30 million in gifts and endowments and established the High Frequency Electronics Laboratory. In 1997, he joined the Electrical Engineering and Computer Science Department, University of California, Irvine, and has been Dean of The Henry Samueli School of Engineering. As Dean he led the efforts to establish The Integrated Nanosystems Research Facility, The Biomedical Engineering Department, The Center for Pervasive Communications and Computing, The California Institute for Telecommunications and Information Technology, supported the establishment of The National Fuel Cell Research Center and initiated The Arts, Computing and Engineering Program. In addition, he is the Principal Investigator of the University of California Irvine Mathematics Engineering Science Achievement Program (MESA) and Co-PI of The National Science Foundation UC Systemwide California Alliance for Minority Participation (CAMP). His research contributions over time include the ﬁrst contributions in the interaction of electromagnetic waves with active surfaces and particles in the early 1970s. He was the ﬁrst to deﬁne and publish on Active and Passive Magnetic Walls (PBG and EBG structures) and their realization with artiﬁcial periodic structures and speciﬁcally arrays of antennas terminated at variable load impedances. He demonstrated how such surfaces can be used for beam scanning, and radar cross section elimination or enhancement. Subsequently he focused on developing, with his students, a single full wave theory for the simultaneous design of microstrip circuits and printed antennas, thus taking into account all wave phenomena and mutual interactions. This work also led to the study of substrate-superstrate effects and anisotropic and gyrotropic substrate materials. This body of research contributed signiﬁcant progress in the use of the MoMs for the development of useful design algorithms for microstrip antennas and circuits. More recently, he and his students focused in the development of percolation theory and its applications in materials and wave propagation in complex media, as well as the design of artiﬁcial materials. Presently, he is working on the integration of the above mentioned experience in the research of electromagnetically metamorphic objects and interfaces. He has more than 250 publications and lectures on a variety of subjects including a popular lecture on “The Genesis and Destruction of The First Research University; The Library/Museum of Alexandria.” Dr. Alexopoulos was the corecipient (with his students) of the Schelkunoff Award in 1985 and 1998.

2008

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Fractile Arrays: A New Class of Tiled Arrays With Fractal Boundaries
**

Douglas H. Werner, Senior Member, IEEE, Waroth Kuhirun, and Pingjuan L. Werner, Senior Member, IEEE

Abstract—In this paper, a new class of antenna arrays are introduced, which we call fractile arrays. A fractile array is deﬁned as any array with a fractal boundary contour that tiles the plane without gaps or overlaps. It will be shown that the unique geometrical features of fractiles may be exploited in order to make available a family of deterministic arrays that offer several highly desirable performance advantages over their conventional periodic planar array counterparts. Most notably, fractile arrays have no grating lobes even when the minimum spacing between elements is increased to at least one-wavelength. This has led to the development of a new design methodology for modular broadband low-sidelobe arrays that is based on fractal tilings. Several examples of fractile arrays will be considered including Peano–Gosper, terdragon, 6-terdragon, and fudgeﬂake arrays. Efﬁcient iterative procedures for calculating the radiation patterns of these fractile are also introduced in this arrays to arbitrary stage of growth paper. Index Terms—Fractal antennas, fractal arrays, broad-band arrays, grating lobes, low-sidelobe arrays.

I. INTRODUCTION

S

EVERAL book chapters and review articles have been published recently that deal with the subject of fractal antenna engineering [1]–[5]. A considerable amount of this literature is devoted to new concepts for antenna arrays that employ fractal geometries in their design. The ﬁrst application of fractal geometry to antenna array theory was proposed by Kim and Jaggard [3], [5], [6], where properties of random fractals were used to develop a design methodology for quasirandom arrays. These quasirandom arrays were shown to possess radiation characteristics capable of bridging the gap between those produced by completely ordered (i.e., periodic) arrays and completely disordered (i.e., random) arrays. The design of multiband and low-sidelobe linear arrays based on a Cantor fractal distribution of elements was considered in [1], [3], [5], [7]. Other properties of Cantor fractal linear arrays have been studied more recently in [2], [3], [5]. The electromagnetic radiation produced by planar concentric-ring Cantor arrays was investigated in [3], [5], [8]. These arrays were generated using polyadic Cantor bars, which are described by their similarity fractal dimension, number of gaps, and lacunarity

Manuscript received June 16, 2003; revised August 18, 2003. The authors are with the The Pennsylvania State University, Department of Electrical Engineering, University Park, PA 16802 USA (e-mail: dhw@psu.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832327

parameter. Planar fractal array conﬁgurations, based on Sierpinski carpets, were also considered in [2], [3], [5], [9]. The fact that Sierpinski carpet and related arrays can be generated recursively (i.e., via successive stages of growth starting with a simple generating array) has been exploited in order to develop rapid algorithms for use in efﬁcient radiation pattern computations and adaptive beamforming, especially for arrays with multiple stages of growth that contain a relatively large number of elements [2], [3], [5]. The Cantor linear and Sierpinski carpet planar fractal arrays have also been shown to be examples of deterministically thinned arrays [2], [3], [5]. More recently, a new type of deterministic fractal array was introduced in [10]–[12] that is based on the Peano–Gosper family of space-ﬁlling curves. The elements of the array are uniformly distributed along a Peano–Gosper curve, which leads to a planar array conﬁguration with parallelogram cells that is bounded by a closed Koch-type fractal curve. These unique properties were exploited in [10]–[12] to develop a design methodology for deterministic arrays that have no grating lobes even when the minimum spacing between elements is increased to at least one wavelength. Hence, these Peano–Gosper arrays are relatively broadband when compared to more conventional periodic planar arrays with square or rectangular cells and regular boundary contours. This type of fractal array differs fundamentally from other types of fractal array conﬁgurations that have been studied previously, such as those reported in [1]–[9], which have regular boundaries with elements distributed in a fractal pattern on the interior of the array. However, in direct contrast to this, the boundary contour of the Peano–Gosper array is fractal but the elements on the interior of the array do not follow a fractal distribution. A new category of fractal arrays, which we call fractile arrays, will be introduced in this paper. A fractile array is deﬁned to be any array which has a fractal boundary contour that tiles the plane. Tilings of the plane using fractal shaped tiles have been considered in [13]–[15]. These fractal tiles, or fractiles, represent a unique subset of all possible tile geometries that can be used to cover the plane without gaps or overlaps. Here we exploit the unique geometrical properties of fractiles to develop a new design methodology for modular broadband low-sidelobe antenna arrays. In Section II-A we demonstrate that the Peano–Gosper arrays recently considered in [10]–[12] may be classiﬁed as fractile arrays. The radiation characteristics of other types of fractile arrays will also be investigated in this paper. These include the tredragon, 6-terdragon, and fudgeﬂake fractile arrays discussed in Section II-B, Section II-C, and Section II-D, respectively.

0018-926X/04$20.00 © 2004 IEEE

WERNER et al.: FRACTILE ARRAYS: A NEW CLASS OF TILED ARRAYS WITH FRACTAL BOUNDARIES

2009

(a)

(b)

Fig. 2. First three stages in the construction of a terdragon curve. The initiator is shown in (a) as a dashed line superimposed on the stage 1 generator. The generator (unscaled) is shown again in (b) as the dashed curve superimposed on the stage 2 terdragon curve. The stage 3 terdragon curve is shown in (c).

(c) Fig. 1. Gosper island fractiles and their corresponding Peano–Gosper curves for (a) stage 1, (b) stage 2, and (c) stage 4.

II. SOME EXAMPLES OF FRACTILE ARRAYS A. The Peano–Gosper Fractile Array The radiation properties of Peano–Gosper arrays have been recently investigated in [10]–[12]. These arrays derive their name from the fact that the elements are uniformly distributed along a space-ﬁlling Peano–Gosper curve. This results in a deterministic planar array conﬁguration composed of a unique arrangement of parallelogram cells that is bounded by a variant of an irregular closed Koch fractal curve. It was shown in [10]–[12] that these arrays exhibit relatively broadband low side-lobe performance when compared to their conventional counterparts. Here, we show that Peano–Gosper arrays are in actuality a type of fractile array. In order to see this, it is convenient to start by considering the sequence of Gosper islands illustrated in Fig. 1. Also shown in Fig. 1 are the Peano–Gosper curves that ﬁll the interior of the associated Gosper islands. The array elements are assumed to be equally spaced along these Peano–Gosper curves [10]–[12]. For example, the generating array represented

by the stage 1 Peano–Gosper curve shown in Fig. 1(a) contains a total of eight elements, while the stage 2 Peano–Gosper array represented by the curve shown in Fig. 1(b) contains 50 elements. Fig. 1(b) indicates that seven stage 1 Gosper islands can be tiled together to form a stage 2 Gosper island. Likewise, seven stage 2 Gosper islands can be tiled together in a similar way to form a stage 3 Gosper island, and so on. Fig. 1(c) shows a stage 4 Gosper island (which consists of seven stage 3 Gosper islands tiled together) as well as the corresponding stage 4 Peano–Gosper curve that ﬁlls its interior. The tiling process depicted in Fig. 1 can be repeated to produce Gosper islands having any desired stage of growth. This implies that Gosper island tiles are self-similar since they may be divided into seven equal tiles that are similar to the whole [13]. Moreover, it follows that the boundary of these Gosper island tiles is represented by a type of Koch fractal curve. It is also obvious from Fig. 1 that these Gosper islands are examples of fractiles since they can be used to tile the plane. Finally, because each Gosper island has a corresponding Peano–Gosper curve that ﬁlls its interior, then we are led to the conclusion that Peano–Gosper arrays do in fact belong to the family of fractile arrays. B. The Terdragon Fractile Array In this section, we will introduce the terdragon fractile array as well as derive a useful compact product representation for the corresponding array factor. The terdragon is a member of the family of space-ﬁlling dragon curves [14]. The ﬁrst three stages in the construction of a terdragon curve are shown in Fig. 2. The initiator for the terdragon curve is indicated in Fig. 2(a) by the dashed line segment of unit length. The generator for the terdragon curve is obtained from the initiator by replacing it with a three-sided polygon as shown in Fig. 2(a), where each side has a length of . Now in order to obtain the stage construction of the terdragon curve shown in Fig. 2(b), each of (shown the three sides of the generator polygon at stage

2010

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 3. Element locations and associated current distributions for the (a) stage 1, (b) stage 3, and (c) stage 6 terdragon fractile arrays. The spacing d consecutive array elements uniformly distributed along the terdragon curve is assumed to be the same for each stage.

between

reproduced in Fig. 2(b) as the dashed curve) are replaced by an appropriately scaled, rotated, and translated copy of the entire generator. This iterative process may be repeated to generate terdragon curves up to an arbitrary stage of growth . For inconstruction of the terdragon curve is stance, the stage shown in Fig. 2(c) superimposed on a copy of the stage curve from Fig. 2(b). The geometry for a stage 1, a stage 3, and a stage 6 fractile array based on the terdragon curve are shown in Fig. 3. Fig. 3 also indicates the location of the elements in plane and their corresponding values of current amthe plitude excitation. For this example, the minimum spacing befor each tween array elements is held ﬁxed at a value of stage of growth. The nonuniform current amplitude distributions arise from the fact that the initiator consists of a uniformly excited two-element linear array with spacing between the ele. Hence, we can consider the generator ments denoted by array shown in Fig. 3(a) to be composed of three copies of the two-element initiator array appropriately rotated and translated. In this case there are two instances where two of the array elements will share a common location. From a physical point of view the two colocated elements can be interpreted as a single element having twice the value of current amplitude excitation. With this in mind the mathematical six-element uniformly excited array model can be replaced by a physically equivalent four-element array model that has a nonuniform current distribution of 1:2:2:1. This process is then repeated to generate higher-order versions of the fractile array. The array factor for a stage terdragon fractile array may be matrices conveniently expressed in terms of a product of

EXPRESSIONS OF x

AND y

IN

TABLE I TERMS OF THE PARAMETERS d

AND

Fig. 4. First stage in the construction of a 6-terdragon fractile. The initiator is shown as the dashed curve superimposed on the stage 1 generator shown as the solid curve.

WERNER et al.: FRACTILE ARRAYS: A NEW CLASS OF TILED ARRAYS WITH FRACTAL BOUNDARIES

2011

Fig. 6. 6-terdragon fractile arrays for (a) stage 4 and (b) stage 6.

Fig. 5. Element locations and associated current distributions for the (a) stage 1, (b) stage 2 and (c) stage 3 6-terdragon fractile arrays. The spacing d between consecutive array elements uniformly distributed along the 6-terdragon curve is assumed to be the same for each stage.

which are pre-multiplied by a vector a vector such that

and postmultiplied by

(1) where (2) (3) (4) (5) (6) (7)

(8) (9) (10)

Fig. 7. The ﬁrst three stages in the construction of a fudgeﬂake fractile. The initiator is shown in (a) as the dashed curve superimposed on the stage 1 generator. The generator (unscaled) is shown again in (b) as the dashed curve superimposed on the stage 2 fudgeﬂake. The stage 3 fudgeﬂake curve is shown in (c).

(12) (11) (13)

2012

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 8. Element locations and associated current distributions for the (a) stage 1, (b) stage 3, and (c) stage 5 fudgeﬂake fractile arrays. The spacing d consecutive array elements uniformly distributed along the fudgeﬂake curve is assumed to be the same for each stage.

between

Note that the parameter represents the scale factor used to generate the terdragon fractile arrays shown in Fig. 3. The values required in (8) are found from of

where

represents the empty set. Note that if then for the corresponding values of and . The values of and for – required to evaluate (9) and (10) are listed in Table I. C. The 6-Terdragon Fractile Array

(14)

Fig. 4 shows the ﬁrst stage in the construction of a fractile array that is based on six copies of the stage 1 terdragon array,

WERNER et al.: FRACTILE ARRAYS: A NEW CLASS OF TILED ARRAYS WITH FRACTAL BOUNDARIES

2013

Fig. 9. Fudgeﬂake fractile array aperture divided into three self-similar subarray apertures.

Fig. 11. Plot of the normalized stage 6 terdragon fractile array factor versus ' = . for = 90 and d

Fig. 10. Plot of the normalized stage 6 terdragon fractile array factor versus = =2 and the for ' = 90 . The dashed curve represents the case where d = . solid curve represents the case where d

Fig. 12. Plot of the normalized array factor versus with ' = 90 for a uniformly excited 18 18 periodic square array. The dashed curve represents = =2 and the solid curve represents the case where the case where d = . d

2

shown in Fig. 3, arranged in the plane around the point at the origin. Therefore, we introduce the terminology 6-terdragon array to denote a fractile array generated from the curve shown , and ) in Fig. 4. The ﬁrst three stages (i.e., in the construction of a 6-terdragon fractile array are shown in Fig. 5. Also indicated in Fig. 5 are the locations of the elements and their corresponding values of current amplitude excitation. The minimum spacing between array elements is held ﬁxed at a for each stage of growth. Fig. 6 shows the geomvalue of etry for a stage 4 and a stage 6 6-terdragon fractile array. This ﬁgure also clearly illustrates how these arrays can be considered as being composed of six associated terdragon subarrays tiled together around a common central point. The array factor for a stage 6-terdragon fractile array may be expressed in terms of the product of matrices which are premultiplied by a vector and postmultiplied by a vector such that (15) (16) (17)

where the matrices respectively.

and

are deﬁned in (2) and (6),

D. The Fudgeﬂake Fractile Array In this section, another type of fractile, known as the fudgeﬂake, is investigated for its potential utility in the design of broadband low-sidelobe antenna arrays. The ﬁrst three stages in the construction of a fudgeﬂake fractile are illustrated in Fig. 7 [14]. The initiator appears as the dashed curve (i.e., the triangle) in Fig. 7(a) superimposed on the stage 1 generator. This generator is shown again in Fig. 7(b) as the dashed curve superimposed on the stage 2 fudgeﬂake, while Fig. 7(c) shows the stage 3 fudgeﬂake with the associated generator from stage 2 superimposed. The geometry and current distributions for a stage 1, stage 3, and stage 5 fudgeﬂake fractile array located in the plane are depicted in Fig. 8. Finally, an example is presented in Fig. 9 that illustrates how a fudgeﬂake fractile array can be divided into three self-similar subarray apertures.

2014

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE II MAXIMUM DIRECTIVITY FOR SEVERAL DIFFERENT TERDRAGON FRACTILE ARRAYS

TABLE III COMPARISON OF MAXIMUM DIRECTIVITY FOR A STAGE 6 TERDRAGON FRACTILE ARRAY WITH 308 ELEMENTS AND AN 18 COMPARABLE SIZE WITH 324 ELEMENTS

2 18 SQUARE ARRAY OF

The array factor for a stage be expressed as

fudgeﬂake fractile array may

III. RESULTS Fig. 10 contains a plot of the normalized array factor (in decifor the stage 6 terdragon fractile bels) versus with array shown in Fig. 3(c). The dashed curve represents the radiation pattern slice for a terdragon fractile array with element while the solid curve represents the spacings of corresponding radiation pattern slice for the same array with . Fig. 11 shows a plot of the normalized array factor for the case where , and . This plot demonstrates that there are no grating lobes present anywhere in the azimuthal plane of the terdragon fractile array, even with elements spaced one-wavelength apart. For comparison purposes, we consider a uniformly excited periodic 18 18 square array of comparable size to the stage 6 terdragon fractile array, which contains a total of 308 elements. Plots of the normalized array factor for the 18 18 periodic square array are shown in Fig. 12 for element spacings of (dashed curve) and (solid curve). A grating lobe is clearly visible for the case in which the elements are periodically spaced one wavelength apart.

(18) where the matrices respectively, and and have been deﬁned in (2) and (6),

(19) (20)

Finally, we point out that the self-similar and associated iterative properties of fractile arrays could be exploited to develop fast algorithms for calculating their driving point impedances. This could be accomplished by following a similar procedure to that introduced in [5] and [16] for the more conventional Cantor linear and Sierpinski carpet planar fractal arrays.

WERNER et al.: FRACTILE ARRAYS: A NEW CLASS OF TILED ARRAYS WITH FRACTAL BOUNDARIES

2015

Fig. 13. Plots of the normalized array factor versus for ' = 0 with mainbeam steered to = 45 and ' = 0 . The solid curve represents the = =2 and the radiation pattern of a stage 6 terdragon fractile array with d dashed curve represents the radiation pattern of a uniformly excited 18 18 = d = =2. Note that terdragon arrays are examples square array with d of almost uniformly excited arrays.

2

Fig. 15. Plot of the normalized stage 5 fudgeﬂake fractile array factor versus = =2 and for ' = 90 . The dashed curve represents the case where d = . the solid curve represents the case where d

may be readily obtained by setting tuting the result into

in (21) and substi-

(22)

This leads to the following expression for the maximum directivity given by [10]:

(23) Table II lists the values of maximum directivity, calculated using (23), for several terdragon fractile arrays with and stages of growth different minimum element spacings . Table III provides a comparison between the maximum directivity of a stage 6 terdragon fractile array and that of a conventional uniformly excited 18 18 planar square array. These directivity comparisons are made for three different values , of array element spacings (i.e., ). In the ﬁrst case, where the element spacing and , we ﬁnd that the maximum is assumed to be directivity of the stage 6 terdragon array and the 18 18 square array are comparable. This is also found to be the case when (see Table III). the element spacing is increased to However, in the third case where the element spacing is in, we see that the maximum directivity for creased to the stage 6 terdragon fractile array is about 9 dB higher than its conventional 18 18 square array counterpart. This is because the maximum directivity for the stage 6 terdragon fractile array increases from 25.6 to 29.8 dB when the element spacing is changed from a half-wavelength to one-wavelength, respectively, while on the other hand, the maximum directivity for the 18 18 square array drops from 26.9 dB down to 20.9 dB. The drop in value of maximum directivity for the 18 18 square

Fig. 14. Plot of the normalized stage 4 6-terdragon fractile array factor versus = =2 and for ' = 90 . The dashed curve represents the case where d = . the solid curve represents the case where d

The array factor of any stage planar fractile array with elements may be expressed in the general form:

(21) where and represent the excitation current amplitude and phase of the th element respectively, is the horizontal position vector for the th element with magnitude and angle , and is the unit vector in the direction of the far-ﬁeld observation point. Therefore, an expression for the maximum directivity of a broadside stage planar fractile array of isotropic sources

2016

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE IV MAXIMUM DIRECTIVITY FOR SEVERAL DIFFERENT 6-TERDRAGON FRACTILE ARRAYS

array may be attributed to the appearance of grating lobes in the radiation pattern. Next, we consider the case where the mainbeam of the terdragon fractile array is steered in the direction corresponding to and . In order to accomplish this, the element phases for the terdragon fractile array are chosen according to (24) Fig. 13 shows normalized array factor plots with the mainbeam steered to and , where the solid curve results from a stage 6 terdragon fractile array and the dashed curve results from a conventional uniformly excited 18 18 square array. The minimum spacing between elements for both arrays is assumed to be a half-wavelength. This comparison demonstrates that the terdragon fractile array is superior to the 18 18 square array in terms of its overall sidelobe characteristics. At this point the radiation characteristics of the 6-terdragon fractile array illustrated in Fig. 6(a) will be investigated. A plot of the normalizad array factor as a function of for the stage 4 . The 6-terdragon fractile array is shown in Fig. 14 for , whereas dashed curve represents the case where the solid curve represents the case where . Again, we see from Fig. 14 that there are no grating lobes present for this array when the minimum spacing between elements is as much as one-wavelength. The values of maximum directivity for several 6-terdragon fractile arrays with different minimum element spacings and stages of growth are listed in Table IV. Table V provides a comparison between the maximum directivity of a stage 4 6-terdragon fractile array and a uniformly ex-

cited 15 15 planar square array for three different values of element spacings. Finally, the last example to be considered will be the stage 5 fudgeﬂake fractile array illustrated in Fig. 8(c). A plot of the normalized array factor for this array is shown in Fig. 15, where the dashed curve and solid curve represent the cases and , respectively. Table VI where lists the values of maximum directivity for several fudgeﬂake fractile arrays with different minimum element spacings and stages of growth , while Table VII provides a comparison between the maximum directivity of a stage 5 fudgeﬂake fractile array and a uniformly excited 18 18 planar square array for element spacings of , and . Therefore, this example provides yet another illustration of the unique feature characteristic of fractile arrays; namely, the fact that they possess very low sidelobes and no grating lobes will appear in the radiation patterns when the minimum spacing is changed from a half-wavelength to at least a full-wavelength. It is also interesting to note that terdragon, 6-terdragon, and fudgeﬂake fractile arrays are all deterministic examples of almost uniformly excited arrays [17]. IV. CONCLUSION A new class of antenna arrays, which we call fractile arrays, has been introduced in this paper. These fractile arrays are characterized by having a fractal boundary contour that tiles the plane without gaps or overlaps. The unique geometrical properties of fractiles have been exploited in order to develop a deterministic design methodology for modular broadband low-sidelobe arrays. The radiation properties of several different

WERNER et al.: FRACTILE ARRAYS: A NEW CLASS OF TILED ARRAYS WITH FRACTAL BOUNDARIES

2017

TABLE V COMPARISON OF MAXIMUM DIRECTIVITY FOR A STAGE 4 6-TERDRAGON FRACTILE ARRAY WITH 211 ELEMENTS AND A 15 COMPARABLE SIZE WITH 225 ELEMENTS

2 15 SQUARE ARRAY OF

TABLE VI MAXIMUM DIRECTIVITY FOR SEVERAL DIFFERENT FUDGEFLAKE FRACTILE ARRAYS

TABLE VII COMPARISON OF MAXIMUM DIRECTIVITY FOR A STAGE 5 FUDGEFLAKE FRACTILE ARRAY WITH 292 ELEMENTS AND AN 18 WITH 324 ELEMENTS

2 18 SQUARE ARRAY

fractile arrays have been investigated including Peano–Gosper, terdragon, 6-terdragon, and fudgeﬂake arrays. Efﬁcient iterative procedures for calculating the radiation patterns of these fractile arrays to arbitrary stage of growth have also been developed in this paper.

REFERENCES

[1] J. L. Vehl, E. Lutton, and C. Tricot, Eds., Fractals in Engineering. New York: Springer-Verlag, 1997. [2] D. H. Werner, R. L. Haupt, and P. L. Werner, “Fractal antenna engineering: The theory and design of fractal antenna arrays,” IEEE Antennas Propagat. Mag., vol. 41, no. 5, pp. 37–59, Oct. 1999.

2018

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[3] D. H. Werner and R. Mittra, Eds., Frontiers in Electromagnetics. Piscataway, NJ: IEEE Press, 2000. [4] J. P. Gianvittorio and Y. Rahmat-Samii, “Fractal antennas: A novel antenna miniaturization technique, and applications,” IEEE Antennas Propagat. Mag., vol. 44, pp. 20–36, Feb. 2002. [5] D. H. Werner and S. Ganguly, “An overview of fractal antenna engineering research,” IEEE Antennas Propagat. Mag., vol. 45, pp. 38–57, Feb. 2003. [6] Y. Kim and D. L. Jaggard, “The fractal random array,” Proc. IEEE, vol. 74, no. 9, pp. 1278–1280, 1986. [7] C. P. Baliarda and R. Pous, “Fractal design of multiband and low sidelobe arrays,” IEEE Trans. Antennas Propagat., vol. 44, pp. 730–739, May 1996. [8] D. L. Jaggard and A. D. Jaggard, “Cantor ring arrays,” Microwave and Opt. Technol. Lett., vol. 19, pp. 121–125, 1998. [9] D. H. Werner, K. C. Anushko, and P. L. Werner, “The generation of sum and difference patterns using fractal subarrays,” Microwave and Opt. Technol. Lett., vol. 22, no. 1, pp. 54–57, July 1999. [10] D. H. Werner, W. Kuhirun, and P. L. Werner, “The Peano–Gosper fractal array,” IEEE Trans. Antennas Propagat., vol. 51, pp. 2063–2072, Aug. 2003. , “A new class of modular broadband arrays based on gosper islands [11] and associated Peano–Gosper curves,” in Proc. IEEE Int. Antennas and Propagation Symp. and URSI North American Radio Science Meeting, vol. 4, Columbus, OH, June 22–27, 2003, pp. 250–253. , “A new design methodology for modular broadband arrays based [12] on fractal tilings,” in Proc. IEEE Topical Conf. Wireless Communication Technology, Honolulu, HI, Oct. 15–17, 2003. [13] B. B. Mandelbrot, The Fractal Geometry of Nature. New York: Freeman, 1983. [14] G. A. Edgar, Measure, Topology, and Fractal Geometry. New York: Springer-Verlag, 1990. [15] B. Grunbaum and G. C. Shephard, Tilings and Patterns. New York: W. H. Freeman and Company, 1987. [16] D. H. Werner, D. Baldacci, and P. L. Werner, “An efﬁcient recursive procedure for evaluating the impedance matrix of linear and planar fractal arrays,” IEEE Trans. Antennas Propagat., vol. 52, pp. 380–387, Feb. 2004. [17] P. Lopez, J. A. Rodríguez, F. Ares, and E. Moreno, “Low sidelobe level in almost uniformly excited array,” Inst. Elect. Eng. Electron. Lett., vol. 36, no. 24, pp. 1991–1993, Nov. 2000.

Douglas H. Werner (S’81–M’89–SM’94) received the B.S., M.S., and Ph.D. degrees in electrical engineering and the M.A. degree in mathematics from The Pennsylvania State University (Penn State), University Park, in 1983, 1985, 1989, and 1986, respectively. He is an Associate Professor in the Department of Electrical Engineering, Penn State. He is a member of the Communications and Space Sciences Lab (CSSL) and is afﬁliated with the Electromagnetic Communication Research Lab. He is also a Senior Research Associate in the Electromagnetics and Environmental Effects Department of the Applied Research Laboratory at Penn State. He is a former Associate Editor of Radio Science. He has published numerous technical papers and proceedings articles and is the author of nine book chapters. He is an Editor of Frontiers in Electromagnetics (Piscataway, NJ: IEEE Press, 2000). He also contributed a chapter for Electromagnetic Optimization by Genetic Algorithms (New York: Wiley Interscience, 1999). His research interests include theoretical and computational electromagnetics with applications to antenna theory and design, microwaves, wireless and personal communication systems, electromagnetic wave interactions with complex media, meta-materials, fractal and knot electrodynamics, and genetic algorithms. Dr. Werner is a Member of the American Geophysical Union (AGU), International Scientiﬁc Radio Union (URSI) Commissions B and G, the Applied Computational Electromagnetics Society (ACES), Eta Kappa Nu, Tau Beta Pi, and Sigma Xi. He received the 1993 Applied Computational Electromagnetics Society (ACES) Best Paper Award and a 1993 URSI Young Scientist Award. In 1994, he received the Pennsylvania State University Applied Research Laboratory Outstanding Publication Award. He received a College of Engineering PSES Outstanding Research Award and Outstanding Teaching Award in March 2000 and March 2002, respectively. He recently received an IEEE Central Pennsylvania Section Millennium Medal. He has also received several Letters of Commendation from Penn State’s Department of Electrical Engineering for outstanding teaching and research. He is an Editor of IEEE ANTENNAS AND PROPAGATION MAGAZINE.

Waroth Kuhirun received the B. Eng. degree from Chulalongkorn University, Thailand, in 1994 and the M.S. and Ph.D. degrees in electrical engineering from Pennsylvania State University, University Park, in 1998 and 2003, respectively. From 1994 to 1995, he worked at Kasetsart University, Thailand. His research interest is in the area of fractal and fractile antenna arrays. Dr. Kuhirun received a scholarship from the Thai Government for his M.S. and Ph.D. studies.

Pingjuan L. Werner (SM’02) is an Associate Professor with the Pennsylvania State University College of Engineering. Her primary research focuses are in the area of electromagnetics, including fractal antenna engineering and the application of genetic algorithms in electromagnetics. Prof. Werner is a Fellow of the Leonhard Center, College of Engineering, The Pennsylvania State University, and a Member of Tau Beta Pi National Egineering Honor Society, Eta Kappa Nu National Electrical Engineering Honor Society, Sigma Xi National Research Honor Society. She received The Best Paper Award from the Applied Computational Electromagnetics Society in 1993.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2019

A New Millimeter-Wave Printed Dipole Phased Array Antenna Using Microstrip-Fed Coplanar Stripline Tee Junctions

Young-Ho Suh, Member, IEEE, and Kai Chang, Fellow, IEEE

Abstract—A new millimeter-wave printed twin dipole phased array antenna is developed at Ka band using a new microstrip-fed CPS Tee junction, which does not require any bonding wires, air bridges, or via holes. The phased array used a piezoelectric transducer (PET) controlled tunable multitransmission line phase shifter to accomplish a progressive phase shift. A progressive phase shift of 88.8 is achieved with the 5 mm of perturber length when the PET has full deﬂection. Measured return loss of the twin dipole antenna is better than 10 dB from 29.5 to 30.35 GHz. Measured return loss of better than 15 dB is achieved from 30 to 31.5 GHz for a 1 8 phased array. The phased array antenna has a measured antenna gain of 14.4 dBi with 42 beam scanning and has more than 11 dB side lobe suppression across the scan. Index Terms—Coplanar stripline (CPS), CPS Tee junction, coplanar transmission lines, dipole antenna, microstrip-to-CPS transition, phase shifter, phased array antenna, piezoelectric transducer phase shifter, twin dipole antenna.

I. INTRODUCTION HASED array antenna systems usually associated with large and complex active device networks for phase shifters, which occupies large portion of the system expenses. Phased array used in military radar system requires low proﬁle for invisibility against opponents. It also needs to be light weight especially in the applications of satellite communications. Correspondingly, the demands for low cost, low proﬁle, small size, light weight, and less complicated phased array antenna systems are increasing nowadays for both commercial and military applications. A printed dipole antenna satisﬁes the beneﬁts of low proﬁle, light weight, low cost and compact size, which is suitable for building phased arrays if proper phase shifters are provided. To construct a printed dipole array, several conﬁgurations have been proposed. Nesic et al. [1] reported a one-dimensional printed dipole antenna array fed by microstrip at 5.2 GHz. Scott [2] introduced a microstrip-fed printed dipole array using a microstrip-to-coplanar stripline (CPS) balun. In [1] and [2], the balun designs were not easy to match the impedance and the structures were too big and complicated to build an

Manuscript received August 1, 2002; revised June 17, 2003. This work was supported in part by the National Science Foundation and NASA Glenn Research Center. Y.-H. Suh was with the Department of Electrical Engineering, Texas A&M University, College Station, TX 77840 USA. He is now with Mimix Broadband Inc., Houston, TX 77099 USA (e-mail: ysuh@mimixbroadband.com). K. Chang is with Department of Electrical Engineering, Texas A&M University, College Station, TX 77840 USA (e-mail: chang@ee. tamu.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832510

P

Fig. 1. CCPS structure (a) original CPS, (b) CCPS, (c) cross-sectional view at A-A’ with ﬁelds distributions of the CCPS for different layers of metallization.

array. In 1998, a wideband microstrip-fed twin dipole antenna was introduced with double-sided structure operating at the frequency range from 0.61 to 0.96 GHz [3]. Zhu and Wu [4] developed a 3.5 GHz twin dipole antenna fed by a hybrid ﬁnite ground coplanar waveguide (FGCPW)/CPS Tee junction. An X-band monolithic integrated twin dipole antenna mixer was reported in [5] with devices directly integrated into the antenna, so no feeding network was necessary. In this paper, a new planar printed dipole phased array antenna using a tunable phase shifter controlled by PET is presented at 30 GHz. The phased array antenna uses a new twin dipole antenna excited by a microstrip-fed CPS Tee junction [8]. The piezoelectric transducer (PET) controlled phase shifter does not require any solid-state devices and their associated driving circuits. The 1 8 twin dipole phased array antenna has compact size, low loss, low cost, light weight and reduced complexity as well as good beam scanning with low side lobe levels. The PET controlled phase shifter was adopted for the low cost phased array antenna systems for the ﬁrst time in [6] and [7]. In this structure, a dielectric perturber controlled by PET is used to introduce a progressive phase shift. The deﬂection takes place at the PET when the proper voltages are applied. Using this property of the PET, a dielectric perturber can have upward and downward movement according to the applied voltages. Consequently, if a transmission line is perturbed by a PET actuated dielectric perturber, its propagation constant

0018-926X/04$20.00 © 2004 IEEE

2020

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Simulated performances comparison at Ka band between conventional and coupled CPS.

will be changed. This phenomenon induces a variable phase shift along the transmission line controlled by PET. In [6] and [7], an end-ﬁre Vivaldi antenna was used for covering a wide bandwidth, and a transition was required to feed antennas. Consequently, the system was large and bulky. The new printed twin dipole phased array using a microstrip-fed CPS combined with a PET phase shifter provides low-cost, low-loss, low-proﬁle, compact-size and low-complexity with simple antenna feeding. II. A MICROSTRIP-FED CPS TEE JUNCTION The twin dipole antenna is fed by a CPS. Since conventional planar transmission line is microstrip line, a microstrip-to-CPS transition is needed to feed the dipole. A microstrip-fed CPS Tee junction without using bonding wires or air bridges was introduced in [8]. In [8], the operating frequency is centered near 3.5 GHz with 0.7 dB insertion loss ranged from 2 to 4.15 GHz. The Tee junction utilized novel coupled CPS (CCPS). This transmission line can have a physical discontinuity while ﬁelds are continuous over the whole transmission line using CCPS. The structure of original CPS and CCPS at 30 GHz is shown in Fig. 1(a) and (b). For the CPS, CCPS, Tee junction and antenna design, IE3D software [9], which uses the method of moments, is employed for full wave electromagnetic simulation. A 31 mil RT/Duroid 5870 substrate with a dielectric constant of 2.33 is used for the antenna and feeding network fabrication. The width ( ) of CPS strip is 0.65 mm and gap ( ) between the strips is 0.5 mm, which has the characteristic impedance of 202 . This impedance is chosen to match a dipole antenna input impedance which will be shown later. As shown in Fig. 1(b), one of the CPS strips is discontinued and is terminated with radial stubs with a rotation angle of 30 and a radius of 0.65 mm for coupling to the bottom layer metallization. The bottom layer metallization, which is coupled from the top layer’s radial stubs, works as a CPS strip shown in Fig. 1(c). The radial stub is used to accomplish the smooth ﬁeld transition. The wideband coupling performance of radial stubs has been reported in the microstrip-to CPS-to-microstrip back-to-back transition for lower frequency operation [10]. The back-to-back transition has a measured 3 dB insertion loss over a frequency range from 1.3 to 13.3 GHz (1:10.2) and return loss

is better than 10 dB. The radial stub provides virtual short to the bottom layer metallization, which depends on the radius of the radial stub. Hence, smaller radius of radial stub gives higher operating frequency with minimal insertion loss and return loss deteriorations compared to the original CPS conﬁguration. Performances of CCPS are simulated with IE3D and compared with those of conventional CPS as shown in Fig. 2. The simulated transmission line length is about 5 mm and the conventional CPS has almost zero insertion loss with that short length transmission line. Fig. 2 shows that the insertion loss of CCPS is deteriorated by about 1 dB as compared with that of conventional CPS for the frequency range from 29.2 GHz to 35 GHz and the return loss is better than 10 dB. Insertion loss deterioration of less than 2 dB covers the wider frequency range from 26.4 GHz to 35 GHz. From the above results, CCPS shows that ﬁelds are continuous all over the transmission line with the aid of radial stub, though a discontinuity is introduced at one of the CPS strips. The structure of microstrip-fed CPS Tee junction at 30 GHz is shown in Fig. 3. The Tee junction has the characteristic at each output port 1 and 2. The input impedance of 202 impedance to the microstrip feed at port 3 is about 101 , which is half of 202 . Radial stubs effectively rotate the electric ﬁelds from parallel to the normal to the substrate to have a good coupling to the bottom metallization, which provides the ground of microstrip line. The Tee junction is simulated with IE3D to verify the performance at 30 GHz. Simulated performance of the Tee junction is shown in Fig. 4. The simulated performance shows that the Tee junction equally splits the power to each CPS port with 1.2 dB insertion loss at 30 GHz. Simulated 2 dB insertion loss bandwidth of the Tee junction is from 27.2 to 34.8 GHz, and the return loss is better than 20 dB. Because of high frequency operation bandwidth restriction of the microstrip-to-CPS transition in [8], the Tee junction is not measured but simulation results quite veriﬁes its performances. III. TWIN DIPOLE ANTENNA USING MICROSTRIP-FED CPS TEE JUNCTION The structure of the twin dipole antenna is illustrated in Fig. 5. The twin dipole antenna utilizes the microstrip-fed CPS

SUH AND CHANG: A NEW MILLIMETER-WAVE PRINTED DIPOLE PHASED ARRAY ANTENNA

2021

Fig. 3. Structure of Ka-band microstrip-fed CPS Tee junction for twin dipole antenna feeding near 30 GHz.

Fig. 4.

Simulated performance of the Tee junction near 30 GHz.

Fig. 5. Structure of printed twin dipole antenna using a microstrip-fed CPS Tee junction.

Tee junction as discussed in Section II. The antenna is placed in front of a reﬂector for uni-directional radiation. The reﬂector is spaced from the antenna at the distance of 1.5 mm (60 mil), which is about 0.15 . The length of dipole is 5.3 mm or 0.53 . The spacing between dipoles was optimized to be because of an insertion loss increase in CCPS in the 0.36 Tee-junction with a long coupled line such as 0.5 , causing a gain drop. Mutual coupling normally takes place when antenna spacing is less than a half wavelength. Twin dipole antenna’s input impedance is supposed to have some reactance due to this coupling effect. By adjusting the reﬂector’s spacing, this reactance can be minimized with a small change in input impedance. The input impedance of a single dipole antenna is around 202 . The strip width ( ) and gap ( ) between strips of CCPS at the CPS Tee junction in Section II are determined to have a CCPS characteristic impedance identical to the dipole antenna input impedance for good impedance matching. Measured return loss of the twin dipole antenna is better than 10 dB from 29.5 to 30.35 GHz as shown in Fig. 6. Measured and simulated return losses have good agreements. For measurements, a quarter-wavelength transformer with limited bandwidth is used and causes small discrepancies between simulated data and measurements. Radiation patterns of the antenna are measured in an anechoic chamber. The measured radiation patterns are shown in Fig. 7. and -plane radiation patterns are quite similar to each other for the twin dipole antenna as discussed in [4]. Measured and -plane gains are about 7.6 and 7.7 dBi with the 3 dB beamwidths of 32 and 34 , respectively. The measured cross-polarizations at broadside are about 47.7 and 42.4 dB down compared with the copolarization levels in and -plane, respectively. Gains and 3 dB beam widths of and -planes are quite close to each other. Some discrepancies of gains and 3 dB beam widths are partly due to the small misalignments of the antenna in millimeter-wave frequencies.

2022

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 6.

Simulated and measured return loss of the twin dipole antenna.

Fig. 7. Measured radiation patterns of the twin dipole antenna.

IV. PHASED ARRAY ANTENNA WITH MULTITRANSMISSION LINE PET CONTROLLED PHASE SHIFTER For the linear phased array, an array factor is a function of the progressive phase shift and the element spacing . The array factor is given as (1) where (2) and is beam scanning angle. where is expressed as is the number of elements. The progressive phase shift causes the radiation emitted from the array to have a constant phase front that is pointing at the angle . This beam scanning angle ( ) is also a function of and , given by (3) The array factor in (1) can also be expressed as (4) below in an alternate, compact and closed form whose function and their

distributions are more recognizable [11]. It is assumed that the reference point is the physical center of the array. (4) The total ﬁeld of array is equal to the ﬁeld of a single element positioned at the origin multiplied by an array factor, which is expressed as (5) From (4), the maximum value of array factor is [11]. Hence maximum achievable gain of the array can be found from (4) and (5), which is expressed as

(6) In (6), the effect of mutual coupling between elements is excluded for the simplicity. Mutual coupling normally degrades arrayed antenna gain. Equation (6) can be used for the gain approximation of the array. To achieve more accurate calculation including mutual coupling effects, a full-wave electromagnetic simulation can be used for antenna array analysis.

SUH AND CHANG: A NEW MILLIMETER-WAVE PRINTED DIPOLE PHASED ARRAY ANTENNA

2023

Fig. 8. Structure of printed dipole phased array antenna controlled by PET (a) top view and (b) side view.

The structure of 1 8 printed twin dipole phased array antenna is shown in Fig. 8. A conventional microstrip power divider with binominal impedance transformers is used for feeding network to cover the wide bandwidth. The bottom metallization provides good ground plane for the microstrip. To obtain the required phase shift, the 101 microstrip line, which has the same input impedance as the twin dipole antenna, is perturbed with a dielectric perturber actuated by PET. The length of dielectric perturber varies linearly from 5 to 35 mm on top of line 2 to line 8 in Fig. 8. The ﬁrst line is not perturbed. The PET is conﬁgured to have no deﬂection (no perturbation) when a DC voltage of 0 V is applied, and full deﬂection (full perturbation) when a DC voltage of 50 V is applied. A 50 mil RT/duroid 6010.2 with a dielectric constant of 10.2 is used as the dielectric perturber. The amount of phase shift is linearly proportional to the length of perturber [7], which is expressed as (7) is the perturber length along the th transwhere, represents the differential propagation conmission line. stant expressed as (8) represents the propagation constant of the where th perturbed transmission line, which is microstrip in this case. Since the ﬁrst perturbed microstrip line (i.e., the second line or line 2) has the minimum perturbed length, the following relationship is obtained. (9)

Fig. 9. Differential phase shift for 5 mm dielectric perturber controlled by PET.

With a dielectric perturber of 5 mm, Fig. 9 shows that a differential phase shift of 88.8 takes place with a 2 dB insertion loss. Narrower microstrip line generates larger phase shift but the insertion loss is increased. Hence, a proper microstrip line’s width should be chosen for having a good phase shift as well as low insertion loss. Table I summarizes the design and measured parameters for the twin dipole phased array. The parameter values in Table I are useful in analytical calculations of the scanning angle ( ), maximally achievable gain, and optimum element spacing ( ) of the phased array. According to (7) and (8), the perturber’s length can be determined for a desired phase shift. A length of 5 mm dielectric perturber produces about 88.8 differential phase shift. Accordingly, the length of each neighboring perturbed line is increased by 5 mm. The length of perturber for the ) is about 35 mm, which gives ﬁnal microstrip ( a differential phase shift of 621.6 .

2024

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE I PARAMETER VALUES OF THE TWIN DIPOLE PHASED ARRAY

TABLE II COMPARISON AMONG ANALYTICAL, SIMULATION, AND MEASURED RESULTS OF THE 1

2 8 PHASED ARRAY

Fig. 10.

Measured return loss of the printed twin dipole phased array antenna.

IE3D analysis shows that a progressive phase shift of 88.8 gives around beam scanning with low side lobe levels. An analytical scan angle can also be obtained using (3), and maximally achievable gain of the phased array can be obtained from (6). The maximum spacing ( ) between elements to avoid grating lobes is expressed as (10) From analytical equations in (3), (6) and (10) and the parameters in Table I, the calculated , maximally achievable gain, and maximum spacing are calculated to be 19.47 , 16.73 dBi, and 7.5 mm, respectively. The results agree very well with IE3D simulation as given in Table II. Measured return loss of the 1 8 twin dipole array is plotted in Fig. 10. The measured return loss is about 41.9 dB at 30.3 GHz for the unperturbed twin dipole phased array antenna. With perturbation by the dielectric perturber, the return loss is about 31.8 dB at 30.7 GHz, which shows a 0.4 GHz frequency shift compared with the unperturbed result. For a bandwidth from 30 to 31.5 GHz, a measured return loss is better than 15 dB.

V. PHASED ARRAY MEASUREMENTS The phased array is measured in an anechoic chamber. As shown in Fig. 8, the antenna is arrayed for the -plane beam scanning. To accomplish bidirectional beam scanning, two triangular perturbers are used side by side [12]. PET actuation for the dielectric perturber is conﬁgured as 0 V for no perturbation (no PET deﬂection) and 50 V for full perturbation (full PET deﬂection). The measured twin dipole phased array antenna gain without perturbation (0 V for PET) is about 14.4 dBi with a 3 dB beam width of 6 as shown in Fig. 11. The fully perturbed antenna with a dielectric perturber controlled by PET shows about ) beam scanning with the gain of 12.2 dBi. 42 ( Side lobe levels of the steered beam are more than 11 dB down compared with main beam. The gains of steered beams are about 2.2 dB down due to the insertion loss incurred by dielectric perturbation. The beam can be dynamically steered depending on the voltages applied to PET because the amount of phase shift changes according to the applied voltages on PET as shown in Fig. 9. The comparison among analytical, simulation, and measured results of the phased array are exhibited in Table II. Beam scanning angle is following closely among analytical, IE3D simula-

SUH AND CHANG: A NEW MILLIMETER-WAVE PRINTED DIPOLE PHASED ARRAY ANTENNA

2025

Fig. 11. Measured perturbation.

H-plane radiation pattern for twin dipole phased array antenna at 30 GHz. Measured beam scanning is from 020

to

+22

with full

tion, and measured results. Measured unperturbed gain is about 2.3 dB lower than analytical or IE3D simulated data. This is due to the insertion loss of power divider and the mutual coupling effects among elements, which normally degrades antenna gain. The measured gains of steered beams are about 2.2 dB down compared to that of unperturbed beam due to the insertion loss incurred by dielectric perturbation. VI. CONCLUSION A new printed twin dipole phased array antenna is developed at 30 GHz using a multitransmission line tunable phase shifter controlled by a PET. The new twin dipole antenna is designed using a microstrip-fed CPS Tee junction. To construct the Tee junction, CCPS is used to have a physical discontinuity at CPS while ﬁelds are continuous all over the transmission line. The Tee junction effectively splits power to each CPS output port with low insertion loss. The PET actuated phase shifter requires only one (one-directional beam scanning) or two (bi-directional beam scanning) applied voltages to produce the progressive phase shift. A PET controlled phase shifter is tested and optimized for the proper phase shift with minimal insertion loss. The twin dipole phased array antenna shows a ) beam scanning with more than 11 dB 42 ( side lobe suppression across the scan. The phased array should ﬁnd many applications in wireless communications and radar systems. ACKNOWLEDGMENT The authors would like to thank C. Wang of Texas A&M University for technical assistance. REFERENCES

[1] A. Nesic, S. Jovanovic, and V. Brankovic, “Design of printed dipoles near the third resonance,” in Proc. IEEE Int. Antennas and Propagation Symp. Dig., vol. 2, Atlanta, GA, 1998, pp. 928–931. [2] M. Scott, “A printed dipole for wide-scanning array application,” in Proc. IEEE 11th Int. Conf. Antennas and Propagation, vol. 1, 2001, pp. 37–40. [3] G. A. Evtioushkine, J. W. Kim, and K. S. Han, “Very wideband printed dipole antenna array,” Electron. Lett., vol. 34, no. 24, pp. 2292–2293, Nov. 1998. [4] L. Zhu and K. Wu, “Model-based characterization of CPS-fed printed dipole for innovative design of uniplanar integrated antenna,” IEEE Microwave and Guided Wave Lett., vol. 9, pp. 342–344, Sept. 1999.

[5] K. L. Deng, C. C. Meng, S. S. Lu, H. D. Lee, and H. Wang, “A fully monolithic twin dipole antenna mixer on a GaAs substrate,” in Proc. Asia Paciﬁc Microwave Conf. Dig., Sydney, NSW, Australia, 2000, pp. 54–57. [6] T. Y. Yun and K. Chang, “A phased-array antenna using a multi-line phase shifter controlled by a piezoelectric transducer,” in IEEE Int. Microwave Symp. Dig., vol. 2, Boston, MA, 2000, pp. 831–833. , “A low-cost 8 to 26.5 GHz phased array antenna using a piezoelec[7] tric transducer controlled phase shifter,” IEEE Trans. Antennas Propagat., vol. 49, pp. 1290–1298, Sept. 2001. [8] Y. H. Suh and K. Chang, “A microsatrip fed coplanar stripline Tee junction using coupled coplanar stripline,” in Proc. IEEE Int. Microwave Symp. Dig., vol. 2, Phoenix, AZ, 2001, pp. 611–614. [9] IE3D, 8.1 ed., Zeland Software Inc., 2001. [10] Y. H. Suh and K. Chang, “A wideband coplanar stripline to microstrip transition,” IEEE Microwave and Wireless Components Lett., vol. 11, pp. 28–29, Jan. 2001. [11] C. A. Balanis, Antenna Theory Analysis and Design, 2nd ed. New York: Wiley. [12] T. Y. Yun, C. Wang, P. Zepeda, C. T. Rodenbeck, M. R. Coutant, M. Y. Li, and K. Chang, “A 10- to 21-GHz, low-cost, multifrequency, and fullduplex phased-array antenna system,” IEEE Trans. Antennas Propagat., vol. 50, pp. 641–650, May 2002.

Young-Ho Suh (S’01–M’02) received the B.S degree in electrical and control engineering from Hong-Ik University, Seoul, Korea, in 1992, and the M.S and Ph.D. degrees in electrical engineering from Texas A&M University, College Station, TX, in 1998, and 2002, respectively. From 1992 to 1996, he worked for LG-Honeywell Co. Ltd., Seoul, Korea, as a Research Engineer. From 1996 to 1998, he worked on developing robust wireless communication systems for GSM receiver under multipath fading channel for his M.S degree. From 1998 to 2002, he was a Research Assistant in the Electromagnetics and Microwave Laboratory, Department of Electrical Engineering, Texas A&M University, College Station, TX, where he was involved in rectenna design for wireless power transmissions, phased array antennas, and coplanar transmission line circuit components development. In May 2002, he joined Mimix Broadband Inc., Houston, TX, as a Senior Microwave Design Engineer, where he is working on state-of-the-art microwave/millimeter-wave active circuit designs including low noise/power ampliﬁers, receivers, transmitters, and transceiver modules for LMDS, point-to-point, point-to-multipoint radio systems in Ka band using GaAs MMICs. His research area includes state-of-the-art millimeter-wave transceiver modules, transitions between dissimilar transmission lines, uniplanar transmission line analysis and components development, microwave power transmission, antennas for wireless communications, and phased array antennas.

2026

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Kai Chang (S’75–M’76–SM’85–F’91) received the B.S.E.E. degree from the National Taiwan University, Taipei, Taiwan, R.O.C., the M.S. degree from the State University of New York at Stony Brook, and the Ph.D. degree from the University of Michigan, Ann Arbor, in 1970, 1972, and 1976, respectively. From 1972 to 1976, he worked for the Microwave Solid-State Circuits Group, Cooley Electronics Laboratory, University of Michigan, as a Research Assistant. From 1976 to 1978, he was employed by Shared Applications, Inc., Ann Arbor, where he worked in computer simulation of microwave circuits and microwave tubes. From 1978 to 1981, he worked for the Electron Dynamics Division, Hughes Aircraft Company, Torrance, CA, where he was involved in the research and development of millimeter-wave solid-state devices and circuits, power combiners, oscillators and transmitters. From 1981 to 1985, he worked for the TRW Electronics and Defense, Redondo Beach, CA, as a Section Head, developing state-of-the-art millimeter-wave integrated circuits and subsystems including mixers, VCOs, transmitters, ampliﬁers, modulators, upconverters, switches, multipliers, receivers, and transceivers. He joined the Electrical Engineering Department of Texas A&M University in August 1985 as an Associate Professor and was promoted to a Professor in 1988. In January 1990, he was appointed E-Systems Endowed Professor of Electrical Engineering. He authored and coauthored several books “Microwave Solid-State Circuits and Applications” (New York: Wiley, 1994), “Microwave Ring Circuits and Antennas” (New York: Wiley, 1996), “Integrated Active Antennas and Spatial Power Combining” (New York: Wiley, 1996), and “RF and Microwave Wireless Systems” (New York: Wiley, 2000). He served as the editor of the four-volume “Handbook of Microwave and Optical Components” (New York: Wiley, 1989 and 1990). He is the Editor of the Microwave and Optical Technology Letters and the Wiley Book Series in Microwave and Optical Engineering. He has published over 350 technical papers and several book chapters in the areas of microwave and millimeter-wave devices, circuits, and antennas. His current interests are in microwave and millimeter-wave devices and circuits, microwave integrated circuits, integrated antennas, wideband and active antennas, phased arrays, microwave power transmission, and microwave optical interactions. Dr. Chang received the Special Achievement Award from TRW in 1984, the Halliburton Professor Award in 1988, the Distinguished Teaching Award in 1989, the Distinguished Research Award in 1992, and the TEES Fellow Award in 1996 from the Texas A&M University.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2027

**Physical Limitations of Antennas in a Lossy Medium
**

Anders Karlsson

Abstract—The dissipated power and the directivity of antennas in a homogeneous, lossy medium are systematically analyzed in this paper. The antennas are ideal and located inside a lossless sphere. In the lossy space outside the sphere, the electromagnetic ﬁelds are expanded in a complete set of vector wave functions. The radiation efﬁciency, the directivity, and the power gain are deﬁned for antennas in a lossy medium, and the optimal values of these quantities are derived. Simple relations between the maximal number of ports, or channels, an antenna can use and the optimal directivity and gain of the antenna are presented. Index Terms—Antenna theory, lossy systems.

and the power gain of antennas are deﬁned and studied for the simpliﬁed geometry where the antenna is enclosed in a lossless sphere. The optimal values of these three quantities are the main results in this paper. The optimal value of the directivity is shown to be related to the maximum number of ports, or channels, of the antenna, a result that holds also in a lossless medium. It is emphasized that in a lossy medium the magnetic dipole is the most radiation efﬁcient antenna, a well known and important result, cf. [11]. II. PRELIMINARIES

I. INTRODUCTION N some applications there is a need for wireless communication with devices in lossy materials. A conductive medium is a low-pass ﬁlter for the electromagnetic waves, and one is then often forced to use low frequencies, or equivalently, long wavelengths. If the space for the antenna is limited it results in an antenna that is small compared to the wavelength. The drawback is that small antennas in lossy materials consume much power, due to the ohmic losses in the near-zone of the antenna. Hence, the design of the antenna and the choice of frequency are delicate problems, where two power loss mechanisms with counteracting frequency dependences are involved. This power problem is addressed in this paper. Antennas in lossy materials are found in various areas. In geophysical applications underground antennas are used, e.g., in bore holes. In marine technology antennas are used for communication with underwater objects. In medical applications there is an increased usage of wireless communication with implants. Implants, e.g., pacemakers, have limited power supply and it is important to use power efﬁcient antennas. Some of the results in this paper are based on the results obtained by Chu [4] and Harrington [8], who investigated physical limitations for antennas in free space. Chu derived the optimal value of the directivity and the optimal value of the ratio between the directivity and the -value of omni-directional antennas and Harrington derived the corresponding results for general antennas. There are a number of other articles that address the optimization of the -value of an antenna, cf. [5], [7], and [12]. For a lossy material it is the dissipated power, rather than the -value, that is the most important quantity in the design of an antenna. In this paper, the radiation efﬁciency, the directivity,

Manuscript received May 9, 2003; revised August 6, 2003. This work was supported by the Competence Center for Circuit Design at Lund University. The author is with the Department of Electroscience, Lund Institute of Technology, S-221 00 Lund, Sweden (e-mail: anders.karlsson@es.lth.se). Digital Object Identiﬁer 10.1109/TAP.2004.832335

I

The antennas are conﬁned in a spherical, lossless region, denoted . They are idealized in the sense that there are no ohmic losses in . The volume is denoted and is an inﬁnite, homogeneous, conducting medium with a complex permittivity (2.1) is assumed. The corresponding where the time-dependence wave number is denoted and is given by (2.2) The permeability impedance in reads is assumed to be real. The wave

(2.3)

III. GENERAL ANTENNAS IN CONDUCTING MEDIA the electric ﬁeld is expanded in In the exterior region spherical vector waves , also referred to as partial waves. These waves satisfy Maxwell’s equations and are complete on a spherical surface. The details of the spherical vector waves are given in Appendix A. The expansion reads (3.1) The corresponding magnetic ﬁeld is given by the induction law

(3.2)

0018-926X/04$20.00 © 2004 IEEE

2028

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

where . Here, is the index for the two different wave types (TE and TM), for waves that are even with respect to the azimuthal angle and for the waves that are odd with respect to is the index for the polar direction, and is the index for the azimuthal angle. For only the partial waves with are nonzero, cf., (A.2). The expansion in (3.1) covers all possible types of time harmonic sources inside . A. Classiﬁcation Antennas that radiate partial waves with are referred to as magnetic antennas, since the reactive part of their radiated complex power is positive, i.e., inductive. Antennas radiating are referred to as electric antennas, partial waves with since they are capacitive when they are small compared to the wavelength. in the expansion (3.1) can The expansion coefﬁcients theoretically be altered independently of each other. Hence, each partial wave corresponds to an independent port of the antenna. The maximum number of ports, or channels, an antenna can use is then equal to the maximum number of partial waves the antenna can radiate. The following classiﬁcation of antennas is used in this paper: Partial wave antenna—antenna that radiates only one partial wave . The antenna has one port. Magnetic multipole antenna of order An antenna that radiates partial waves with and index . The max. imum number of ports is Electric multipole antenna of order An antenna that raand index . The maximum diates partial waves with number of ports is . An antenna that radiates Magnetic antenna of order and with . The partial waves with . maximum number of ports is Electric antenna of order An antenna that radiates and with . The partial waves with . maximum number of ports is An antenna that radiCombined antenna of order ates partial waves with and . The . maximum number of ports is B. Rotation of an Antenna If an antenna is rotated, the new set of radiated partial waves is determined by the rotational matrix for the vector waves, cf. [3]. That matrix is diagonal in the index and in the index , but not in the other two indices and . Thus, a magnetic multipole antenna of index is still a magnetic multipole antenna of index , after it is rotated. This type of invariance under rotation is true for all types of antennas in Section III-A, except for the partial wave antenna. The invariance is utilized in Section IV to determine the optimal values of the directivity and power gain. , is A partial wave antenna that radiates the partial wave under a rotation transformed to an antenna that radiates several , where can be both , , and can partial waves . take the values

C. The Power Flow The complex power by radiated from an antenna is given

(3.3) is the radial unit vector, where is the surface is the complex conjugate of the magnetic ﬁeld. The and complex power is decomposed as (3.4) is the power dissipated in the The active part of the power , whereas and are the time averages of the region stored magnetic and electric energies in the exterior region. The impedance and admittance of the antenna are related by the power relation to the complex power (3.5) where and are the complex current and voltage that feeds the antenna, respectively. The star denotes complex conjugate. should be added For a nonideal antenna the powers inside to the left-hand side of (3.5). The complex power radiated from a combined antenna of follows from (A.4) and (A.5), and from (3.1)–(3.3) order

(3.6) The complex powers of the other types of antennas in Section III-A are special cases of (3.6). The normalized complex power, , of multipole antennas of order depends only on the indices and . If the transmitted complex power of such an antenna is denoted and the corresponding then impedance is denoted

(3.7)

where

is the wave impedance. and

D. Asymptotic Values of

When the asymptotic behavior of the Hankel functions, (A.6), implies that the asymptotic values of the radiated complex power, cf., (3.6) and of the impedance, cf., (3.5), are (3.8)

KARLSSON: PHYSICAL LIMITATIONS OF ANTENNAS IN A LOSSY MEDIUM

2029

As

the limiting values of the Hankel functions yield

(3.9) The asymptotic values in (3.8) and (3.9) are valid for all of the antennas in Section III-A. The values are illustrated in Fig. 1. In the rest of the pape,r only active power will be considered and for this reason the term power is understood to mean active power. E. Far-Field and Directivity The far-ﬁeld amplitude is deﬁned as (3.10) The far-ﬁeld amplitude of a combined antenna of order is given by the asymptotic values of the spherical Hankel functions, cf., (A.6)

Fig. 1. Argument of the impedance, arg(Z ) of magnetic multipole antennas of order l = 1 (lower dashed line), 2 (middle dashed), and 3 (upper dashed) and for the electric multipole antennas of order l = 1 (upper solid line), 2 (middle solid), and 3 (lower solid). The frequency is 400 MHz, " = 50, and = 1 S/m, corresponding to an argument of the wave impedance arg( ) = 0 and 0:37 rad. The asymptotic values in (3.8) and (3.9) are reached when a , respectively.

1

!

(3.11) The far-ﬁeld amplitude of the antenna corresponding to (3.1) is . thus completely deﬁned by the coefﬁcients The directivity is deﬁned in the same way as for a lossless medium, cf., [10]. The directivity of the combined antenna is obtained from the far-ﬁeld amplitude of the antenna and from the orthogonality of the vector spherical harmonics, cf., (3.12), as shown in (A.4), at the bottom of the page, where , and where max is with respect to and . Hence, also the directivity is completely deﬁned by the expansion coefﬁcients. The far-ﬁeld amplitudes and the directivities of the other antennas in Section III-A follow from (3.11) and (3.12). F. Radiation Efﬁciency and Power Gain For antennas in a lossless space the radiation efﬁciency, , is deﬁned as the ratio of the power radiated from the antenna to the power put into the antenna. This deﬁnition is not applicable here since the antenna is ideal and hence, the efﬁciency would be one. A possible alternative deﬁnition for an ideal antenna in , where is the a lossy material is the quotient is the power radiated power radiated from the antenna and through a spherical surface of radius . That ratio indicates how

much of the power fed to the antenna is radiated in the far-zone. In the far-zone

(3.13) for a combined antenna of order , as seen from (3.6) and (A.6). The radiated powers of the other types of antennas are special cases of this expression. In order to have a deﬁnition of radiation efﬁciency that is independent of the radius , the following dimensionless quantity is used (3.14) where (3.15) The radiated power at a distance from an antenna is expressed in terms of the radiation efﬁciency and the input power, , as . The notation is in accordance with most antenna literature. It should not be confused with the notation for the wave impedance.

(3.12)

2030

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

The asymptotic value of

for large radii

is (3.16)

which is in agreement with the graph of the efﬁciency in Fig. 2. The radiation efﬁciencies of multipole antennas of order are and according to (3.6), (3.14), and (3.15) they denoted read as in (3.17), shown at the bottom of the page. The product of the directivity and the radiation efﬁciency, , is proportional to the quotient of the maximum power ﬂow density in the far-zone and the input power to the antenna. It is referred to as the power gain of the antenna and the notation (3.18) is adopted. This deﬁnition is in concordance with the power gain of antennas in lossless media, also referred to as the and maximum value of the gain, cf. [2]. The notations are below used for the power gains of magnetic and electric antennas, respectively.

Fig. 2. Radiation efﬁciency, , of six different multipole antennas. The dashed curves are for the magnetic antennas and the solid curves are for electric antennas. The curves are for l = 1 (upper), l = 2 (middle), and l = 3 (lower). These curves emphasize that the magnetic dipole is the most radiation efﬁcient antenna. When a = 5 mm the magnetic dipole is approximately 10 dB more efﬁcient than the electric dipole. The frequency is 400 MHz, " = 50, and = 1 S/m.

IV. OPTIMIZATION Optimization of an antenna is in this context to ﬁnd the of the radiated partial waves such that a amplitudes speciﬁed quantity is optimized. The techniques used by Chu and Harrington, cf. [4] and [8] can be used to derive the optimal (i.e., maximal) values of the directivity and of the power gain, , of general spherical antennas in a lossy medium. Harrington showed that the optimal value of the directivity for a com, i.e., half of bined antenna of order in vacuum is the number of ports for the antenna. That proof holds also for conductive media. For convenience a derivation of the optimal directivity, analogous to the one given by Harrington, is given in Appendix B. The other derivations are left to the reader. Optimization of the radiation efﬁciency is to minimize the power fed to an antenna for a given power ﬂow in the far-zone, regardless of the directivity. Optimization of the directivity is to maximize the power ﬂow density in one direction in the far-zone, for a given total power ﬂow in the far-zone, regardless of the power fed to the antenna. Optimization of the power gain is to maximize the power ﬂow density in one direction in the far-zone, for a given power fed to the antenna. For a lossy material it can be shown from (3.17) that the op, for any antenna is timal value of the radiation efﬁciency, the one obtained for a magnetic dipole. This is seen in Fig. 2.

In Appendix B it is shown that the optimal directivity of an is electric or magnetic antenna of order

The corresponding value for a combined antenna of order is . For magnetic and electric multipole antennas of order the and , respeccorresponding results for the directivities tively, are (4.1) The relation (4.2) is a result of the fact that sets of partial waves of different index are independent of each other. It is notable that the optimal value of the directivity of an electric or a magnetic multipole antenna of order one, i.e., a dipole antenna, is 1.5. This value is the same as the directivity of each partial wave antenna of order one. For higher order antennas the directivity of a partial wave antenna of order is always smaller than the maximum directivity of the multipole antenna of order .

(3.17)

KARLSSON: PHYSICAL LIMITATIONS OF ANTENNAS IN A LOSSY MEDIUM

2031

The optimal directivity of a general antenna that consists of a combination of independent partial wave antennas, where , has a lower the maximum order of any of the antennas is and upper bound

Equality is only achieved for a combined antenna of order . Notice also that two times the optimal value of the directivity is an upper bound for the number of independent ports an antenna can have. is presented. Using the Next the optimal power gain for a same method as in Appendix B the optimal values of magnetic antenna and an electric antenna of order can be derived

Fig. 3. Power gain G of three magnetic and three electric antennas. The dashed curves are for the magnetic antennas and the solid curves for the electric = 1 (lower), l = 2 (middle), and l = antennas. The curves are for l 3 (upper). The frequency is 400 MHz, " = 50, and = 1 S/m.

(4.3) and are given by (3.17) and (4.1), and where is the optimal power gain of a multipole antenna of order . The optimal value of the power gain of a combined antenna of order equals the sum of the optimal gains of the electric and the , i.e. magnetic antenna of order (4.4) According to (3.16), the asymptotic values as are

by adding higher order multipoles, but for small antennas the improvement compared to the dipole antennas is negligible. , one should Graphs like that in Fig. 3 indicate what order, use for an electric or magnetic antenna. In that way they also indicate the number of useful ports of the antenna. VI. CONCLUSION The main results in the paper are the optimal values of the radiation efﬁciency, the directivity, and the power gain of antennas conﬁned in a lossless sphere. Only ideal antennas are treated in this paper. Real antennas have ohmic losses in the wires that reduce the radiation efﬁciency as well as the power gain. However, that power problem is associated with the actual antenna design and is out of the scope of this paper. A comprehensive study of the design of antennas in lossy materials is found in [9]. The purpose with the optimal values of the radiation efﬁciency, the directivity, and the power gain is to give the antenna designer relative measures and theoretical limitations of the properties of antennas. Optimization of the radiation efﬁciency of an antenna is to minimize the dissipated power for a given power ﬂow in the far-zone. The most radiation efﬁcient antenna is the magnetic dipole. The radius of the sphere should be as large as possible. Optimization of the directivity of an antenna is to maximize the power ﬂow density in one direction in the far-zone for a given total power ﬂow in the far-zone. For an electric antenna the optimal directivity is or magnetic antenna of order and the amplitudes of the radiated partial waves are given by (B.5). The maximum number of ports the antenna can use is twice the optimal directivity. The optimal value of the directivity is independent of frequency and of the material . In theory one can achieve any directivity, even for small in . However, for a small anantennas, by a suitable choice of and tenna the dissipated losses increase very rapidly with it costs a lot of power to obtain high directivity.

(4.5)

V. NUMERICAL EXAMPLES From the formulas in this paper it is straightforward to write short programs that illustrate the difference between the antennas in Section III-A. The three graphs given here are for antennas at 400 MHz, located in a material that is similar to muscles in a body. The conceivable application is implanted devices with wireless communication, even though the inﬁnite is somewhat unrealistic. The conductivity lossy region S/m and the relative permittivity is . In is Fig. 1 the phase of the impedance of six different multipole antennas is plotted as a function of the radius . The argument is 0.37 radians. of the wave impedance of the material in It is seen that the asymptotic values in (3.8) and (3.9) are approached for large and small values of , respectively. In Fig. 2 the radiation efﬁciency is given as a function of for the same six multipole antennas. The ﬁgure clearly shows that for a small radius the magnetic dipole is the mm it is more most efﬁcient antenna. For a radius than 20 dB more efﬁcient than the electric dipole, and 30 dB . In Fig. 3, the better than the magnetic quadrupole is plotted for electric and magnetic antennas power gain , and . One always obtains a larger gain with

2032

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Optimization of the power gain of an antenna is to maximize the power ﬂow density in one direction for a given input power to the antenna. For an electric antenna or magnetic antenna of the optimal power gain is given by (4.3). The power order . A graph like that in Fig. 3 gain increases with increasing . indicates the most suitable value of APPENDIX A VECTOR WAVES The deﬁnition of spherical vector waves can be found in different textbooks, e.g., [6] and [8]. In this paper they are deﬁned using vector spherical harmonics, cf. [1]

APPENDIX B OPTIMAL DIRECTIVITY The optimization problems of ﬁnding the maximum value the directivity, is a multivariable optimization problem. First assume the following function of variables (B.1) where are given real numbers and are given positive real numbers. This function has a maximum when all of its ﬁrst order are zero. That leads derivatives with respect to to the following relations for the variables (B.2) The corresponding maximum value of is (B.3) Now consider electric antennas and magnetic antennas of order and let and denote the corresponding directivities. Without loss of generality the maximum power ﬂow density is assumed to be in some direction given by the spherical angles and , and in that direction the polarization of the corresponding wave is assumed to be parallel to some unit vector . The optimal value of the directivity is independent of the angles and , and of the vector , due to the invariance under rotation described in Section III-B. If , then and are identiﬁed as the real quantities

(A.1) The following deﬁnition of the spherical harmonics is used:

(A.2) where and take the values (A.3) In the current application the index will never take the value 0, since there are no monopole antennas. The vector spherical harmonics constitute an orthogonal set of vector function on the unit sphere (A.4) where the integration is over the unit sphere and where . The outgoing divergence-free spherical vector waves are deﬁned by (A.5), show at the bottom of the page, where is the spherical Hankel function of the second kind. The asymptotic behavior in the far-zone and the limiting values in the near-zone of the spherical Hankel functions are

(B.4) According to (B.2), the optimal directivity is obtained when (B.5) The optimal value of the directivity is given by (B.3) (B.6) Next, is expressed in terms of and . Since is independent of (B.6) in from 0 to . The result is as one may integrate

(B.7) (A.6)

(A.5)

KARLSSON: PHYSICAL LIMITATIONS OF ANTENNAS IN A LOSSY MEDIUM

2033

where . Furthermore, is also independent of and , and the relation above can be integrated over the unit sphere. The orthonormality of the vector spherical harmonics, (A.4), results in (B.8) This is in accordance with the result in [8]. Notice that that (B.5) and (B.6) can be generalized. Assume independent partial wave a general antenna consisting of antennas that are to be fed so that the directivity function , cf. [10], is optimized in a prescribed direction and with a prescribed polarization of the radiated wave. Then a slight modiﬁcation of (B.5) and (B.6) gives the amplitudes of the antennas and the value of the optimal directivity function. It also follows that the mean value of the optimal directivity and , is . Hence, the optimal function, with respect to . value of the directivity is always greater than or equal to REFERENCES

[1] G. Arfken, Mathematical Methods for Physicists, 3rd ed. Orlando, FL: Academic Press, 1985. [2] C. A. Balanis, Antenna Theory, 2nd ed. New York: Wiley, 1997. [3] A. Boström, G. Kristensson, and S. Ström, “Transformation properties of plane, spherical and cylindrical scalar and vector wave functions,” in Field Representations and Introduction to Scattering, V. V. Varadan, A. Lakhtakia, and V. K. Varadan, Eds, Amsterdam: Elsevier Science Publishers, 1991, ch. 4, pp. 165–210.

[4] L. J. Chu, “Physical limitations of omni-directional antennas,” Appl. Phys., vol. 19, pp. 1163–1175, 1948. [5] R. E. Collin, “Minimum Q of small antennas,” J. Elect. Waves Applicat., vol. 12, pp. 1369–1393, 1998. [6] J. E. Hansen, Ed., Spherical Near-Field Antenna Measurements. ser. Number 26 in IEE electromagnetic waves series, Stevenage, U.K.: Peregrinus, 1988, ISBN: 0-86 341-110-X. [7] R. C. Hansen, “Fundamental limitations in antennas,” Proc. IEEE, vol. 69, no. 2, pp. 170–182, 1981. [8] R. F. Harrington, Time Harmonic Electromagnetic Fields. New York: McGraw-Hill, 1961. [9] R. W. P. King and G. S. Smith, Antennas in Matter, Cambridge, London, U.K.: MIT Press, 1981. [10] J. D. Kraus, Antennas, 2nd ed. New York: McGraw-Hill, 1988. [11] H. A. Wheeler, “Fundamental limitations of a small vlf antenna for submarines,” IRE Trans. Antennas Propagat., vol. 6, pp. 123–125, 1958. [12] , “Small antennas,” IEEE Trans. Antennas Propagat., vol. 23, no. 4, pp. 462–469, 1975.

Anders Karlsson was born in 1955, Gothenburg, Sweden. He received the M.Sc. and Ph.D. degrees from Chalmers University of Technology, Gothenburg, Sweden, in 1979 and 1984, respectively. Since 2000, he has been a Professor at the Department of Electroscience, Lund University, Lund, Sweden. His research acivities comprehend scattering and propagation of waves, inverse problems, and time-domain methods. Currently, he is involved in projects concerning propagation of light in blood, wireless communication with implants, and design of passive components on silicon.

2034

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Minimum Norm Mutual Coupling Compensation With Applications in Direction of Arrival Estimation
**

C. K. Edwin Lau, Raviraj S. Adve, Senior Member, IEEE, and Tapan K. Sarkar, Fellow, IEEE

Abstract—This paper introduces a new mutual coupling compensation method based on the minimum norm solution to an underdetermined system of equations. The crucial advantage over previous techniques is that the formulation is valid independent of the type of antenna element and provides good results in situations where signal strengths vary considerably. In using the matrix pencil algorithm to estimate the directions of arrival, the examples show that the proposed method results in signiﬁcantly lower bias than the traditional open circuit method. The analysis of mutual coupling is also applied in the context of a Code Division Multiple Access communication system. Index Terms—Code division multiaccess, direction of arrival estimation, matrix pencil, MUSIC, mutual coupling compensation.

I. INTRODUCTION

D

IRECTION of arrival (DOA) estimation is an important feature of smart antenna arrays. It could serve as a fundamental building block for applications such as space division multiple access (SDMA) and Enhanced 911 (E911), the proposed wireless emergency service [1]. Several algorithms have been proposed for DOA estimation, including the popular MUSIC-type techniques, ESPRIT [1] and matrix pencil (MP) [2]–[4]. These signal processing algorithms have been shown to provide accurate estimates, even in moderate signal to noise (SNR) conditions. The problem is that these signal processing algorithms generally ignore the electromagnetic behavior of the receiving antenna. The receiver is assumed to be an ideal, equispaced, linear array of isotropic point sensors. In this case, the array samples, but does not reradiate the incident signals. Each signal can be associated with a linear phase front, the slope of which is directly related to the DOA. Most signal processing techniques rely heavily on this assumption. In practice, this ideal situation cannot be met. The elements of the array must be of some nonzero size. The elements sample and reradiate the incident ﬁelds, causing mutual coupling. Mutual coupling distorts the linear phase front of the incoming signal, signiﬁcantly degrading performance [5]–[7]. Only in the case of a single incoming signal is the phase front somewhat retained. However,

Manuscript received July 23, 2002; revised September 7, 2003. This research was supported by a grant from the Nortel Institute for Telecommunications, University of Toronto. C. K. E. Lau and R. S. Adve are with the University of Toronto, Toronto, ON M5S 3G4, Canada (e-mail: rsadve@comm.utoronto.ca). T. K. Sarkar is with Syracuse University, Syracuse, NY 13244 USA. Digital Object Identiﬁer 10.1109/TAP.2004.832511

for arrays with strong mutual coupling, the phase front is signiﬁcantly corrupted and the DOA estimate is inaccurate. Any practical implementation of DOA estimation therefore requires compensation for mutual coupling. Research into compensating for the mutual coupling has been based mainly on the idea of using open circuit voltages, ﬁrst proposed by Gupta and Ksienski [5]. The authors argue that due to the lack of a terminal current, the open circuit voltages are free of mutual coupling. However, as shown in [7], this only reduces the effects of mutual coupling. The technique presented there is more effective in suppressing mutual coupling effects [7], [8]. A big drawback with the approaches of [5] and [7], is that they are valid for only linear dipoles. The work of [5] is valid only for a linear array of half wavelength dipoles spaced apart by half a wavelength. The work of [7] is restricted to linear arrays of linear dipoles, though of arbitrary length and spacing. In this paper we introduce the use of a minimum norm technique, based on the technique in [7], for general arrays with arbitrary elements. As an aside we also extend the open circuit technique of [5] to arbitrary arrays. The method of moments (MoM) is used to accurately model the interactions between antenna elements. In the minimum norm approach, the MoM admittance matrix is used to estimate the incident ﬁelds, with minimum energy, that would generate the received voltages. Unlike in [9], this technique does not require the solution to the entire MoM problem. The compensation matrix depends only on the MoM admittance matrix and can be calculated a priori to reduce computation load. In this paper, we use the MP [2] and the popular MUSIC [1], [10] DOA estimation algorithms to compare various compensation methods. Section II presents the model for mutual coupling using antenna analysis based on the MoM. This eventually leads to the formulation of minimum norm mutual coupling compensation method. Section III presents examples illustrating the performance of the open circuit and the minimum norm methods in case of a equispaced, linear array of dipoles. Section IV ends with some conclusions and a summary of the contributions presented here. II. MUTUAL COUPLING AND COMPENSATION Most DOA estimation algorithms including MP and MUSIC assume an ideal, linear array of isotropic sensors. Unfortunately, such an ideal sensor is clearly not realizable. A practical antenna array comprises elements of some physical size. Such elements

0018-926X/04$20.00 © 2004 IEEE

LAU et al.: MINIMUM NORM MUTUAL COUPLING COMPENSATION WITH APPLICATIONS

2035

sample and reradiate incident ﬁelds that interact with other elements, i.e., the elements are mutually coupled. Mutual coupling severely degrades the accuracy of the DOA estimator [6]. Any implementation of DOA estimation must account for the mutual coupling between elements. In a practical antenna array, the received signals are the voltages measured across the load at the port of each element. To deal with mutual coupling, researchers originally proposed processing these measured voltages to obtain the open circuit voltages, the voltages if all the ports were open circuited [5], [6]. Open circuiting the ports reduces the currents on the elements, consequently the reradiated ﬁelds and therefore the mutual coupling. However, as shown in [7], this methodology is valid only when all signals have similar strengths. In [7], we use a MoM analysis to compensate for mutual coupling. That technique is very effective, but is valid only for a linear array of parallel dipoles. We present here a technique that is theoretically valid for all kinds of arrays. Based on a minimum norm solution to an undetermined system of MoM equations, the technique makes no assumptions regarding the type of antenna, or the spacing between elements. However, for simplicity, this methodology is presented here for a linear array of dipoles. We begin with a brief review of the analysis technique, as the MoM analysis for dipole arrays is well known [7], [11]. The review included here sets the stage for the minimum norm solution. A. System Model We begin with the general formulation of the MoM based on subdomain basis functions for a receiving antenna array of -elements. The central assumption is that only a single basis function contributes to the current at the port of each element in the array. The incident electric ﬁeld is related to the currents on the antenna through a linear operator [12] (1) The current is approximated by a set of subdomain basis functions, , with basis functions per element, i.e. (2) testing where is the th current coefﬁcient. Using a set of , and a convenient deﬁnition of inner product, (1) functions, can be reduced to a matrix equation (3) (4) where the the th element of matrix and the are (5) (6) The matrix is the matrix with zero endiagonal entries, where the th tries other than the th element of

Fig. 1.

Linear array of wire dipoles terminated in loads Z .

entry corresponds to the port. This th entry is the load impedance of the corresponding element. The matrix is the MoM impedance matrix. Assuming a single basis function corresponds to the current at a port of the array, from (4), the measured voltages, affected by mutual coupling are given by (7) is the submatrix of correThe matrix , a compressed version of , is sponding to the ports. diagonal matrix of port impedances. is a the matrix of dimensionless entries. Note that the entries of are directly related to the incident ﬁelds and are free of mutual coupling. In this paper, this general formulation is applied to a linear array of dipoles. It must be emphasized that this choice is not fundamental to the theoretical development here and is made only for purposes of illustration. Consider a wire dipole antenna -directed elements as shown in Fig. 1. Each element array of . To anhas a centrally located port terminated in impedance alyze this array we use sinusoidal basis functions. Each element segments of equal length. To satisfy the is divided into requirement that only a single basis function corresponds to the current on the array is chosen to be odd. Based on a Galerkin formulation, the weighting and testing functions are the same. The entries for the MoM voltage and impedance matrices are available in [7], [11]. B. Open Circuit Voltages The principal idea of [5] is to use the open circuit voltages instead of the measured voltages for further signal processing. However, the theory is valid only for half wavelength dipoles with half wavelength spacing. In the more general case, one can use the MoM analysis in conjunction with the Thevenin and Norton equivalent circuits to obtain the open circuit voltages. to be the inverse of Deﬁne the MoM admittance matrix . Note that this is not the same mathe impedance matrix trix, , in (4). Also deﬁne a new matrix whose entries are those rows and columns of the MoM admittance mathat correspond to ports, i.e. trix (8) (9) (10)

2036

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Example 3.1. MP, using uncompensated voltages.

Fig. 3. Example 3.1. MP, using open circuit voltages.

The open circuit voltages are then related to the short circuit currents as (11) and the measured voltages to the short circuit currents as (12) Eliminating the short circuit currents from (11) and (12) yields the open circuit voltages (13) In the following sections the open circuit voltages we refer to are obtained from the measured voltages using (13). C. Minimum Norm Compensation Formulation As shown in [7], using the open circuit voltages only somewhat reduces the effects of mutual coupling. In [7], we reconstruct a part of the MoM voltage vector under the assumption of a linear dipole array. The motivation comes from the fact that, from (5), the MoM voltages are directly related to the incident ﬁelds and so are free of mutual coupling. In the general case, from (7), the equation relating the measured and MoM voltages is underdetermined and the cannot be reconstructed exactly. However, one can ﬁnd the minimum norm solution to this equation. This solution provides the vector with the minimum two-norm (minimum energy) that would result in the received voltages. The resulting vector is an estimate of the MoM voltage vector. Using (7), the minimum norm solution to the MoM voltage vector is (14) is the conjugate transpose (Hermitian) of matrix . where corresponding to the ports may be used for Entries in further signal processing.

Fig. 4. Example 3.1. MP, using minimum norm compensation.

Physically, the compensation procedure may be interpreted as ﬁnding the signal with minimum energy that results in the measured voltages. Since the MoM analysis and so the matrix may be obtained a priori, the computation load to use (14) is no greater than ﬁnding the open circuit voltages or using the technique of [7]. In the following section, we compare the performance of the two compensation methods in various settings. III. NUMERICAL EXAMPLES In this section, we present numerical examples to illustrate the workings of the two compensation techniques, the open circuit and minimum norm methods. The application here is DOA estimation. The ﬁrst two examples deal with the DOA estimation of multiple signals and demonstrates the impact of mutual coupling and compares the two compensation techniques. The third example deals with the impact of mutual coupling on code division multiple access (CDMA) communications in particular, and the effectiveness of mutual coupling compensation in this

LAU et al.: MINIMUM NORM MUTUAL COUPLING COMPENSATION WITH APPLICATIONS

2037

TABLE I COMPARING OPEN CIRCUIT AND MINIMUM NORM TECHNIQUES. EQUAL SIGNAL STRENGTHS

case. Due to mutual coupling, the signal level at each element may be different. The SNR is deﬁned here as the average SNR at all ports of the array, i.e., in adding white, complex Gaussian noise at each element, the power level is chosen to set an average SNR. In all examples using MP, the pencil parameter is . set to A. Three Signals of Equal Strength This example uses a seven element array with interelement . The MoM analysis uses 7 unknowns per elespacing of ment, i.e., a total of 49 unknowns are used. The array receives three signals from 40 , 90 and 140 . Each signal has a SNR of 1 dB. The MP algorithm uses only a single snapshot. The plots shown here use the results of 1000 independent trials. Fig. 2 shows a histogram of the results of using MP without any compensation for mutual coupling. 38 times, the estimation procedure fails completely by resulting in imaginary angles. This happens because MP estimates the complex phase before estimating the direction . In 38 instances, function becomes greater than 1. the argument to the As is clearly seen in the ﬁgure, the DOA estimation is very poor with very large errors. Figs. 3 and 4 plots the performance after compensation for mutual coupling. Fig. 3 plots the use of open circuit voltages while Fig. 4 plots the results of using the minimum norm technique. In both ﬁgures, the hugely improved performance over the uncompensated case is very clear. Neither technique results in any imaginary angles. Note that because of the accurate performance, we can estimate a standard deviation, which for all cases is approximately 3.5 . As Table I shows, the crucial difference between the two compensation techniques is in the bias. The bias resulting from using the minimum norm compensation approach is signiﬁcantly smaller than using the open circuit voltages. This is because using the open circuit voltages only implies the lack of a terminal current. Physically, there is still a nonzero current on the dipole arms. These currents reradiate, resulting in some residual mutual coupling. Fig. 5 explains the improved performance of the minimum norm technique over the open circuit approach. The ﬁgure plots of the phase front of the three incoming signals in the various scenarios of this example. It plots the phase at each element in the ideal case, in the case of no mutual coupling compensation, using the open circuit approach and using the minimum norm solution. Without compensation, the phase information is signiﬁcantly corrupted, explaining the erroneous results. Both compensation techniques correct this somewhat. However, clearly the minimum norm solution is better than

Fig. 5.

Example 3.1. Phase front of three incoming signals.

Fig. 6. Example 3.2. MP, using uncompensated voltages.

using the open circuit voltages. This explains, from the phase point of view, why the two compensation methods work and why the proposed approach is better than the traditional open circuit approach. B. Three Signals of Unequal Strength In this example we use the same array as in the ﬁrst example with the the signal bearings at 40 , 70 , and 140 , with SNR’s of 7, 15, and 5 dB respectively. We use 10 000 independent trials.

2038

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 7.

Example 3.2. MP, using open circuit voltages.

Fig. 9.

Example 3.2. MUSIC.

Fig. 8.

Example 3.2. MP, using minimum norm compensation. Fig. 10. Example 3.2. MUSIC. The middle signal is at 60 .

Fig. 6 shows a histogram of MP estimate without any mutual coupling compensation. 2269 estimates result in imaginary angles. Clearly the remaining estimates are not of any practical use. Figs. 7 and 8 are results of using MP compensated with the open circuit and minimum norm approaches respectively. Both compensation methods improve the estimation dramatically. All the imaginary angles are recovered. Similar to the previous example, the open circuit method exhibits a larger bias than the minimum norm approach. The bias is even stronger in this example than the last one as the signal at 70 is relatively strong and closer to the 40 signal. Fig. 9 shows the pseudo-spectrum generated by MUSIC without compensation, using the open circuit voltages and with minimum norm compensation. In all cases, 15 time samples are used to estimate the covariance matrix. As can be seen, with either compensation technique, the resolution improves and the bias is reduced. Again, the bias in the estimation is less with minimum norm method than that with open circuit method. This is in agreement with the examples presented

for the MP algorithm. If the signal at 70 is moved to 60 , as shown in Fig. 10, the results are even more dramatic. If no compensation is used, the signals at 40 and 60 merge. But after the compensation, the two spikes are recovered. Again, using the open circuit voltages results in a greater bias than the minimum norm method. Figs. 11 and 12 show the results if the strength of the signal at 70 is increased to 25 dB. The results are shown in . We can see from the ﬁgures that the bias in not signiﬁcant when using the minimum norm method in Fig. 12. When using the open circuit method, the bias in the weaker signals is 2 and 3.7 . Table II summarizes our statistical ﬁndings of this example. C. Mutual Coupling Compensation in CDMA Communications One motivation for this research is position location in wireless communication systems. Here we focus on a CDMA system. In applying the MP technique to a practical array in a CDMA based communication setting, a curious fact emerges.

LAU et al.: MINIMUM NORM MUTUAL COUPLING COMPENSATION WITH APPLICATIONS

2039

Fig. 11.

Example 3.2. MP, using open circuit voltages. Signal at 70 is 25 dB.

Fig. 12. Example 3.2. MP, using minimum norm compensation. Signal at 70 is 25 dB.

The CDMA processing gain provides some resistance to mutual coupling. This is because, after the matched ﬁlter, there is effectively only one signal plus relatively weak residual interference. With only one signal impinging on the array, the linear phase front is not fatally corrupted and it is possible to estimate the DOA. This is true particularly of arrays with moderate mutual coupling. To illustrate this effect, we use the same example as in Section III-A. However, each signal is spread with a spreading gain of 128. We use four signal samples per chip. For a fair comparison, the power of each signal is reduced by the spreading gain. Using the ﬁlter matched to the ﬁrst signal, two of three signals are suppressed. Note that in using MP to estimate the DOA of this signal after the matched ﬁlter, we set the number of signals . This also eliminates a drawback associated to one, i.e., with MP, the restriction on the number of signals that can be estimated simultaneously [13]. MP is applied without compensating for mutual coupling. Fig. 13 plots the histogram of the resulting estimates. In comparison to Fig. 2 the accuracy is dramatically improved. No estimate results in imaginary angles. In fact, the accuracy is comparable to using the open circuit voltages as in Fig. 3. It must be emphasized that this resistance to mutual coupling is only an approximation. Depending on the accuracy required, compensation for mutual coupling can still play an important role. Fig. 14 plots the results of using the minimum norm approach. The performance is improved with signiﬁcantly reduced bias.

Fig. 13. Example 3.3. CDMA/MP, using uncompensated voltages.

IV. CONCLUSION Practical implementations of DOA estimation must deal with the problem of mutual coupling between antenna elements. The work of [7] introduced the concept of reconstructing a part of the MoM voltage vector. We extend this concept here and develop a very effective technique based on the minimum norm solution to an underdetermined system of equations. The approach is to ﬁnd the signals, with minimum energy, that would

create the mutually coupled measured signals. The overhead associated with the compensation procedure is limited to a matrix multiplication. In testing the proposed approach, the technique proves to be more accurate than the classical open circuit approach. The minimum norm technique reduces the bias in the estimates because the phase response is reconstructed more accurately than when using open circuit voltages. In applying DOA estimation speciﬁcally to CDMA communications a curious fact emerges. If DOA estimation is applied after the matched ﬁlter, the CDMA spreading gain results in the desired signal plus residual interference. The phase front of a single signal is not signiﬁcantly corrupted and so the resulting DOA estimation, without compensation is fairly accurate. However, one cannot conclude that mutual coupling compensation is not required. Applying compensation further improves performance. Since the additional cost is restricted to a matrix multiplication, the resulting performance gains would probably outweigh the cost of implementing mutual coupling compensation.

2040

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE II COMPARING OPEN CIRCUIT AND MINIMUM NORM TECHNIQUES. UNEQUAL SIGNAL STRENGTHS.

[6] C.-C. Yeh, M.-L. Leou, and D. R. Ucci, “Bearing estimations with mutual coupling present,” IEEE Trans. Antennas Propagat., vol. 37, pp. 1332–5, Oct. 1989. [7] R. S. Adve and T. K. Sarkar, “Compensation for the effects of mutual coupling on direct data domain algorithms,” IEEE Trans. Antennas Propagat., vol. 48, pp. 86–94, Jan. 2000. [8] M. Ali and P. Wahid, “Analysis of mutual coupling effect in adaptive array antennas,” in Proc. IEEE Antennas and Propagation Soc. Int. Symp., June 2002. [9] K. M. Pasala and E. M. Friel, “Mutual coupling effects and their reduction in wideband direction of arrival estimation,” IEEE Trans. Aerospace and Electron. Syst., vol. 30, pp. 1116–1122, Apr. 1994. [10] R. O. Schmidt, “Multiple emitter location and signal parameter estimation,” IEEE Trans. Antennas Propagat., vol. 34, pp. 276–290, Mar. 1986. [11] B. J. Strait, T. K. Sarkar, and D. C. Kuo, “Special programs for analysis of radiation by wire antennas,” Syracuse Univ., Tech. Rep. AFCRL-TR-73-0399, 1973. [12] R. F. Harrington, Field Computation by Moment Methods. Melbourne, FL: Kreiger, 1982. [13] C. K. E. Lau, R. S. Adve, and T. K. Sarkar, “Combined CDMA and matrix pencil direction of arrival estimation,” in Proc. IEEE Vehicular Technology Conf., 2002, pp. 496–499. Fig. 14. Example 3.3. CDMA/MP, using minimum norm compensation. Edwin C. K. Lau received the B.A.Sc. and M.A.Sc. degrees, both in electrical engineering, from the University of Toronto, Toronto, ON, Canada, in 2000 and 2003, respectively. He is one of the participants of the Communications and Information Technology Ontario (CITO) Research Partnerships Program. His area of research includes retrodirective antennas, microwave circuit and antenna design, and direction of arrival estimation algorithm.

In summary, we have presented a practical and accurate minimum norm mutual coupling compensation method. The new approach proves to more accurate than the traditional open circuit approach. This method can theoretically also be applied to arrays of arbitrary elements.

REFERENCES

[1] J. C. Liberti Jr and T. S. Rappaport, Smart Antennas for Wireless Communications: IS-95 and Third Generation CDMA Applications. Englewood Cliffs, NJ: Prentice Hall, 99. [2] Y. Hua and T. K. Sarkar, “Matrix pencil method for estimation parameters of exponentially damped/undamped sinusoids in noise,” IEEE Trans. Acoust. Speech and Signal Processing, vol. 38, pp. 814–24, May 1990. [3] J. E. F. del Rio and T. K. Sarkar, “Comparison between the matrix pencil method and the Fourier transform for high-resolution spectral estimation,” Digital Signal Processing: A Review Journal, vol. 6, pp. 108–125, 1996. [4] R. S. Adve, O. M. Pereira-Filho, T. K. Sarkar, and S. M. Rao, “Extrapolation of time domain responses from three dimensional objects utilizing the matrix pencil technique,” IEEE Trans. Antennas Propagat., vol. 45, pp. 147–156, Jan. 1997. [5] I. J. Gupta and A. A. Ksienski, “Effect of mutual coupling on the performance of adaptive array,” IEEE Trans. Antennas Propagat., vol. 31, pp. 785–91, Sept. 1983. Raviraj S. Adve (S’88–M’97–SM’03) received the B.Tech. from the Indian Institute of Technology, Bombay, in 1990 and the Ph.D. degree from Syracuse University, Syracuse, NY, in 1996, all in electrical engineering. His dissertation, on the impact of mutual coupling on the performance of adaptive antenna arrays, received the Syracuse University “Outstanding Dissertation Award” in 1997. From 1997 to August 2000, he was a Senior Research Engineer with Research Associates for Defense Conversion (RADC) Inc., Marcy, NY, working on contract with the Air Force Research Laboratory, Sensors Directorate, Signal Processing Branch, Rome, NY. He is currently an Assistant Professor in the Department of Electrical and Computer Engineering, University of Toronto. He has also been a consultant to Stiefvater Consultants. His research interests are in practical adaptive signal processing algorithms for wireless communication and airborne radar systems.

LAU et al.: MINIMUM NORM MUTUAL COUPLING COMPENSATION WITH APPLICATIONS

2041

Tapan K. Sarkar (S’69–M’76–SM’81–F’92) received the B.Tech. degree from the Indian Institute of Technology, Kharagpur, in 1969, the M.Sc.E. degree from the University of New Brunswick, Fredericton, NB, Canada, in 1971, and the M.S. and Ph.D. degrees from Syracuse University, Syracuse, NY, in 1975. From 1975 to 1976, he was with the TACO Division, General Instruments Corporation. He was with the Rochester Institute of Technology, Rochester, NY, from 1976 to 1985. He was a Research Fellow at the Gordon McKay Laboratory, Harvard University, Cambridge, MA, from 1977 to 1978. He is now a Professor in the Department of Electrical and Computer Engineering, Syracuse University. He has authored or coauthored more than 210 journal articles and numerous conference papers and has written 28 chapters in books and ten books, including his most recent, Iterative and Self Adaptive Finite-Elements in Electromagnetic Modeling (Boston, MA: Artech House, 1998). His current research interests deal with numerical solutions of operator equations arising in electromagnetics and signal processing with application to system design. Dr. Sarkar is a Registered Professional Engineer in the State of New York. He is a member of Sigma Xi and the International Union of Radio Science Commissions A and B. He received one of the ”best solution” awards in May 1977 at the Rome Air Development Center (RADC) Spectral Estimation Workshop. He received the Best Paper Award of the IEEE Transactions on Electromagnetic Compatibility in 1979 and in the 1997 National Radar Conference. He received the College of Engineering Research Award in 1996 and the Chancellor’s Citation for Excellence in Research in 1998 at Syracuse University. He received the title Docteur Honoris Causa from Universite Blaise Pascal, Clermont Ferrand, France in 1998 and the medal of the city of Clermont Ferrand, France, in 2000. He was an Associate Editor for feature articles of the IEEE Antennas and Propagation Society Newsletter, and he was the Technical Program Chairman for the 1988 IEEE Antennas and Propagation Society International Symposium and URSI Radio Science Meeting. He is on the editorial board of Journal of Electromagnetic Waves and Applications and Microwave and Optical Technology Letters. He has been appointed a U.S. Research Council Representative to many URSI General Assemblies. He was the Chairman of the Intercommission Working Group of International URSI on Time Domain Metrology from 1990 to 1996.

2042

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**A Phase-Space Beam Summation Formulation for Ultrawide-Band Radiation
**

Amir Shlivinski, Member, IEEE, Ehud Heyman, Fellow, IEEE, Amir Boag, Senior Member, IEEE, and Christine Letrou, Member, IEEE

I. INTRODUCTION

Abstract—A new discrete phase space Gaussian beam (GB) summation representation for ultrawide-band (UWB) radiation from an aperture source distribution is presented. The formulation is based on the theory of the windowed Fourier transform (WFT) frames, wherein we introduce a novel relation between the frequency and the frame overcompleteness. With this procedure, the discrete lattice of beams that are emitted by the aperture satisﬁes the main requirement of being frequency independent, so that only a single set of beams needs to be traced through the medium for all the frequencies in the band. It is also shown that a properly tuned class of iso-diffracting (ID) Gaussian-windows provides the “snuggest” frame representation for all frequencies, thus generating stable and localized expansion coefﬁcients. Furthermore, due to the ID property, the resulting GBs propagators are fully described by frequency independent matrices whose calculation in the ambient environment need to be done only once for all frequencies. Consequently, the theory may also be expressed directly in the time-domain as will be presented elsewhere. The localization implied by the new formulation is demonstrated numerically for an UWB focused aperture. It is shown that the algorithm extracts the local radiation properties of the aperture source and enhances only those beams that conform with these properties, i.e., those residing near the phase space Lagrange manifold. Further localization is due to the fact the algorithm accounts only for beams that pass within a few beamwidths vicinity of the observation point. It is thus shown that the total number of beams is much smaller than the Landau Pollak bound on the aperture’s degrees of freedom. Index Terms—Beam summation representations, frame theory, Gaussian beams (GBs), phase space, ultrawide-band (UWB), windowed Fourier transform (WFT).

B

NOMENCLATURE GB ID UWB WFT Gaussian beam. isodiffracting. ultrawide-band. windowed Fourier transform.

Manuscript received November 21, 2001; revised June 12, 2003. The work of E. Heyman was supported in part by the Israel Science Foundation under Grant 216/02 and in part by the Air Force Ofﬁce of Scientiﬁc Research (AFOSR) under Grant F49620-01-C-0018. The work of A. Boag was supported in part by the Israel Science Foundation under Grant 577/00. A. Shlivinski was with the School of Electrical Engineering, Tel Aviv University, Tel Aviv 69978, Israel. He is now with the Department of Electrical Engineering, University of Kassel, 34109 Kassel, Germany. E. Heyman and A. Boag are with the School of Electrical Engineering Tel Aviv University, Tel Aviv 69978, Israel. C. Letrou is with GET/INT, CNRS SAMOVAR, UMR 5157, France. Digital Object Identiﬁer 10.1109/TAP.2004.832513

EAM based phase-space formulations are an important tool in the wave theory since they provide a systematic framework for ray-based construction of spectrally uniform local solutions in complex conﬁgurations [1]–[3]. In these formulations, the ﬁeld is expanded into a phase-space spectrum of beams that emanate at a given set of points and directions in the source domain, and thereafter are tracked locally in the medium (cf. Fig. 1). The advantages of the beam formulations over the more traditional representations are: 1) unlike the plane waves, the beam propagators can be tracked locally in inhomogeneous media or through interactions with interfaces, and unlike rays, they are insensitive to the geometrical optics (GOs) transition zones; 2) the formulations are a priori localized in the vicinity of the phase-space skeleton of GOs (the so-called Lagrange manifold; see Section IV-C) since only those beam propagators that pass near the observation point actually contribute there. Thus, beam representations combine the algorithmical ease of GOs with the uniform features of spectral representations, and therefore have been used recently in various applications [4]. An important property of these formulations is that the spectrum of beam propagators is overcomplete and thus may be a priori discretized as, for example, in the Gabor series representation. This attractive feature has led to the utilization of Gabor-based beam algorithms in various applications involving radiation, scattering and inverse scattering in complex environments [1], [5]–[12]. The Gabor representation is critically complete, i.e., the phase-space grid is constrained by the Gabor condition , where and are the step sizes of the spatial and spectral discretization [see (9)]. Consequently, it suffers from two inherent difﬁculties: a) Nonlocality and instability of the expansion coefﬁcients: The “analysis function” (the dual or biorthogonal function), which is used to calculate the expansion coefﬁcients tends to be highly nonlocal and irregular, and in certain cases it is even not integrable (see Fig. 3(f) and [13]–[17]), hence, the Gabor coefﬁcients represent a nonlocal and unstable sampling of the ﬁeld [1], [3]. b) Frequency-dependent beam lattice: The beam lattice obtained in the Gabor representation changes with frequency [18], hence a different set of beams needs to be tracked for each frequency (as opposed to the situation in Fig. 1). This difﬁculty stems from the Gabor condition on the phase space grid: choosing to be constant for all frequencies (i.e., constant beam initiation points) results in frequency dependent beam directions, and vice versa [see discussion accompanying (20), (21)].

0018-926X/04$20.00 © 2004 IEEE

SHLIVINSKI et al.: PHASE-SPACE BEAM SUMMATION FORMULATION FOR ULTRAWIDE-BAND RADIATION

2043

Fig. 2. Coverage of an overcomplete phase space grid with = 0:5 using the Gaussian window (14) with = 0:5 ; 1; and 2 (dotted, solid, and dash-dotted lines, respectively). Heavy dots: Grid points.

Fig. 1. Discrete phase space beam summation representation for wideband radiation from an extended aperture source.

Difﬁculty (b) makes the conventional Gabor-based beam formulation inapplicable for UWB applications. Difﬁculty (a) makes it inconvenient even for monochromatic ﬁelds. For monochromatic applications, difﬁculty (a) has been circumvented recently by using a frame-based beam summation representation [19], [20]. The overcomplete nature of this representation smoothes out and localizes the dual function, ending up with stable and local coefﬁcients at the expense of having to calculate more coefﬁcients and trace more beam propagators. This poses a tradeoff in the choice of the oversampling ratio versus the stability and localization of the representation. A reasonable solution has been found at an oversampling of order 4/3 or larger for one dimensional (1-D) apertures. In this paper, we introduce a novel scheme wherein the frame formulation accommodates the difﬁculty under (b) for UWB ﬁelds [21], [22]. The scheme is based on the fact that the overcomplete frame removes the Gabor constraint and hence by a proper scaling of the overcompletness with the frequency (see Section III-A) one may construct a frequency independent beam lattice, so that the same set of beams is used for the entire relevant frequency spectrum as schematized in Fig. 1. In Section IV, it is shown further that the iso-diffracting Gaussian beams (ID-GBs) provide the “snuggest” frame basis for all frequencies (thus providing local and stable coefﬁcients). These windows, which in fact have been introduced in a different context [23]–[26], also simplify the beam calculations since, the resulting ID-GB propagators are fully described by frequency independent matrices whose calculation in the ambient environment need to be done only once for all frequencies. Consequently, these beams can be transformed in closed form into the time domain, where they describe the so-called ID pulsed beams [23]–[28]. Based on this property, we have also introduced in [29] a new discrete phase space beam summation representation for short-pulse ﬁelds directly in the time domain (full papers will be published elsewhere). The paper is organized as follows: Section II reviews the relevant elements of frame theory, starting with general frames (Section II-A) and then concentrating on the windowed Fourier

transform (WFT) frames that form the basis of the beam summation representation (Sections II-B and II-C). Using analysis and numerical examples, we identify the relevant frame parameters for a snug and stable representation. The UWB beam summation representation is presented in Section III, starting with the formulation of the frequency independent beam lattice (Section III-A) followed by the parameterization of the ID-GB windows for an UWB snuggest representation and a thorough analysis of the algorithm (Section IV). The localization and frequency independence issues implied by the new formulation are demonstrated numerically in Section V for an UWB focused aperture. It is shown that the algorithm extracts the local radiation properties of the aperture source and enhances only those beams that conform with these properties, i.e., those residing near the phase space Lagrange manifold. The presentation concludes in Section VI with an extensive summary of the algorithm, the various considerations in choosing the expansion parameters. II. ELEMENTS OF FRAME FORMULATIONS This section presents a brief review of the relevant theory of frames, starting in Section II-A with general frames and then proceeding in Section II-B with a detailed analysis of the WFT frames that form the basis of our beam summation representation. Extensive treatments of frame theory can be found in [16], [17], yet the results are presented here with a new slant that is more relevant to our UWB wave analysis. A. Frames, Dual Frames and Frame Representation of Signals The theory of frames has been introduced originally in [30], but it has gained renewed interest recently since it provides a framework for an advanced phase-space signal processing [16], [17]. Deﬁnition [17, Sec. 3.2]: A family of functions , , is called a frame if there exist “frame in a Hilbert space bounds” such that for all (1) In the context of this paper, we shall be interested in the function with the inner product for space , where the asterisk denotes complex conjugate. For the

2044

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 3. Exact and approximate dual functions '(x) and ' (x) (solid and dotted lines) corresponding to the Gaussian window (14) for 6 different values of . In all cases, the phase space grid has been chosen to match to the window according to (16), and the unit cell boundaries at x=2 are depicted as two vertical lines.

6

2-D aperture distributions considered later on in this paper, the frame will be obtained by a Cartesian multiplication of the 1-D frames in each of the Cartesian coordinates. is overcomplete, thus, it is not orthogGenerally, the set onal and not even a basis. It does not have to be normalized and the bounds and may depend on the relative magnitude of the elements in . The frame operator is deﬁned as (2) Clearly is self adjoint. Rephrasing (1) in operator conventions, yields the bounds on , [17, Sec. 3.2], where is the identity operator. Since has a lower bound , it has an inverse. Applying to , yields the set (3) which is also a frame with bounds , and frame operator [17, Proposition 3.2.3]. is termed the dual frame and is the dual frame operator. A frame representation for is given by [17, eq. 3.2.8] (4) with a dual expression using . It should be noted that the frame representation (4) is not unique, i.e., there are other sets of coefﬁcients that can be used in (4) instead of and still express , yet the latter minimizes the norm of the coefﬁcient series [17, Proposition 3.2.4].

The frame representation requires a choice of an appropriate for a given application, and a calculation of its frame dual via (3). The computations involve the inversion of , which can be performed via several methods, e.g., an iterative Neumann series procedure [17, Sec. 3.2] or a projection of on a ﬁnite Hilbert space (e.g., a set of sufﬁciently dense sampling points) discussed in the Appendix . The frame calculations to be presented below have been performed via the latter approach, yet, we shall brieﬂy comment on the former since it explains some properties of frames, which will be used in our UWB beam representation of Sections III and IV. where and Recasting as is a constant, leads to a Neumann series expansion of (5) This series converges if verges if mizes . Since from (1) , it follows that the series in (5) con[31]. The optimal value that miniis readily found to be (6) tends to 1 so that The series converges fast if the ratio tends to 0. Such frame is called snug, whereas if the frame is tight [17]. For a snug frame, one may use the term in (5), i.e. (7)

SHLIVINSKI et al.: PHASE-SPACE BEAM SUMMATION FORMULATION FOR ULTRAWIDE-BAND RADIATION

2045

In general however, and are not known analytically, hence for each frequency may be time conthe calculation of suming. In our UWB formulation, we shall therefore choose a parameter range where one can use an analytic approximation [see (11)]. This not only simpliﬁes the frequency for and domain calculations but also leads to closed form expressions in the time domain theory [29]. B. WFT Frames In two types of frames are mostly used: the wavelet frames and the WFT frames (also termed Weyl-Heisenberg frames or Gabor frames). In this paper, we utilize the latter. , the eleChoosing a proper window function ments of the frame and of its dual are given by [16], [17] (8a) (8b) . where we use conveniently the vector index The parameters deﬁne the spatial and spectral displaceis centered at the lattice point in ment units, i.e, phase-space. For the set to constitute a frame, the it is necessary that the unit-cell area be smaller than , i.e. (9) The parameter describes the overcompleteness or the redunis the oversampling factor). The frame dancy of the frame ( is overcomplete for and it is critically complete in the , where it becomes a basis. A necessary conGabor limit dition on [17, Proposition 3.4.1] basically states that the phase space should be covered without “holes”, i.e., and all its translations should provide a full coverage of the real axis with . no gaps, with a similar condition on its Fourier transform needs to be calculated for a given The dual window and a phase space lattice . One numerical approach is outlined in the Appendix . An exact reference solution can also be computed via the Zak transform for rational oversampling [33]. Though not used in computations below, the Neumann series approach (5)–(7) is helpful here since it explains properties that will be employed in constructing the UWB beam representation of (6) involves of Sections III and IV. In general, ﬁnding and (see elaborate calculations for the frame bounds Appendix). For WFT frames, we may use the bounds [17, eq.. 3.4.2] (10) to ﬁnd an approximation to (6): , so that the term in the Neumann series (5) for becomes (cf. (7)) (11) is not only the limit of for , but it also approximates over a wide range of provided that the window is matched to the lattice over that range as discussed in (12). We over the “optimal” value in (7) since unlike prefer (11) with

, (11) is known analytically and does not require numerical calculations; this becomes essential when the frame expansion needs to be calculated for is used for UWB ﬁelds where each frequency in the band. Equation (11) also yields simple analytic expressions when the formulation is transformed to obtain a beam representation in the time domain [29]. The properties of the frame and of its dual depend on two main parameters: the “overcompleteness” parameter of (9) and the “matching” parameter that describes how well the spatial and spectral distributions of ﬁt into the phase space lattice. It is deﬁned as (12) and are the spatial and spectral widths , the spatial and spectral window coverage of . When are balanced, but if or , of the unit cell the window is spatially narrow or wide, respectively (i.e., spectrally wide or narrow), with respect to the unit cell (see Fig. 2). As we shall show in Figs. 3 and 5, for a given , the snuggest frame and, thus, the most localized dual function are obtained when . Finally, in view of (4), the WFT frame representation for is given by (13a) where

(13b) From (13b), are samples of the WFT of with respect to , the “analysis function” at the phase space points while (13a) synthesizes in terms of all the phase space translations of the “synthesis function” . In general, it is desired that and its dual be localized both spatially and spectrally, so that (13a) expresses in term with local and stable of localized phase space constituents coefﬁcients . Balian-Low theorem [17, Theorem 4.1.1] and [32] implies in this case is required to be strictly less than 1. C. Special Case: Gaussian Windows We consider the Gaussian window (14) which is normalized to . Noting that the spatial and and , respecspectral widths are tively, the matching parameter (12) is found to be (15) where the second expression follows via (9). The cell dimensions that “match” this window are obtained by setting , giving (16)

2046

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 4. Relative error ' ' = ' between the exact ' and its approximation ' of (11), as a function of .

k 0

kk k

Fig. 5. Exact dual function ' corresponding to the Gaussian window (14) for = 0:5 with three values of : = 0:5 ; 1; and 2 (dotted, solid and dash-dotted lines, respectively).

The dual function corresponding to (14) is depicted in Fig. 3 for a wide range of the overcompleteness parameter . For each , the cell dimensions have been chosen via (16) to “match” the window, and the resulting cell’s boundaries at are shown as vertical lines. The exact functions , calculated by the method in the Appendix , are shown by the full lines, while the dotted of (11), which basically lines are the approximate functions replicate up to a proportionality constant. One observes that , provides a good approximation of , in the range , becomes increasingly less localized and while for it tends to the Gabor biorthogonal function in the limit [1], [13], [14], which is seen to be nonlocal and irregular. The quality of the approximation is further examined in Fig. 4 which as a function of . depicts the relative error The role of the matching parameter in (12) is explored in Fig. 5, which depicts the numerically calculated for and for three values of : , 1, and 2 (i.e., have been chosen such that is spatially narrow, matched, and wide with respect to the lattice, respectively; cf. Fig. 2). As expected, . This the smoothest and most localized is obtained for observation is further supported by Fig. 6, which shows that for [see (6)] a given the frame is snuggest with minimal when . III. WIDEBAND BEAM SUMMATION FORMULATION The WFT frame discussed above is used now to construct the UWB discrete beam summation representation. Such representation should satisfy the following requirements: (a) It should utilize a frequency independent beam lattice so that the beam axes do not have to be retraced for each frequency. (b) The window function should properly scale with the frequency to provide the snuggest frame for the entire frequency band. (c) The resulting GB propagators should be trackable in the ambient environment. The method is presented in the context of radiation into the due to a given UWB scalar ﬁeld uniform half space in the plane, with frequency band distribution . The coordinate conventions in the 3-D space with denoting the coordinates transare (Fig. 1). verse to , and the radiated ﬁeld is denoted as

Fig. 6. Frame-bound ratio B=A as a function of , calculated numerically (via the method in the Appendix ) for the Gaussian window (14) with = 0:5 (dash-dotted line) and = 0:75 (solid line). For each , B=A is minimal (snuggest frame) at = 1.

We use the caret to denote frequency domain ﬁeld constituents ; time-domain conwith a suppressed time-dependence stituents which will be considered in subsequent publications have no caret. The method is presented here in a general format, while explicit expressions for the ID Gaussian window functions, which provide an UWB snug frame representation will be given in Section IV. A. Frequency Independent Phase-Space Beam Lattice We start by deﬁning the plane wave spectrum of the initial distribution, denoted by a tilde (17) where is the wave number and is the wave speed. The frequency normalized spectral wavenumber is used instead of the conventional spectral wavenumber since is a pure geometrical constituent that deﬁnes the spectral propagation direction in a frequency independent sense.

SHLIVINSKI et al.: PHASE-SPACE BEAM SUMMATION FORMULATION FOR ULTRAWIDE-BAND RADIATION

2047

We therefore deﬁne the spatial-spectral phase space

as (18)

The UWB phase space grid is, thus, deﬁned in the

domain by

is localized around , the coefﬁcients Assuming that are the local spectrum of sampled at the grid points . is obtained now by replacing The radiated ﬁeld for in (23a) by the “beam propagators” , giving (24)

(19) where are the unit cell dimensions, and we use conveniently a vector index notation . In general, the unit cell dimensions along the and axes need not be the same as long as they satisfy the overcompleteness condition (21) in each coordinate. Here, however, we shall not utilize these options. This grid deﬁnes the origins and directions of the beam lattice hence it is required to have the same grid for all frequencies in the relevant band, while providing an overcomplete coverage of the domain. To construct this grid we ﬁrst choose a reference frequency such that (the choice of will be discussed in Section IV.B), and then choose to be critically complete at , i.e., it satisﬁes the Gabor condition at (20) The same grid is then used for all given by . The unit cell area is

denote the ﬁelds radiated into by the synthesis where . They may be described, for example, by the windows Kirchhoff integral (25) where with Alternatively, representation is the free space Green’s function, . can be expressed by a plane wave

(26) where , with is the spectrum (17) of

(21) Thus, scaling the overcompleteness parameter with the frequency such that yields a frequency independent . beam lattice for all B. Wideband Beam Expansion of the Field Next, for a proper window function , we introduce the WFT frames in [cf. (8a) and (8b)] and its dual for all

being the spectrum of the “mother” , and is the spectral wavenumber in the direction. If is wide on a wavelength scale then behave like in the collimated beams whose axes emerge from the points plane in the frequency independent directions deﬁned via . Propagating beams where is the spectral width of occur only for ( for collimated beams). For , propagate plane and decay exponentially tangentially along the with . In practice, we ignore all the evanescent beams and use , thereby expressing the in (24) only the beams with radiated ﬁeld as a discrete superposition of beams that emerge in the aperture plane and in all real from all lattice points lattice directions . IV. ID GAUSSIAN WINDOWS The ID Gaussian windows are the preferable basis functions for the UWB beam synthesis, either in the frequency or time domains, because they have the following properties: 1) Their width is scaled with the frequency in a speciﬁc fashion, termed “isodiffracting” [see (31)] that provides the snuggest frame for all frequencies in the relevant band [see (33)]. 2) They give rise to GB propagators that can be tracked locally in the ambient environment. In view of the ID scaling, the beam propagation through inhomogeneous nondispersive media or through interfaces need to be calculated only at a reference frequency, and can then extrapolated to all other frequencies [23], [25], [26]. 3) The corresponding time-domain phase-space propagators have closed form expressions, known as the ID pulsed beams [23], [25], [26].

(22a) (22b) and its dual are obtained from a The window Cartesian multiplication of the 1-D functions, i.e., . Note that in general, and do not have to be the same as long as they are valid windows in each coordinate for all the relevant frequencies. Nevertheless, here assume . that Referring to the synthesis (13), the WFT frame representation plane for all is given by of the ﬁeld in the (23a)

(23b)

2048

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

The ID windows have the general form [25], [34] where is a frequency independent, comis positive deﬁnite and plex symmetric matrix such that is the transpose of . The positive deﬁniteness of guarantees that has a Gaussian decay as the distance from the origin increases. For the sake of simplicity, we consider here the special case of a real symmetric window (27) is a frequency-independent parameter whose where optimal value for the present application will be determined following (33). This parameter is the collimation (or Rayleigh) distance of the GB that emerges from this window [see (30)], and the term ID implies that this distance is the same for all frequencies. A. The Phase-Space Propagators: ID-GBs The phase-space propagators are calculated by substituting into (26). For large (collimated beams), the integral can be evaluated asymptotically as detailed in [34], [35]. For a given , the result can be expressed in the most physically appealing format by utilizing the beam coordinates deﬁned for a given phase-space point : is in the a coordinate along the beam axis that emerges from plane in the direction as discussed after (26). The coordinates transverse to the beam axis on the plane are chosen such that the projection of coincides with the direction of , while where the over-circle denotes unit vectors along the corresponding axes. With this choice, the linear phase implied direcby the window function in (22a) is operative in the direction. These coordinates are related tion but not in the to the system coordinates via

. The collimation distances are given, accordingly, by (30) These determines the waists the wavefront radii of curvature in the cuts, gles , the beamwidths , and the diffraction anvia

(31) Thus, since we are interested in collimated beams with narrow spectral spread that can be tracked analytically via paraxial for all in the relevant models, we shall choose in the frequency band and for the largest relevant tilt angle source data (see further discussion in Section IV-B). The expression in (29) can be extended to UWB propagation in inhomogeneous media or transmission through curved interfaces. The result has the generic ID-GB form [23], [25], [26]

(32) This expression describes the beam along its (generally curved) propagation trajectory, which is the GOs ray that emerges from in a direction speciﬁed by , with and being the axial is the wavespeed along and transversal coordinates. Here and this axis while the 2 2 complex symmetric matrices are found by solving a Ricatti type equation along the axis. It can readily be shown that is positive deﬁnite for all , hence the quadratic form in the exponent exhibits a increases away from the beam axis, and Gaussian decay as [26]. that It thus follows that the ID-GB is determined by the frequency and which need to be solved independent functions once for all frequencies. B. Choosing the Frame Parameters The optimal value of is determined by considering the matching parameter in (12). Using and , where , yields the frequency independent result . Setting for a “balanced” window which provides the snuggest frame (see Section II-B and Fig. 5), we obtain (33) where in the second and third expressions we utilized (20). Thus, the ID Gaussian window (27) with related to and

(28)

Utilizing these coordinates, we ﬁnd by saddle point integration is a GB of the form of (26) that

(29) This expression is an astigmatic GB with principle axes , , 2, with waist at and collimation distances . The astigmatism is caused by the beam tilt which reduces direction by a factor the effective initial beamwidth in the

SHLIVINSKI et al.: PHASE-SPACE BEAM SUMMATION FORMULATION FOR ULTRAWIDE-BAND RADIATION

2049

Fig. 7. Relative error of the GB approximation (29) to the exact beam propagators, shown as a function of the tilt angle n for several values of the collimation parameter kb, at k = 0:1. Without loss of generality, the beams error is are tilted only in the x direction (i.e., n = (sin n ; 0)). The calculated at a distance z = 400 from the source by integrating the beam (the ﬁeld structure proﬁle normal to the axis along the beam coordinate x along x is not affected by the beam tilt; see (30).

Fig. 8. Ray trajectories for the focused aperture in (40). The ﬁgure also shows four observation points where the ﬁeld will be tested: (a) (near zone) r = (x; z ) = (0; 50), (b) (cusp point) (0, 400), (c) (far zone) (0,600), and (r ) (d) (shadow zone) (200 360). The phase space observation manifolds corresponding to these points are depicted in Fig. 11.

O

via (33), provides the snuggest WFT frame for all frequencies in the band. This leads to the following considerations for choosing the : frame parameters , and will be First, is chosen to be sufﬁciently large so that collimated even at the lowest end of the frequency spectrum to justify the GB solution. Recalling (31), should satisfy (34) and on the source It follows that the choice of depends on spectral (directional) spread. If the source has a wide spectral range that give rise to large propagation angles, it is required to be sufﬁciently large to choose the beam collimation in the source so that GB be collimated even at the largest spectrum. Another factor that should be considered in this context is that the quality of the GB approximation (29) to the exact beam propagators (given by either the Kirchhoff or the plane wave integrals in (25)–(26) with Gaussian initial condition) deteriorates for large tilt angles . This is demonstrated in Fig. 7 which also shows that for a given , the quality of the GB approximation improves by increasing the collimation . Thus, it is also reto be sufﬁciently large so that from Fig. 7 quired to choose is acthe GB approximation of the propagators at the largest curate to a desired level. where is a constant to Next, we set be chosen, as a tradeoff between the desired accuracy and the numerical efﬁcacy, see the discussion in Section V. From (21), , so that should be as small as possible to minimize the oversampling. For analytic simplicity, on the small other hand, should be sufﬁciently large making so that the dual ID window can enough for all be approximated by the 2-D extension of (11) (35)

where we also used . Choosing, for example, yields hence from Fig. 4 the error in via the approxcalculating the expansion coefﬁcients at of (35) is bounded by 4%, and it becomes smaller imate as gets smaller. The ﬁeld reconstruction for error is also bounded by this level: see Figs. 10 and 12–14. yields Choosing a larger improves the accuracy (e.g., and a 2% error bound). Note that actually Fig. 4 depicts the error for the 1-D frames considered in Section II-C; for the 2-D frames considered here, the error is larger by a factor of roughly 1.5. is localized (apClearly for the cases noted above, proximately a Gaussian) and the expansion coefﬁcients are local and stable as required. Finally, it is important to note though that the UWB beam expansion is valid in the entire band even if is taken to be smaller, yet the approximation (35) can be used only for the is small enough (say, lower frequencies where ). are determined subject to these consideraOnce and is determined via (33). Note that tions, the lattice is proportional to , hence in choosing one should also consider the desired degree of spatial localization which sets a limit on . C. Phase-Space Localization An important feature of the phase-space representation is the a priori localization of and around well deﬁned regions in domain. We assume that the initial distribution is given the by (36) where the “amplitude” and “phase” functions and are slowly varying functions of on a wavelength scale. The local direction of radiation at a given and is given by (37)

2050

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 9. Coefﬁcients a at k = 0:25 and 0.5, shown in the m ; n plane for m ; n = (0; 0). The expansion parameter b has been taken as: (a,b) b = jRj=2 = ^ 200, (c,d) b = jRj = 400, and (e,f) b = 2jRj = 800. The corresponding grids (; ) are determined by b via (33). The gray scale is in decibels. x

In the continuous phase space of (18), this relation deﬁnes the (the phase-space skeleton so called Lagrange submanifold of GOs). are samples As mentioned after (23b), the coefﬁcients of the local spectrum of with respect to the window at . If in (23b) is wide on the the phase space grid points wavelength scale, then it senses the local radiation properties of . Consequently the coefﬁcients are nonnegligible only for

points near the discrete Lagrange manifold [1], [3], [25], [34], [36] (38) depends on the The number of nonnegligible elements near width of . This limits the number of beams that are excited by the aperture.

SHLIVINSKI et al.: PHASE-SPACE BEAM SUMMATION FORMULATION FOR ULTRAWIDE-BAND RADIATION

2051

Fig. 10. Real part of the reconstructed ﬁeld in the aperture, shown only for x 0 (for x < 0, the ﬁeld is symmetrical). (a) k = 0:4 and (b) k = 0:5. Full lines: ^ the exact ﬁeld. Dashed and dotted lines: the reconstructed ﬁeld using the exact and the approximated dual function (the analysis window) ' , respectively. The dashed lines are almost indistinguishable from the full lines).

The effective range of summation in (24) is constrained further since only those beams that pass near actually contribute to the ﬁeld. For a given in a homogeneous region, these beams correspond to in the vicinity of the discrete observation manifold [1], [25], [34] (39) which simply relates the initiation points and directions of these beams. The width of the contributing zone near depends on the spatial width of at the observation point. As will be demonstrated in Section V.C, it is sufﬁcient to account only for those beams that pass within the three beam-widths vicinity of . The simultaneous constraints above implement the a priori localization. V. NUMERICAL EXAMPLES As an example, we consider the radiation from an UWB focusing aperture distribution (40) and zero otherwise, with frequency band and . The parameter is the radius of curvature of the wavefront in the aperture and it is taken here to be , (see Fig. 8). The so that the ﬁeld focuses to a cusp at . Such apertures are usually width of the aperture is characterized by the Fresnel number [39], [40] (41) is the Fresnel length, while the -number deﬁnes the spectral width. Thus, the given distribution is characterized by large Fresnel numbers ranging from 16 at to 32 at . A. Wideband Phase Space Coefﬁcients , i.e., . We choose the reference frequency to be recommended after (35), This value is smaller than where

Fig. 11. Observation manifolds corresponding to some typical observation points. They are plotted on the phase space map of the coefﬁcients a for the ^ case b = 400 and k = 0:5 in Fig. 9(d). The indexes (a,b,c,d) correspond to the observation points in Fig. 8.

O

which leads to . It was chosen in order to explore the error in calculating the expansion coefﬁcients using of (35) instead of the exact . With this the approximate yields noticeable error in the choice, using the approximate (where ) while being sufﬁciently ﬁeld for accurate for , whereas using the exact (which is calculated numerically for each ) yields accurate results everywhere (see numerical experiments in Figs. 10 and 12–14). We used three choices of the window parameter and . For , the beam propagators start , before the focal zone, while for to diverge around they are still collimated there. Note that in all cases the [see discussion beams are collimated and satisfy after (35)]. For each , the cell dimensions are calculated via (33). have been calculated using The expansion coefﬁcients (23b) and the results are shown in Fig. 9 for the three values of and for two frequencies. The coefﬁcients are shown in the plane for a slice of the phase space where . In all the ﬁgures, the dominant coefﬁcients are concenof (38), which yields, for trated near the Lagrange manifold

2052

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 12. Radiated ﬁeld at three frequencies k = 0.25, 0.4, and 0.5, shown in the z = 200 plane for x 0 (the ﬁeld for x < 0 is symmetrical). The algorithm utilizes threshold level " = 0.03 and summation over all beams passing within a three beamwidths vicinity of the observation point. Dashed and dotted lines: The ^ ﬁeld synthesized by using the exact and the approximated dual function (the analysis window) ' , respectively. Solid line: Exact reference solution.

Fig. 13.

As in Fig. 12, but for z = 360.

the distribution in (40), . In the discrete phase space, this condition becomes using (33) (42) Note that this condition is frequency independent, but the width of the strip of coefﬁcients and their magnitudes depend on the frequency. Also note the difference in the unit cell dimensions for the different values of . One also observes large contributions along the phase space corresponding to the diffraction at the end lines point of the aperture. They were termed diffraction manifolds [1], [25], [36]. The phase space characteristics can be explored analytically by evaluating the coefﬁcients approximately, using from (35). Substituting into (23b) yields the following integral for the coefﬁcients

One may readily verify that the dominant coefﬁcients in (44) are indeed aligned along the discretized Lagrange manifold of (42). B. Reconstruction of the Aperture Field The quality of the representation is explored ﬁrst by considering the ﬁeld reconstruction in the aperture. We apply thresholding to the set , i.e., we consider only the signiﬁcant coefwhere is a small error ﬁcients that satisfy parameter, thus reducing the number of elements in the summa, i.e., we neglect the terms in tion. Speciﬁcally we use Fig. 9 below . We compare reconstruction using the exact coefﬁcients (dashed lines), and using the approximate ones that are obtained of (35) [dotted by processing the data with the approximate lines]. In the former case, the error is mainly due to the thresh[for the reconstructed ﬁeld is olding at a level almost indistinguishable from the exact ﬁeld (solid line)]. In the , which is due to the latter, there is a noticeable error for , whereas for where , error in (35) for the reconstruction error is smaller than 6% as predicted in Fig. 4 and the discussion after (35). One also observes a weak Gibbs . We thereeffects at the aperture truncation points and 0.5 since the quality of the fore consider here only approximate reconstruction is excellent for . or In view of these results it is recommended to use in order to obtain accurate results while using the even approximate dual function of (35).

(43) For grid points that are far away from the aperture boundaries, the end points effect in the integral can be neglected leading to : the closed form result for

(44)

SHLIVINSKI et al.: PHASE-SPACE BEAM SUMMATION FORMULATION FOR ULTRAWIDE-BAND RADIATION

2053

Fig. 14.

As Fig. 12, but for z = 600.

C. Radiated Field The radiated ﬁeld has been calculated via the beam summation formula (24) using the ID-GB propagators of (29), and the results are shown in Figs. 12–14. In view of our UWB phase space construction, the beam trajectories and the beam parameters are frequency independent. The summation includes only [see discussion after (26)]. the propagating beams with As noted in Section V-B, the number of beams excited is reduces at an error level (i.e., by applying thresholding to ). Furthermore, only those beams that pass “near” a given observation point are summed in (24): As follows from the numerical results below, it is sufﬁcient to include only those passing within the three beam-widths vicinity of , where the beamwidth is given in (31). Viewed from a phase-space perspective, the beams that pass near are deﬁned by the of (39), which is illustrated in observation manifold Fig. 11 for the four observation points (a)–(d) that are indicated in Fig. 8, and the algorithm selects the beams that are located within . Dominant contributions a three beam-widths vicinity of at a given are therefore obtained only from those phase-space of with the points in the vicinity of the intersection of Lagrange and diffraction manifolds of the beam amplitudes (shown by the gray scale in Fig. 11). Considering as an example the near zone point (a), the beam contributions obtained from the intersection of in Fig. 11 with the Lagrange manifold describe the GOs ﬁeld near the aperture center, while those obtained from the intersection with the diffraction manifolds describe edge diffraction contributions at . At the focal point (b), on the other hand, in Fig. 11 is essentially parallel to the Lagrange manifold and the ﬁeld is described by signiﬁcant beam contributions from the entire aperture. Finally, at point (d) in the shadow zone, the beams along the corresponding in Fig. 11 are weakly excited hence there are no sgniﬁcant GOs contributions but there are contributions from the intersection of with the diffraction manifolds. The total numbers of beams used in the formulation are summarized in Table I. For further discussion, see Section V-D. Figs. 12–14 depict cross sectional cuts of the UWB ﬁeld calculated at several distances in the near, intermediate and far and zones, and at three different frequencies . The dashed and dotted lines compare, respectively, the ﬁelds calculated by using the exact and the approximated

TABLE I NUMBER OF BEAMS USED IN OUR ALGORITHM. 2ND COLUMN: NUMBER OF EXCITED BEAMS AFTER THRESHOLDING AT LEVEL " = 0:03. 3RD-6TH COLUMNS: NUMBER OF BEAMS USED TO SYNTHESIZE THE FIELD AT THE OBSERVATION POINTS (A)–(D) IN FIG. 8

of (35), respectively. As in Fig. 10, the former provides accurate beam amplitudes and excellent agreement with the exact solution (full lines) for the entire frequency band: the minor noticeable discrepancies are due to the thresholding of at a level , and are essentially eliminated by using . of leads to accurate results for Using the approximate but a noticeable error for . Note that the beam formulation provides a good representation for the ﬁeld near the shadow ), for boundary of edge diffraction (see Fig. 12 around the ﬁeld in the focal zone in Fig. 13, and for the transition to the far zone in Fig. 14. D. The Phase Space Degrees of Freedom The manifestation of localization in discrete phase-space representations like the discrete Fourier transform or the beam summation, can be quantiﬁed in view of the Landau-Pollak (LP) bound that speciﬁes the number of discrete degrees of freedom required for a given ﬁeld to be approximated to a prescribed accuracy [3]. This issue is explored here in the context of our UWB beam summation representation. Clearly, the oversampling increases the number of terms, but on the other hand, the snug localization extracts the local features of the ﬁeld and renders the phase space coefﬁcients highly localized near the Lagrange manifold. Consequently the number of degrees of freedom may be signiﬁcantly reduced relative to the LP bound. Further savings are achieved for UWB ﬁelds since the beams need to be traced only once at a reference frequency. that We start with the LP dimension deﬁnes the degrees of freedom in a discrete representation of a 2-D (aperture) domain, with and being the spatial and spectral cross sections of the ﬁeld distribution, respectively [3]. yields Setting (45)

2054

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

This expression describes a complete phase space coverage and visible spectrum it is obtained, for example, by assigning spatial points in the aperture where points to each of the . A smaller number is obtained if one includes only the GOs radiation from the aperture and excludes edge diffraction. Referring to the aperture source distribution in (40) we use and obtain (46) is the Fresnel number of (41). where In the beam summation formulation, the actual number of elements is further reduced since it is sensitive to the local structure of the phase space. We consider only the nonnegligible where is a small coefﬁcients that satisfy error parameter. Referring to Fig. 9, the number of phase is then a sum of elements space coefﬁcients of (38) and along the discrete Lagrange (or GOs) manifold elements near the diffraction manifold. The latter will be neglected in the estimates in view of the bound , which applies asymptotically for large apertures ( is proportional to the linear dimension while is proportional to the area). From (44), the relevant coefﬁcients are enclosed within a 4-D tube near (47) For a given , (47) represents the interior of a circle in the of radius plane, which is centered at on . The in that circle is equal to its area number of points . Hence accounting for all the points in the aperture yields an estimate on (48) with , where the second expression follows from (33). It follows that the minimal number of ele. ments is obtained with Comparing (48) with (46) one obtains the asymptotic relation , i.e., the phase space beam representation consists, asymptotically, of a smaller number of elements. This property follows from the fact that the beam formulation extracts the local radiation property of the source and thus includes only the elements along the Lagrange manifold, whereas LP is a global bound ( bounds the spectral cross section of the entire aperture ﬁeld). appearing in (48) which One should also note the term represents the frame oversampling. Finally, the number of elements needed to describe the ﬁeld at a given may be further reduced if out of all the phase space points in (48) one keeps only those corresponding to beams that pass near , as discussed in Section V-C, and summarized in Table I. Applying (48) to the present conﬁguration, using threshold we obtain and for level and , respectively. These numbers are smaller than

those in the second column of Table I for the beams that are actually excited by the aperture at the same threshold level. The difference is essentially due to the edge diffraction beams that are included in the calculations of Table I but not in (48). Recalling the discussion before (47), it is expected that for larger apertures in (48) the relative difference between the estimate for and the results of Table I will diminish.

VI. SUMMARY AND CONCLUDING REMARKS A novel discrete phase space beam summation representation for UWB radiation from extended source distributions was introduced. The representation is based on a WFT frame analysis of the aperture source distribution. The formulation comprises the following key features: 1) It utilizes a frequency independent beam lattice (unlike the conventional Gabor scheme), emerging from a discrete set of points and orientations in the aperture. This important feature is achieved by introducing relation (21) between the overcompleteness and the frequency. 2) The ID-GB are shown to provide snuggest frames for all frequencies, provided that the beam parameter and the phase-space grid are related via (33). 3) The ID-GB propagators are also fully described by frequency independent parameters [see (28) and (32)]. Consequently, the calculation of the ID-GB propagation in the ambient environment need to be done only once for all frequencies. 4) The expansion coefﬁcients are samples of the WFT of the source distribution with respect to the dual (or “analysis”) window. An important parameter is the reference frequency which is larger than , the highest frequency in the source spectrum. If is chosen to be sufﬁciently large ( with ), the overcompleteness is greater than for all and the dual window can be approximated by the ID-GB window as in (35). This greatly simpliﬁes the calculations since the dual function does not have to be calculated numerically for each frequency in the band (the numerical procedure for calculating the dual function numerically is described in the Appendix). Furthermore, in this case the dual function can be transformed in closed form to the time domain, thus providing a starting point for the new time domain theory that has been brieﬂy reported in [37], [38]. 5) Following item 4, the overcompleteness poses a tradeoff between analytical simplicity and numerical efﬁcacy. Referring to [19], [20] is sufﬁcient for expansion coefﬁcient localization. If, like in our case, an approximate dual function is desired, we have found that is preferable since in this case the error due to the approximate dual function is bounded by 6% at , and it is lower for where (Fig. 4 and numerical examples in Section V). Choosing reduces the error bound at to 3%. Large increases the overcompleteness and the number of elements, but simpliﬁes the calculations of the wideband expansion coefﬁcients (item 4) as well as the ﬁeld calculations (items 1 and 3). Furthermore, the snug localization of the formulation renders the phase space representation highly localized near the Lagrange manifold (the GOs skeleton),

SHLIVINSKI et al.: PHASE-SPACE BEAM SUMMATION FORMULATION FOR ULTRAWIDE-BAND RADIATION

2055

hence, depending on the source properties, the number of elements in the expansion can be signiﬁcantly reduced relative to the LP bound (see Section V-D). In view of the frequency independence of the various characteristics above, the formulation can be transformed in closed form into the time domain, where the propagators are the ID pulsed-beams and the expansion amplitudes are found via a new discrete local slant-stack (or Radon) transform. Initial results of these new formulations have been reported in [29], [37], [38], but full details will be published separately. Referring to item 1, we note that the oversampling increases at lower frequencies so that the numerical efﬁciency there decreases (see item 5). Thus, although the bandwidth of the formulation above can be arbitrarily large, we have introduced in [37], [38] a modiﬁed formulation for excitations with bandwidth . For these cases, we larger than one octave, divide the excitation band into a hierarchy of one-octave subbands, and apply the UWB formulation above in each band while choosing the parameters in each band such that the spatial and spectral discretizations are obtained by a multilevel binary decimation of those at the highest band. The result is a self-consistent representation wherein the beam sets at the lower bands are decimated subsets of those at the highest band, so that only the set of beam-propagators at the highest frequency band needs to be traced, while for the lower bands one may use properly decimated subsets. The resulting multiband algorithms in the UWB frequency domain and in the time domain have been reported in [37], [38], but full papers are in preparation. We are also engaged in extending the expansion algorithm to circular cylinders, mainly in connection with applications to indoor propagation. This subject will be reported elsewhere. APPENDIX DUAL FRAME CALCULATION VIA AN EXPANSION IN A FINITE HILBERT SPACE We present the numerical algorithm used for calculating the dual window of the WFT frames shown in Figs. 3 and 5 and of the frame bounds in Fig. 6. Two other approaches, namely the Neumann series and the Zak transform, have been discussed after (8b). , which The dual window is deﬁned in (8b) by with in view of (2) yields explicitly . To calculate , we expand it as , where , is a valid expansion set in a ﬁnite that covers the support of (i.e., can Hilbert space be a basis, a set of sampling delta functions, or a frame set), and are unknown coefﬁcients. We therefore obtain (49) Recalling that if for , this equation is satisﬁed (50) where , is a column vector and is a column vector of nulls except for the term corresponding that equals 1. has columns but an inﬁnite to

number of rows, however, it may be truncated since from [41] for . This property holds for any we have frame, yet it may readily be veriﬁed for the present WFT frame is by recalling the essentially ﬁnite support of . Since a frame set, (50) is overdetermined and its solution may be obtained by the singular value decomposition (SVD) [42], yielding , being the pseudo inverse obtained after truncating the low singular values (see [43]). A convenient choice for the expansion set is the Galerkin set in which case is Hermitian. In this case, the largest and smallest singular values of are the upper and lower frame bounds and , respectively [41]. REFERENCES

[1] B. Z. Steinberg, E. Heyman, and L. B. Felsen, “Phase space beam summation for time-harmonic radiation from large apertures,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 8, pp. 41–59, 1991. [2] , “Phase space beam summation for time dependent radiation from large apertures: Continuous parameterization,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 8, pp. 943–958, 1991. [3] J. M. Arnold, “Phase-space localization and discrete representations of wave ﬁelds,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 12, pp. 111–123, 1995. ˇ [4] V. Cervený, “Gaussian beam synthetic seismograms,” J. Geophys., vol. 58, pp. 44–72, 1985. [5] M. J. Bastiaans, “The expansion of an optical signal into a discrete set of Gaussian beams,” Optik, vol. 57, pp. 95–102, 1980. [6] P. D. Einziger, S. Raz, and M. Shapira, “Gabor representation and aperture theory,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 3, pp. 508–522, 1986. [7] J. J. Maciel and L. B. Felsen, “Gaussian beam analysis of propagation from an extended plane aperture distribution through dielectric layers: I-Plane layer,” IEEE Trans. Antennas Propagat., vol. 38, pp. 1607–1617, 1990. , “Gaussian beam analysis of propagation from an extended plane [8] aperture distribution through dielectric layers: II-Circular cylindrical layer,” IEEE Trans. Antennas Propagat., vol. 38, pp. 1618–1624, 1990. [9] R. J. Burkholder and P. H. Pathak, “Analysis of EM penetration into and scattering by electrically large open waveguide cavities using Gaussian beam shooting,” Proc. IEEE, vol. 79, pp. 1401–1412, 1991. [10] H. T. Chou, P. H. Pathak, and R. J. Burkholder, “Novel Gaussian beam method for the rapid analysis of large reﬂector antennas,” IEEE Trans. Antennas Propagat., vol. 49, pp. 880–893, 2001. [11] B. Rao and L. Carin, “Hybrid (parabolic equation) (Gaussian beam) algorithm for wave propagation through large inhomogeneous regions,” IEEE Trans. Antennas Propagat., vol. 46, pp. 700–709, 1998. , “Beam-tracing-based inverse scattering for general aperture an[12] tennas,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 16, pp. 2219–2231, 1999. [13] M. J. Bastiaans, “Gabor’s expansion of a signal into Gaussian elemntary signals,” Proc. IEEE, vol. 68, pp. 538–539, 1980. , “Signal descriptions by means of a local frequency spectrum,” [14] Proc. SPIE, Transformation Opt. Signal Processing, vol. 373, pp. 49–62, 1981. [15] A. J. E. M. Janssen, “Bargmann transform, Zak transform and coherent states,” J. Math. Phys., vol. 23, no. 5, pp. 720–731, 1982. [16] I. Daubechies, “The wavelet transform, time-frequency localization and signal analysis,” IEEE Trans. Inform. Theory, vol. 36, pp. 961–1005, 1990. , Ten Lectures on Wavelets. Philadelphia, PA: SIAM Publ., 1992, [17] CBMS-NSF Series in Applied Mathematics. [18] B. Z. Steinberg and E. Heyman, “Phase space beam summation for time dependent radiation from large apertures: Discretized parametrization,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 8, pp. 959–966, June 1991. [19] D. Lugara and C. Letrou, “Alternative to Gabor’s representation of plane aperture radiation,” Electron. Lett., vol. 34, pp. 2286–2287, 1998. , “Printed antennas analysis by a Gabor frame-based method of mo[20] ments,” IEEE Trans. Antennas Propagat., vol. 50, pp. 1588–1597, 2002. [21] A. Shlivinski, E. Heyman, A. Boag, D. Lugara, and C. Letrou, “Gaborframe phase space beam summation formulation for wideband radiation from extended apertures,” in Proc. URSI Trianum Int. Symp. Electromagnetic Theory, Victoria, Canada, May 2001, pp. 56–58. [22] D. Lugara, C. Letrou, A. Shlivinski, E. Heyman, and A. Boag, “The frame based Gaussian beam summation method: Theory and application,” Radio Sci., vol. 38, no. 2, pp. VIC27/1–15, 2003.

2056

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[23] E. Heyman, “Pulsed beam propagation in an inhomogeneous medium,” IEEE Trans. Antennas Propagat., vol. 42, pp. 311–319, 1994. [24] E. Heyman and T. Melamed, “Certain consideration in aperture synthesis for ultra-wideband/short-pulsed ﬁelds,” IEEE Trans. Antennas Propagat., vol. 42, pp. 518–525, 1994. [25] E. Heyman and L. B. Felsen, “Gaussian beam and pulsed beam dynamics: Complex source and spectrum formulations within and beyond paraxial asymptotics,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 18, pp. 1588–1611, 2001. [26] E. Heyman, “Pulsed beam solutions for propagation and scattering problems,” in Scattering: Scattering and Inverse Scattering in Pure and Applied Science, R. Pike and P. Sabatier, Eds. New York: Academic, 2002, vol. 1, ch. 1.5.4, pp. 295–315. [27] S. Feng and H. G. Winful, “Spatiotemporal transformations of isodiffracting ultrashort pulses by nondispersive quadratic phase media,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 16, pp. 2500–2509, 1999. [28] M. A. Porras, “Nonsinusoidal few-cycle pulsed light beams in free space,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 16, pp. 1468–1474, 1999. [29] A. Shlivinski, E. Heyman, A. Boag, A. Fluerasu, and C. Letrou, “A discretized-phase-space pulsed beam representation for time dependent radiation,” in Proc. URSI Trianum Int. Symp. Electromagnetic Theory, Victoria, Canada, May 2001, pp. 71–73. [30] R. J. Dufﬁn and A. J. Schaeffer, “A class of nonharmonic Fourier series,” Trans. AMS, vol. 72, pp. 341–366, 1952. [31] K. Gröchenig, “Acceleration of the frame algorithm,” IEEE Trans. Signal Processing, vol. 41, pp. 3331–3340, 1993. [32] Gabor Analysis and Algorithms: Theory and Applications, H. G. Feichtinger and T. Strohmer, Eds., Birkhäuser, Boston, MA, 1998. [33] M. J. Bastiaans, “Gabor’s signal expansion and the Zak transform for continuous-time and discrete-time signals: critical sampling and rational oversampling,” Eindhoven Univ. Technol., Faculty Elect. Eng., Eindhoven, Netherlands, EUT Report 95-E-295, ISBN 90-6144-295-8, 1995. [34] E. Heyman and T. Melamed, “Space-time representation of ultra-wideband signals,” in Advances in Imaging and Electron Physics, P. W. Hawkes, Ed. New York: Academic, 1998, vol. 103, pp. 1–66. [35] T. Melamed, “Phase space beam summation: A local spectrum analysis of time dependent radiation,” J. Electromagn. Waves Appl., vol. 11, pp. 739–773, 1997. [36] J. M. Arnold, “Rays, beams and diffraction in a discrete phase space: Wilson bases,” Optics Express, vol. 10, no. 16, pp. 716–722, Aug. 2002. [37] A. Shlivinski, E. Heyman, A. Boag, and C. Letrou, “Frame-based beamsummation algorithms for ultra wideband radiation from extended apertures. Part I: Formulations in the multi-frequency domain,” and “Part II: Time domain formulation,” in Ultra-Wideband, Short-Pulse Electromagnetics, E. Mokole, K. Gerlach, and M. Kragalott, Eds. New York: Plenum Press, 2003, vol. 6, pp. 101–122. , “A frame based phase-space beam and pulsed beam summation [38] formulations for ultra wideband/short pulse radiation,” in Proc. XXVII General Assembly Int. Union of Radio Science (URSI), Maastricht, The Netherland, August 2002, manuscript #699, pp. 1–4. [39] A. E. Siegman, Lasers. Mill Valley, CA: Univ. Sci. Books, 1986. [40] J. J. Stamness, Waves in Focal Regions: IOP Publishing, 1986. [41] A. Teolis and J. J. Benedeto, “Local frames and noise reduction,” Signal Processing, vol. 45, pp. 369–387, 1995. [42] G. H. Golub and C. V. Loan, Matrix Computations, 2nd ed. Baltimore, MD: The Johns Hopkins Univ. Press, 1989. [43] T. Strohmer, “Approxiamtion of dual Gabor frames, window decay, and wireless communications,” App. Comp. Harmonic Anal., vol. 11, pp. 243–262, 2001.

Ehud Heyman (S’80–M’82–SM’88–F’01) was born in Tel Aviv, Israel, in 1952. He received the B.Sc. degree in electrical engineering from Tel Aviv University, Israel (summa cum laude) as Valedictorian, the M.Sc. degree in electrical engineering (with distinction) from The Technion—Israel Institute of Technology, Haifa, and the Ph.D. degree in electrophysics from the Polytechnic Institute of New York (now Polytechnic University), Brooklyn, in 1977, 1979, and 1982, respectively. In 1983, he joined the Department of Physical Electronics of the Faculty of Engineering, Tel Aviv University where he is now a Professor of electromagnetic theory and Heads the School of Electrical Engineering. From 1991 to 1992, he was on sabbatical at Northeastern University, Boston, MA, the Massachusetts Institute of Technology, Cambridge, and the A. J. Devaney Association, Boston. He spent several summers as a Visiting Professor at various universities. He has published over 80 articles and has been an Invited Speaker at many international conferences. His research interests involve analytic methods in wave theory, including high-frequency and time-domain techniques for propagation and scattering, short-pulse antennas and pulsed beams, inverse scattering and target identiﬁcation, imaging and synthetic aperture radar propagation in random medium. Prof. Heyman is a Member of Sigma Xi and the Chairman of the Israeli National Committee for Radio Sciences (URSI). He is an Associate Editor of the IEEE Press Series on Electromagnetic Waves and was an Associate Editor of the IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION. While at the Polytechnic Institute he was a Research Fellow and later a Postdoctoral Fellow, as well as a Rothschild, a Fullbright, and a Hebrew Technical Institute Fellow.

Amir Boag (S’89–M’91–SM’96) received the B.Sc. degree in electrical engineering and the B.A. degree in physics (both summa cum laude) in 1983, the M.Sc. degree in electrical engineering in 1985, and the Ph.D. degree in electrical engineering in 1991 from The Technion—Israel Institute of Technology, Haifa. From 1991 to 1992, he was on the Faculty of the Department of Electrical Engineering, The Technion—Israel Institute of Technology. From 1992 to 1994, he was a Visiting Assistant Professor with the Electromagnetic Communication Laboratory, Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign. In 1994, he joined Israel Aircraft Industries as a Research Engineer where he became a Manager of the Electromagnetics Department in 1997. Since 1999, he has been with the Department of Physical Electronics, the School of Electrical Engineering, Tel Aviv University, Israel. He has published more than 40 journal articles and presented more than 70 conference papers on electromagnetics and acoustics. His research interests are in electromagnetic theory, wave scattering, and design of antennas and optical devices.

Amir Shlivinski (S’98–M’04) was born in Tel-Aviv, Israel, in February 1969. He received the B.Sc. (cum laude), M.Sc. (summa cum laude), and Ph.D. (with distinction) degrees in electrical engineering, all from Tel-Aviv University, Israel, in 1991, 1997, and 2003, respectively. From 1991 to 1999, he worked as a research and development Electromagnetic Engineer, and between 1999 and 2003, he was a full-time Ph.D. student and a Teaching Assistant at Tel-Aviv University. Currently, he is a Postdoctorate Fellow at the Department of Electrical Engineering, University of Kassel, Kassel, Germany. His main ﬁelds of interest are electromagnetics, wave theory and antenna theory, with emphasis on analytic methods and time-domain phenomena.

Christine Letrou (M’96) received the Diplôme d’Ingénieur of Institut National des Télécommunications (INT), Evry, France, in 1982 and the Docteur-Ingénieur and Ph.D. degrees from Université Paris XI, Orsay, France, in 1985 and 1988, respectively. Since then, she has been an Assistant Professor at INT in charge of microwaves, electromagnetics and antennas, teaching, and research activities. Currently, she is also a Member of the CNRS Laboratory SAMOVAR (UMR 5157), Paris, France. Her main research interests are in phase-space methods development, antennas and quasioptical devices and systems design, and propagation modeling for high bit rate communication systems.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2057

**Theoretical Considerations in the Optimization of Surface Waves on a Planar Structure
**

Samir F. Mahmoud, Senior Member, IEEE, Yahia M. M. Antar, Fellow, IEEE, Hany F. Hammad, and Al P. Freundorfer, Senior Member, IEEE

Abstract—The problem of optimum excitation of surface waves on a grounded dielectric slab by means of slots in the ground plane is considered. By adopting a two-dimensional (2-D) model, analysis lead to closed forms for the power launched as surface waves and power leaked as radiation. Input admittance of a single slot source and mutual admittance between two slots are derived and utilized to design a three element Yagi array of slots to achieve a prescribed ratio of forward to backward surface wave power. As a development of the 2-D model, we allow ﬁnite extent of slot excitation by assuming a Gaussian E-ﬁeld distribution across the slot. The effect of the Gaussian width on the excited surface wave power is studied. The analysis is relevant to the study of surface waves on printed circuits. Speciﬁcally, it applies to the implementation of power combiners based on quasioptical slab beam that have been recently introduced in the literature for use in the millimeter wave band. Index Terms—Millimeter wave power combiners, planar structures, quasioptical power combiners, surface waves.

(a)

I. INTRODUCTION

(b) Fig. 1(a). 2-D model of a grounded dielectric slab of relative permittivity " with a slot source in the ground plane. (b). Grounded dielectric slab fed by a Yagi array of three slots. The middle slot of width “s” is the driven one and the other two are director and reﬂector slots having widths s and s , respectively. The spacing between the driven slot and the director/reﬂector slots are “c ”=c , respectively.

E

XCITATION of surface waves on planar integrated microwave circuits is often considered as an adversary effect causing power loss and undesired coupling. However there exist situations when the main objective is to efﬁciently excite a surface wave with least possible leakage, or radiated power. One recent example of these situations is the implementation of quasioptical slab beam power combiners in the millimeter band, in which surface waves are the means of power transport [1]–[4]. These combiners depend on the efﬁcient excitation of the dominant surface wave mode inside a dielectric slab. Recent investigations by the authors [3], [4] have suggested the use of coplanar waveguide (CPW) driven slots as the most suitable surface wave launchers for monolithic fabrication in the millimeter regime. In this paper an attempt is made to establish the theoretical foundation for operation of the slab beam power combiner. Accordingly we seek to maximize the surface wave excited by a slot dipole on the ground plane of a grounded dielectric slab. Starting with a two-dimensional (2-D) model of the grounded slab and the slot, a rigorous theory is presented in Section II that leads to closed forms for the excited surface wave and leakage powers. The input

admittance of a single slot and mutual admittance between two slots are derived in Section III. Numerical examples are given in Section IV including design examples of Yagi slot arrays that achieve high surface wave front to back ratio. In Section V, we alleviate the assumption of uniformly excited slot by allowing a Gaussian distribution of the E-ﬁeld inside the slot. This breaks the two dimensionality of the problem, but still allows the derivation of closed forms for the surface wave and radiated powers. Comparison is made between the present theory and experimental implementations as given by the authors in earlier work [4]. II. THEORY OF GROUNDED SLAB EXCITATION BY A SLOT SOURCE (2-D MODEL) We start by considering a 2-D model of the problem where a grounded dielectric slab of uniform thickness is assumed to extend inﬁnitely parallel to the - plane. A -directed slot of width “ ” (along ) in the ground plane is used to excite the slab as shown in Fig. 1(a). The slot itself is excited by and therefore, acts as a uniform -oriented electric ﬁeld an inﬁnite magnetic line source of magnetic current given by (volts). To limit the problem to the 2-D model, we assume that the slot is inﬁnite in the direction and

Manuscript received January 5, 2003; revised July 9, 2003. S. F. Mahmoud is with the Electrical Engineering Department, Kuwait University, Kuwait (e-mail: samir@eng.kuniv.edu.kw). Y. M. M. Antar is with the Department of Electrical and Computer Engineering, Royal Military College of Canada, Station Forces Kingston, ON K7K 7B4, Canada (e-mail: antar-y@rmc.ca). H. F. Hammad and A. P. Freundorfer are with the Electrical and Computer Engineering Department, Queen’s University, Kingston, ON K7L 3N6, Canada. Digital Object Identiﬁer 10.1109/TAP.2004.832498

0018-926X/04$20.00 © 2004 IEEE

2058

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

is uniformly excited. In Section V, we consider a ﬁnite slot excitation. The excited ﬁelds can be expressed by a discrete set as well as a of (say ) surface wave modes traveling along continuous spectrum over the transverse wavenumber (along ) that account for radiated ﬁelds. This is known as the transverse spectral representation for the ﬁelds [5]–[8]. This representation is adopted here since it facilitates the determination of the modal amplitudes as will be seen. Since the slot acts as a -oriented uniform magnetic line source, ﬁelds are obviously of and electric ﬁeld components TM type with magnetic ﬁeld and . Assuming a time harmonic excitation of the form , and that the source lies in planes, general directions are expressions for the total ﬁelds in the

leading to the modal equation for a surface wave across for a pseudomode as and determines (5) represents Obviously, the range of the spectrum ﬁelds with active radiation power while the range corresponds to the evanescent part of the ﬁeld. While a pseudomode does not satisfy the radiation condition on its own, the sum of pseudomodes making up the radiated and evanescent ﬁelds does satisfy the radiation condition, as it should [6]. The modal . electric ﬁeld vector is Orthogonality relationships can be established among the surface wave modes and pseudomodes and can be expressed by [8]

(1) (6)

(2)

unless whence it is equal to unity, where the is the usual Dirac Delta function. In addition there is while orthogonality between a surface wave mode and a pseudomode. is easily obtained as The surface wave factor (7) where the modal ﬁelds have been normalized such that . The second integration in (6) is a bit more difﬁcult to evaluate. Following [8], it is useful to change that integration to a contour integral as follows:

superscript and signs apply for and where the , respectively, and , and, are surface wave mode and pseudomode amplitudes. The integration is taken over the wavenumber . Here , are the transverse magnetic -ﬁeld component and the vector electric ﬁeld of the th surface wave mode; namely for the surface wave mode

(3) . and , , Here, the attenuation rate of the th mode in air, and is the free space wavenumber. , are the ﬁelds of a pseuOn the other hand, domode [6]–[8] with a transverse real wavenumber . A pseudomode is the result of an incident plane wave on the slab and a reﬂected one; namely

(8) Evaluating the right-hand side (RHS) and using the identity , we get in (6) as

(9) (4) where: , and . Obviously, (3) and (4) ensure the continuity of the -magnetic ﬁeld . The associated electric ﬁeld is obacross the interface tainable from (3) and (4) through Maxwell’s equations. The requires that be continuous continuity of It is worth noting that, in the absence of any loss in the strucis real and stands for the surface wave power ﬂowing ture, is real and represents spectral power along . Meanwhile , while in the range , density over the range is pure imaginary and so is , which then represents reactive power density.

MAHMOUD et al.: THEORETICAL CONSIDERATIONS IN THE OPTIMIZATION OF SURFACE WAVES

2059

Now consider a -directed inﬁnite slot of inﬁnitesimally exists. narrow width ; Fig. 1, over which a uniform This acts as a magnetic line source of magnetic current . The ﬁelds generated by this source and take the general form in (1) and (2) and the coefﬁcients are to be determined from the electric ﬁeld discontinuity and the magnetic ﬁeld continuity across the source; namely

accounting for the ﬁnite width of the slot and the variation of over the slot. Therefore

(10) Inserting and from (1) and (2) and using the orthogonality relations in (6), we get the modal amplitudes as

(15) which is real over the range of where integration. Performing the integrals over and , and substifrom (9), we arrive at tuting for

(11) Now, we are able to obtain both the guided surface wave power ) and radiation power in simple summation and in(in both tegral form, respectively. Namely (12) and

(16) Note, that the square bracketed term accounts for the ﬁnite width , rendering of the slot. As tends to , this term behaves as the whole integrand to behave as , which ensures a convergent integral. The results obtained so far also allow us to determine the between two parallel slots of given mutual admittance widths and given spacing “ .” For “ ” sufﬁciently larger than is given by slot widths, the mutual admittance

(17) (13) where is the magnetic ﬁeld at slot 2 due to a voltage applied to slot 1. Note that in the range , and in the range . IV. NUMERICAL RESULTS The percentage power launched in surface waves relative to the total power delivered by the source is computed from (12) and (13) versus normalized frequency for different values of the relative permittivity . The frequency is limited to allow for the propagation of only a single surface wave mode. Deﬁning , a single mode a normalized frequency as operation occurs when . Results in Fig. 2 show a monotonic increase of the percentage surface wave power with up where there is a broad maximum whose value to about increases with the substrate . For example a peak value of 88% is attained for , and 68% for . and susceptance per one free The slot conductance space wavelength along are plotted versus in Figs. 3 and and the 4. The conductance displays a peak around peak value depends on the substrate relative dielectric constant, while has a maximum slope near the peak of . While the slot conductance is independent of the slot width for a narrow enough slot, the slot susceptance changes considerably with the slot width as seen in Fig. 4. The break up of the slot in surface and , are plotted in wave and radiation components; Fig. 5 for two values of . The surface wave conductance at. For , both and tains its maximum around

and are given by (7) and (9), respectively. where It is constructive at this point to compare between the transverse wavenumber spectral representation with the more conventional longitudinal wavenumber spectral representation. The latter has been used for several decades by Wait [9], Fuller and Wait [10] and many others, and recently pursued by Bhattacharyya [11]. While the transverse spectral representation explicitly displays the surface wave and the radiation spectra separately, the longitudinal spectral representation does not. However, the latter representation can be converted to the former by changing the path of integration on the real axis of the longitudinal wavenumber (say ) to the complex plane where the poles contribution gives the surface wave modes and the branch cut gives the radiation part [9]–[11]. III. SLOT ADMITTANCE The slot admittance (Siemens per unit length along ) is given by (14)

is given by (1). It turns out that the slot conwhere is given, as expected, by the sum of surface wave ductance and radiation powers divided by . As for the slot susceptance , per unit length, care should be taken in evaluating (14) by

2060

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Percentage surface wave power excited by a slot source versus k d " 1. The relative permitivity " is a normalized frequency F varying parameter.

p 0

Fig. 3. Slot conductance in milli-mho per a free space wavelength versus normalized frequency F . The relative permitivity " is a varying parameter.

are reduced, but is reduced at a slightly higher rate. This explains the shift of the maximum percentage surface wave power to higher values of than 1.6 (see Fig. 2). These observations agree with experimental work conducted by the authors in [4]. A. Design of a Yagi Slot Array Having obtained the self and mutual admittance of slots, one can design an array of such slots to achieve maximum front direction) ratio of excited surface wave. ( ) to backward ( A three-element Yagi slot array is shown in Fig. 1(b). It is composed of one fed slot, of width “ ” and parasitic director and and . The separation between the reﬂector slots of widths fed slot and each of the director and reﬂector slots is denoted by and , respectively. The results of two design examples are shown in Fig. 6 where the relative forward and backward surface wave powers are plotted against frequency. Each array is numerically optimized with respect to the slot widths and the spacing between the elements for a maximum front to back ratio of surface wave power at the center frequency. As seen in Fig. 6, It is possible to achieve a front to back ratio better than 20 dB over a bandwidth of 2.7% (Design 1) or 4.4% (Design 2) around the . normalized frequency The theoretical results displayed in Fig. 6 can be compared to simulated and measured data reported by the authors in [4, Fig. 12], where a three element Yagi array launcher is implewith 2.54 mm thickmented on a Duroid substrate at a frequency of 11.8 GHz. ness. The resulting A comparison between the present theory, represented by design 1 and 2 of Fig. 6 and the simulated results in [4] is given in Table I in terms of the bandwidth of the 20 dB and 15 dB front to back ratio. Although the theory is based on a 2-D model, there is a good agreement with the simulated results in [4]. V. FIELDS OF A FINITE LENGTH SLOT So far we have carried out a 2-D analysis. Now we wish to consider a more realistic situation where the slot source, which is equivalent to a magnetic line source, is effectively of ﬁnite length along . We shall continue to assume that the slab is having inﬁnite width in the direction. The magnetic current

Fig. 4. Slot conductance and susceptance versus F for " = 9:8. The slot susceptance varies with the relative slot width parameter s=d.

Fig. 5. Surface wave and radiation conductance of a slot versus F . " takes the values 3 and 9.8.

on the ground surface form; i.e.

is considered to have a Gaussian

(18)

MAHMOUD et al.: THEORETICAL CONSIDERATIONS IN THE OPTIMIZATION OF SURFACE WAVES

2061

are derived The other ﬁeld components; , , , and from (20) through well-known relations [8]. The normalizing and are still deﬁned as in (6), but now they factors are functions of . After some manipulations, we get

(22) where and are those given by (7) and (9). The amplitudes of the surface wave modes and pseudomodes are obtained by using (18) and (19), and applying the boundary conditions (10) (23) and

TABLE I FBR BANDWIDTH OF THREE ELEMENT YAGI; COMPARISON BETWEEN THEORY AND SIMULATION

Fig. 6. Forward and backward surface wave power versus frequency f =f for two designs. The center frequency corresponds to F = 1:9 and " = 9:8. With reference to Fig. 1(b), the parameters of the two designs are: Design 1: s = 0:1d, s = 0:1d, and s = 0:15d, and c = 2:066d, and c = 0:817d. Design 2: s = 0:1d, s = 0:125d, and s = 0:2d, and c = 2:066d and c = 0:817d.

(24) Combining these equations with (20) and (21), we obtain a closed form expression for . Next considering the TE waves, similar expressions for the TE part of the spectrum can be replacing . Namely [(23), (24)] apply to derived, with TE modes after multiplying the RHS by . We are now in a position to obtain the ﬁelds and powers launched as surface waves. Assuming the propagation of a single TM mode, inside the dielectric we can derive the component of layer as

which is a source of effectively a ﬁnite length. The ﬁelds generated by this source are no longer independent of . Working in domain we use the Fourier transform to get the spectral (19) The ﬁelds generated by this source vary along as instead of being independent of as before. The surface wave modes and pseudomodes are now a mixture of both TM and TE to parts. As it is well known the TE surface wave exists . Considering ﬁrst the TM to ﬁelds all only when components can be derived from only. Similar to the mode as expansion in (1) we can write

(25) A similar expression exists for the TE ﬁeld component with an extra term inside the integral term. This manifests the . fact that this component vanishes when and to get [12] Expression (25) can be evaluated for

(20) whose -inverse Fourier transform is (26) (21) In (20), and of the electric ﬁeld of the pseudomode with transverse tudinal wavenumbers are: . are the -component th surface wave mode and a wavenumber . The longiand where is the Gaussian beam width at a distance from the slot, and . A similar expression is obtained for the TE ﬁeld component that has an extra mul. It is seen that the Gaussian beamwidth tiplying increases linearly with and the ﬁeld magnitude decays with as as expected for the diffraction of a Gaussian beam [12].

2062

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Now turning attention to the surface wave power, we can write (27) should now be taken as the sum of of the TM where and TE ﬁeld. Note however that the TE mode does not exist . when Substituting from (25) in (27), we get after some manipulations (28) where

Fig. 7.Surface wave conductance versus F for a Gaussian excited slot with k w = =2 and . Two values of " ; 3.0 and 9.8, are considered.

It is instructive to note that in the limit , tends tends to zero. In this case, (28) reduces to (12) to one and which applies to an inﬁnite uniformly excited slot. Thus, the and account for the Gaussian distribution of the terms source. This suggests that we deﬁne a Gaussian SW efﬁciency . Namely equal to

=2 and and " = 9:8 and 3.0.

Fig. 8. Gaussian efﬁciency versus F for different slot Gaussian width k

w=

(29) We can deﬁne an effective length of the Gaussian source as the length of a uniformly excited slot that would produce the same TM surface wave power as the Gaussian excited slot. Denoting , we have from (28) that length by (30) A plot of the surface wave conductance versus and different normalized frequency is shown for different normalized Gaussian width in Fig. 7. Here is the effective wavenumber on the dielectric slab and is taken equal to . It is worth noting here that the TE mode contri. For the example given bution to the power is zero for in Fig. 7, the contribution of the TE mode is less than 5% up to for and less than 8.5% for . As it is the case with uniform excitation (Fig. 5), the surface wave . It increases with except conductance peaks around for low values of . The Gaussian efﬁciency parameter deﬁned , which is in [(29), (30)] is plotted in Fig. 8. It is seen that , increases with up to a saturaalso equal to depends on the value of as tion level. The variation with seen in the ﬁgure.

VI. CONCLUSION Rigorous analysis of surface wave excitation and radiation from a grounded dielectric slab driven by a slot source has been presented. The analysis is relevant to the design of quasioptical slab beam power combiners that use surface waves to transport power on a dielectric slab. Adopting a 2-D model of the slab and the slot, closed form expressions for the surface wave and radiation powers have been derived. In addition, slot self-admittance and mutual admittance of two parallel slots have been derived. This facilitates the design of Yagi slot arrays aiming at achieving of excited surface waves. maximum front to back ratio ratio better than 20 dB can Numerical results show that be achieved over a bandwidth of 4%. The analytical results are supported by previously published simulation and experimental work by the authors [4]. In order to improve the 2-D model, the case of a Gaussian -excited slot is treated. Analysis shows that the surface wave power decays linearly with distance traveled along the slab for much greater than the Gaussian beamwidth. An effective length of the Gaussian slot is derived. Although the present theory has been applied to a single homogeneous dielectric slab, extension to an inhomogeneous slab, or a multiplayer slab, is straightforward. Such extension should lead to the study of the interplay between surface wave and radiated powers on printed circuits. This study is underway.

MAHMOUD et al.: THEORETICAL CONSIDERATIONS IN THE OPTIMIZATION OF SURFACE WAVES

2063

REFERENCES

[1] J. Harvey, E. R. Brown, D. B. Rutledge, and R. A. York, “Spatial power combining for high-power transmitters,” IEEE Microwave Mag., pp. 48–59, Dec. 2000. [2] A. R. Perkons, Y. Qian, and T. Itoh, “TM surface wave power combining by planar active-lens ampliﬁer,” IEEE Trans. Microwave Theory Tech., vol. 46, pp. 775–783, June 1998. [3] H. F. Hammad, A. P. Freundorfer, and Y. M. M. Antar, “CPW slot antenna for TM slab mode excitation,” presented at the Proc. IEEE Antenna and Propagation and URSI Int. Symp., Boston, MA, July 2000. [4] H. F. Hammad, Y. M. M. Antar, A. P. Freundorfer, and S. F. Mahmoud, “Uni-planar CPW-fed slot launchers for efﬁcient TM0 surface wave excitation ,” IEEE Trans. Microwave Theory Tech, vol. 51, pp. 1234–1240, Apr. 2003. [5] E. Bahar, “Scattering of VLF radio waves in curved earth-ionosphere waveguide,” Radio Sci., vol. 3, no. 2, pp. 145–154, 1968. [6] V. V. Shevchenco, Continuous Transitions in Open Waveguides: The Golem Press, 1971, ch. 1, sec. 1–7. [7] S. F. Mahmoud and J. C. Beal, “Scattering of surface waves at a dielectric discontinuity on a planar waveguide,” IEEE Trans. Microwave Theory Tech., vol. MTT-23, pp. 193–198, 1975. [8] S. F. Mahmoud, “Electromagnetic waveguides; theory and applications,” in IEE Electromagnetic Waves Series 32. Stevenage, U.K.: Peregrinus, 1991, sec. 4.3. [9] J. R. Wait, Electromagnetic Waves in Stratiﬁed Media. New York: Pergamon, 1970, ch. 6, pp. 33–35. [10] J. A. Fuller and J. R. Wait, “A pulsed dipole in the earth,” J. Appl. Phys., vol. 10, pp. 238–270, 1976. [11] A. K. Bhattacharyya, “Characteristics of space and surface waves in a multilayered structure,” IEEE Trans. Antennas Propagat., vol. 38, pp. 1231–1238, Aug. 1990. [12] L. C. Shen and J. A. Kong, Applied Electromagnetism, 2nd ed. Boston, MA: PWS-Kent, 1987, sec. 8.3.

ACKNOWLEDGMENT S. F. Mahmoud acknowledges the support of Kuwait University for providing him with Sabbatical leave to perform this research.

Yahia M. M. Antar (S’73–M’76–SM’85–F’00) was born on November 18, 1946, in Meit Temmama, Egypt. He received the B.Sc. (Hons.) degree in 1966 from Alexandria University, Egypt, and the M.Sc. and Ph.D. degrees from the University of Manitoba, Winnipeg, Canada, in 1971 and 1975, respectively, all in electrical engineering. In 1966, he joined the Faculty of Engineering at Alexandria University,where he was involved in teaching and research. At the University of Manitoba he held a University Fellowship, an NRC Postgraduate and Postdoctoral Fellowships. From 1976 to 1977, he was with the Faculty of Engineering, University of Regina. In June 1977, he was awarded a Visiting Fellowship from the Government of Canada to work at the Communications Research Centre, Department of Communications, Shirley’s Bay, Ottawa, where he was involved in research and development of satellite technology with the Space Electronics group. In May 1979, he joined the Division of Electrical Engineering, National Research Council of Canada, Ottawa, where he worked on polarization radar applications in remote sensing of precipitation, radio wave propagation, electromagnetic scattering and radar cross section investigations. In November 1987, he joined the staff of the Department of Electrical and Computer Engineering, Royal Military College of Canada, Kingston, ON, Canada, where he is now a Professor of electrical and computer engineering. He is presently the Chairman of the Canadian National Commission (CNC), International Scientiﬁc Radio Union (URSI), holds adjunct appointment at the University of Manitoba, and has a cross appointment at Queen’s University in Kingston. He has authored or coauthored over 100 journal papers on these topics, and supervised or cosupervised over 45 Ph.D. and M.Sc. theses at the Royal Military College and Queen’s University, of which three have received the Governor General Gold Medal. His current research interests include polarization studies, integrated antennas, microwave, and millimeter wave circuits. Dr. Antar is a Fellow of the Engineering Institute of Canada (FEIC). He received the 2003 RMC Excellence in Research Prize. In May 2002, he became the holder of a Canada Research Chair (CRC) in Electromagnetic Engineering. He is an Associate Editor of the IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION and Associate Editor (Features) of the IEEE ANTENNAS AND PROPAGATION MAGAZINE.

Samir F. Mahmoud (S’69–M’73–SM’83) graduated from the Electronic Engineering Department, Cairo University, Cairo, Egypt, in 1964 and received the M.Sc. and Ph.D. degrees from the Electrical Engineering Department, Queen’s University, Kingston, ON, Canada, in 1970 and 1973, respectively. During academic year 1973 to 1974, he was a Visiting Research Fellow at the Cooperative Institute for Research in Environmental Sciences (CIRES), Boulder, CO, doing research on communication in tunnels. He spent two sabbatical years, 1980 to 1982, between Queen Mary College, London and the British Aerospace, Stevenage, U.K., where he was involved in design of antennas for satellite communication. Currently he is a Full Professor at the Electrical Engineering Department, Kuwait University. Recently, he has visited several places including Interuniversity Micro-Electronics Centre (IMEC), Leuven, Belgium, and spent a sabbatical leave at Queen’s University and the Royal Military College of Canada, Kingston, ON Canada, from 2001 to 2002. His research activities have been in the areas of antennas, geophysics, tunnel communication, e.m wave interaction with composite materials and microwave integrated circuits. Dr. Mahmoud is a Fellow of the Institution of Electrical Engineers (IEE), London, U.K. He was a recipient of the Best IEEE/ Microwave Theory Technology Paper for 2003.

Hany F. Hammad received the B.Sc. degree with honors from Ain Shames University, Cairo, Egypt, in 1994 and the M.Sc. and Ph.D. degrees from Queen’s University, Kingston, ON, Canada, in 1997 and 2002, respectively. His Ph.D. thesis was ranked as the “Outstanding Thesis of Engineering and Applied Science Division” at Queen’s University. His research areas of interests are the analysis and design of antennas and microwave integrated circuits.

Al P. Freundorfer (M’90) received the B.A.Sc., M.A.Sc., and Ph.D. degrees from the University of Toronto, ON, Canada, in 1981, 1983, and 1989, respectively. In 1990, he joined the Department of Electrical Engineering, Queen’s University, Kingston, ON, Canada. Since then he has done work in nonlinear optics of organic crystals, coherent optical network analysis as well as microwave integrated circuits. Currently he is focusing his attention on monolithic microwave circuits used in lightwave systems with bit rates in excess of 20 Gb/s and on monolithic millimeter wave integrated circuits used in wireless communications.

2064

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**Generalized System Function Analysis of Exterior and Interior Resonances of Antenna and Scattering Problems
**

Long Li and Chang-Hong Liang, Senior Member, IEEE

Abstract—The generalized system function ( ), directly associated with radiated and scattered ﬁelds, is presented to effectively analyze the exterior and interior resonances of antenna and scattering systems in this paper. ( ) is constructed by using the model-based parameter estimation technique combined with the complex frequency theory. The behaviors of the exterior and interior resonances can be distinguished by analyzing the characteristics of pole-zero of ( ) in a ﬁnite operational frequency band. The intensity of the exterior resonance can be effectively estimated in terms of values and residues at the complex resonant frequencies. The truly scattered ﬁelds from a closed conducting region can be obtained by eliminating the poles corresponding to the interior resonances from ( ). Some examples of the practical antenna arrays and scattering systems are given to illustrate the application and validity of the proposed approach in this paper. Index Terms—Generalized system function, exterior and interior resonances, complex resonant frequency, factor, model-based parameter estimation (MBPE).

I. INTRODUCTION

W

ITH the increasingly complicated electromagnetic environment, the interaction and mutual coupling between antennas and scatterers become more and more severe that sometimes give rise to the strong electromagnetic oscillation phenomena. Therefore, the study of resonance behaviors in the electromagnetic compatibility (EMC) has been an interesting and challenging problem for years [1]–[6]. The -ﬁeld integral equation (EFIE) and the -ﬁeld integral equation (MFIE) have been used extensively to analyze antennas radiation and scattering from perfectly conducting bodies. It is well known that bodies with closed conducting regions can support interior resonance at certain discrete frequencies where both the -ﬁeld and -ﬁeld integral equations fail to calculate the scattered (external) ﬁeld [7]. Theoretically, the undeterminable component of the surface current associated with the cavity mode does not radiate. However, due to truncation error and numerical error effects, at these frequencies the cavity mode is both very weakly excited and radiated very weakly, so the matrix problem was found to have a different structure from that of the functional equation problem [8]. Some techniques

have been proposed for dealing with these numerical problems [8]–[12]. Most of these methods generate a system of equations that has a unique solution for the current and external ﬁelds at all frequencies. presented in circuit The complex resonant frequency theory [13], [14] is ﬁrstly introduced to antenna and scattering systems in this paper, which relates the real resonant frequency with radiated or scattered losses. The generalized , directly associated with radiated and system function, scattered ﬁelds, is presented to effectively analyze the exterior and interior resonances of the antenna and scattering problems in this paper, which is constructed by using model-based parameter estimation (MBPE) technique. The MBPE [15]–[19] is a form of “smart” curve ﬁtting, with broad applications to a fast analysis of radiation patterns or RCS of antennas or scatterers in a widely operating bandwidth. By analyzing the characteristics of poles and zeros of, we can determine the exterior and interior resonant frequencies of antenna and scattering systems efﬁciently. The complex frequency method for is also presented calculating antenna or scattering external in this paper. Furthermore, The exterior resonance strength can be effectively estimated by the values of and residues at the complex resonant frequencies. The truly scattered ﬁelds from a closed conducting region can be got by eliminating the poles corresponding to the interior resonances from the generalized system function. Some examples and discussion, parallel dipoles antenna, two conducting objects scattering system and an inﬁnitely long elliptical cylinder scattering problem are given in this paper. II. COMPLEX RESONANT FREQUENCY For an arbitrary lossy resonant system, the complex resonant frequency [13], [14] can be introduced and written as (1) is a real resonant frequency of the system, rewhere pressents the losses of the resonant system. In general sense, the electric ﬁeld can be written as (2)

Manuscript received July 11, 2003; revised October 25, 2003. This work was supported by the National Natural Science Foundation of China under Contract 69931030. The authors are with the School of Electronic Engineering, Xidian University, Xi’an 710071, Shaanxi, China; (e-mail: Lilong@mail.xidian.edu.cn; Chhliang@xidian.edu.cn). Digital Object Identiﬁer 10.1109/TAP.2004.832319

With the presence of the losses, the energy stored in the resonant system will decay at a rate proportional to the average energy presented at any time, so that (3)

0018-926X/04$20.00 © 2004 IEEE

LI AND LIANG: GENERALIZED SYSTEM FUNCTION ANALYSIS OF EXTERIOR AND INTERIOR RESONANCES

2065

where is the average energy present at . But the rate of must equal the power loss, so that decrease of (4) In addition, an important parameter specifying selectivity, and performance in general, of a resonant system is the quality factor, . A general deﬁnition of applicable to all resonant system is (5) Substituting (5) into (4), we can easily get (6) Therefore, the general expression of is

and represent the where coefﬁcients of numerator and denominator polynomials, respectively. Note that or can be normalized to 1 in denominator coefﬁcients. Thus, (8) has unknown complex coefﬁcients. represents the complex frequency . It is obvious that MBPE utilizes the rational function approximation and extends it into complex frequency domain, which provides an appropriate tool for analyzing the resonance characteristics of antenna and scattering systems from the point of view of complex frequency. According to the uniform approximation theory [21], the error of MBPE interpolation is minimum when or , and the properties of existence and uniqueness of rational function approximation can be demonstrated [22]. Based on the theory of signals and systems, we know a particularly important and useful class of linear time-invariant (LTI) systems is those for which the input and output satisfy a linear constant-coefﬁcient differential equation of the form [23]

(9) (7) It can be seen that the introduction of uniﬁes the resonant of a resonant system, and each complex frequency and resonant frequency corresponds to one resonant mode. It is well known that antenna or scattering system is essentially equivalent to a lossy network. Assume the system media are represents the radiated or scattered power lossless, the loss from the antennas or scattering bodies. The average stored energy denotes the sum of stored electric ﬁeld and magnetic ﬁeld energies around the antennas or scatterers, which is independent of the radiated energies from the antennas or scatterers [20]. Therefore, the complex resonant frequency is applicable to not only the resonant cavity in the closed system, but also the antenna and scattering resonant problems in the open system. If the system media are lossless, the complex resonant frequency corresponding to a nonradiated mode (cavity mode) will . reduce to the real resonant frequency III. GENERALIZED SYSTEM FUNCTION CONSTRUCTED BY MBPE The MBPE is a smart curve ﬁtting technique [15]–[19], which has been widely applied to the fast analysis of radiation patterns or RCS of antennas or scatterers over a wide frequency band. MBPE makes use of low-order analytical formulas as ﬁtting models, while the unknown coefﬁcients for the ﬁtting model are obtained by matching it to multipoint sampled values [18] or ﬁtting it to frequency derivatives of the function at one or two frequencies [15]. In this paper, MBPE is mainly used to construct the generalized system function associated with electromagnetic ﬁelds in the complex frequency domain. According to the observed objects, one form of a ﬁtting model that is commonly employed in MBPE is represented by Padé rational function as follows: (8) and represent the input and output time funcwhere tions, respectively. Taking the Laplace transform of (9), we obtain

(10) is commonly referred to as the system function or, alternatively, the transfer function. Many properties of LTI systems are closely associated with the characteristics of the system function in the plane. It is very interesting that (10) is consistent with (8) formed by MBPE in mathematical representation. In physical sense, (10) represents the system function, which is the Laplace transform of impulse response of LTI systems. In the analysis of antenna or scattering electromagnetic systems, the ideal source models [24] of voltage, current, or unit plane wave are commonly utilized as the excitation functions, and the frequency responses of antenna properties, such as the current distribution , input impedance , radiation patterns , , etc., can be thought of as RCS, or near ﬁelds the output functions. In this case, the output functions just correspond to the impulse responses of the antenna or scattering system in time domain. If we make use of MBPE technique to approximate the output function frequency responses, (8) is characterized by the system function. Therefore, the generaldirectly associated with the radiated ized system function or scattered ﬁelds can be constructed by MBPE technique in a limited operational bandwidth with a model containing a ﬁnite number of suitably chosen complex poles, which describes the intrinsic characteristics of the antenna or scattering systems. ) Equation (8) can be further factored into the form (let

(11)

2066

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

where is scale factor and are the complex poles and zeros of the generalized system function, rerepspectively. We know the denominator polynomial of resents the characteristic polynomial of antenna or scattering systems. The zeros of the denominator polynomial, namely the , deﬁne the locations of the natural resonances of poles of antenna or scattering systems, involving the exterior and interior resonances. It is worth pointing out that only the exterior resonance can be interpreted as intrinsic to the scatterer in principle. According to the stability of the electromagnetic systems, we know the true poles of the systems should reside in the left half of the complex frequency plane. The validity of the complex poles obtained by (11) will be discussed in the following section. It is assumed that the poles are all simple. This has been numerically substantiated. A partial fraction expansion yields (12) where represents the complex pole and is the corresponding residue. Therefore, many properties of antenna and scattering electromagnetic systems can be characterized by a few pole locations with the corresponding residues. When the antenna and scattering systems are regarded as the multiport networks, assumed input ports and output ports, the generalized system function matrix can be similarly constructed by MBPE based on the linear superposition principle, which can be expressed as

Fig. 1. Two parallel dipoles system.

. . .

. . .

. . .

. . .

. . . (13)

Fig. 2. Frequency response of system function (far E -ﬁeld) magnitude.

Namely (14) where represents the generalized system function mais referred to as the subsystem function. All trix, and true poles of the generalized system function matrix deﬁne the natural resonances of the antenna or scattering systems. is a square matrix when , and thus the poles are the solutions to the following: (15)

IV. EXTERIOR AND INTERIOR RESONANT FREQUENCIES AND FACTOR We know that a ﬁnite number of suitably chosen complex poles of deﬁne the natural resonances of the antenna or scattering electromagnetic open systems. Assume the pole , the corresponding partial fraction of the generalized system function can be expressed as (17) The time response corresponding to the complex pole is (18)

where symbol indicates taking the determinant of matrix. While combining MoM with MBPE technique deals with the antenna or scattering problems, the generalized impedance macan be obtained in -plane in terms of the theory of trix generalized networks [24], which describes the intrinsic characteristics of the system structures and is independent of the complicate excitations and loads. Therefore, the complex poles of the antenna or scattering systems are determined by (16)

Comparing (18) with (2), we can see that the residue represents the complex magnitude of the electric ﬁeld at , if at one the complex frequency response of the electric ﬁeld point in space is chosen as the generalized system function. The relationship of the complex pole with the complex resonant frequency presented in the previous section is (19)

LI AND LIANG: GENERALIZED SYSTEM FUNCTION ANALYSIS OF EXTERIOR AND INTERIOR RESONANCES

2067

TABLE I ZEROS, POLES, AND RESIDUES OF THE GENERALIZED SYSTEM FUNCTION

Fig. 3. Frequency response of the near E -ﬁeld magnitude.

Therefore, the resonant frequency and open system can be easily obtained

of the electromagnetic

(20) Obviously, by calculating the complex poles of the generalized based on the physical models, we can disystem function rectly get the resonant frequency and corresponding . It is well-known that the quantitative analysis of electric and magnetic ﬁeld energies stored in the near-ﬁeld zone of the antennas or scatterers is very difﬁcult to give, and thus the calculation of antenna or scattering external has also been an interesting and challenging problem for years [20], [25]–[27]. In this paper, the complex frequency method combined with the generalized system function is used to calculate the antenna or scattering efﬁciently, which has been illustrated by the later numerical tests. For a scattering from a closed perfect conducting region, the total current ﬂowing on the surface is not determined by EFIE or MFIE and the incident external ﬁeld at the interior resonant frequencies. Theoretically, the undeterminable component of the surface current that associated with the cavity mode does not radiate. Therefore, the poles corresponding to the interior resonances should locate on the positive imaginary axis of the plane in principle. However, due to truncation error and numerical error effects, the cavity mode is both very weakly excited and radiated very weakly [8]. So these poles do not strictly occur on the imaginary axis of the plane but reside in the left half of the plane off imaginary axis very small. To get the truly scattered ﬁeld, these poles corresponding to the interior resonances must be eliminated.

Fig. 4. Comparison of the electric ﬁeld magnitude distribution around dipoles at resonance with nonresonance (a) resonance and (b) nonresonance.

By analyzing the characteristics of poles and zeros of the generalized system function and combining with adaptability of MBPE, we can accurately predict the occurrence of resonance phenomena and determine the exterior and interior resonant frequencies of the antenna and scattering systems. The intensity of resonance can be effectively estimated by the values of and residues at the complex resonant frequencies. Only when both the external resonant and the residues are larger, are the resonance phenomena characterized by the strong peak ﬁeld in the near region and large frequency sensitivity in the far ﬁeld region of the antennas and scattering bodies. It should be pointed out that the complex poles referred above must be the true and stable poles of the antenna or scattering electromagnetic systems. A discussion on the validity of the poles of the generalized system function constructed by MBPE is given here. On the one hand, according to the stability of the practical antenna and scattering systems, these complex poles

2068

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

must locate in the left half of the plane. Therefore, the poles occurred in the right half of the plane must be invalid poles of the system. On the other hand, in MBPE, if the Padé rational functions with different numerator and denominator orders are used to construct the generalized system function, we might get some different complex poles. One knows that the true poles are corresponding to the complex natural exterior resonant frequency of systems, which should be independent of the form of ﬁtting function and the orders of the rational function. Thus, the locations of the true poles are stable or invariant. However, the other poles besides those residing in the right half of the plane will vary with the orders of the rational functions, which must be invalid poles and referred to as “parasitical” poles. In addition, as previous discussion, the poles corresponding to the interior resonances are not strictly occur on the imaginary axis of the plane but reside in the left half of the plane off imaginary axis very small. The truly external scattered ﬁeld can be obtained by eliminating these poles from (12). V. APPLICATIONS AND DISCUSSION In the following examples, some special resonance behaviors will be analyzed by the generalized system functions directly associated with the radiated or scattered ﬁelds of some local regions, involving far ﬁelds and near ﬁelds. The results of numerical tests show that the exterior resonance phenomena are very remarkable by virtue of the strong interaction and mutual coupling between antennas or scatterers, and the interior resonance behaviors in EFIE give rise to a false scattered ﬁeld. A. Test 1 Parallel Dipole Antennas Consider the two parallel dipoles system shown in Fig. 1. Dipole 1 will be excited by the ideal voltage source, and the terminal of dipole 2 shorted. The length both of them is m, with radius m. The distance between them is m. The frequency response of the radiated electric in the far zone ﬁeld of observation point at is chosen as the output function, i.e., the generalized system . The MBPE technique is applied to the antennas function system over a frequency range of 15–25 MHz, using the radiated electric ﬁeld data obtained from a numerically rigorous method of moments (MoM) computer codes based on EFIE. The Padé rational function is chosen to set the numerator order and the denominator order . Fig. 2 shows the frequency response of the generalized system function constructed by MBPE, with comparisons being made of the MoM result. As can be seen from Fig. 2, the two curves are nearly graphically indistinguishable. In this case, only six sampling frequencies are required for the MBPE technique. The actual sampling points that were used are indicated by dots on the plots contained in Fig. 2. It is interesting that all of the ﬁtting frequencies are sampled before the resonant frequency, but the resonant behavior can be found efﬁciently due to the adaptability of MBPE technique. It is worth pointing out that MoM direct calculation using 102 unknowns took 28 s to calculate the frequency response at 100 frequency values from 15 to 25 MHz. MBPE took a total of 2 s to generate the solutions with 0.1 MHz increment, poles and zeros.

Fig. 5.

Two conducting bodies scattering system.

Fig. 6. Frequency responses of the generalized system function matrix (scattered near ﬁeld. (a) E component and (b) E component.

The characteristics of zeros, poles, and corresponding residues of the generalized system function are shown in Table I. Note that the data in the table have been transformed from to (MHz). The facts show that there are two true and stable complex poles in the antenna system within the ﬁnite operation frequency band, which are marked by asterisks in the Table I. On the basis of the complex frequency theory presented in the previous section, from (20), the exterior resonant frequencies and of the two parallel dipoles system are obtained, respectively, as follows:

LI AND LIANG: GENERALIZED SYSTEM FUNCTION ANALYSIS OF EXTERIOR AND INTERIOR RESONANCES

2069

TABLE II ZEROS, POLES, AND RESIDUES OF THE SUB-SYSTEM FUNCTION

H (s)

TABLE III ZEROS, POLES, AND RESIDUES OF THE SUB-SYSTEM FUNCTION

H (s )

Fig. 8. Plane wave incident upon an inﬁnite long perfect conducting elliptical cylinder.

presented in [20], [24], based on the Foster for ﬁnding reactance theorem

Fig. 7. Scattered electric ﬁeld magnitude distributions in and plane and (b) plane. at resonance (18.169 MHz). (a)

XZ

YZ

XZ

Y Z plane

It can be found that the exterior resonance behavior to occur at the frequency 19.6921 MHz with high , as shown in Fig. 2. It is worth pointing out that the calculation of antenna is deﬁnite and efﬁcient by using the complex frequency method. To demonstrate the validity of , a classical formula

has been used to calculate the antenna of the two parallel dipoles system. Utilizing a ﬁrst-order accurate difference approximation to the partial frequency derivative of the reactance matrix , we obtain at the resonant frequency 19.6921 MHz, which is very closed to the result of the complex frequency method. It can be seen that the other resonant mode, 18.4607 MHz, makes a little contribution to the resonance behavior in this case for the low . To further understand the behavior of the resonance, we calculated the frequency response of the electric ﬁeld magnitude at the observation point in the vicinity of the dipole 1 indicated by dot in Fig. 1. It can be seen that the behavior of the resonance is also remarkably embodied by the phenomenon of strong peak ﬁeld in the near zone of the antenna system, as shown

2070

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE IV ZEROS, POLES, AND RESIDUES OF THE GENERALIZED SYSTEM FUNCTION (E

)

in Fig. 3. The comparison of the electric ﬁeld magnitude distribution in the xz-plane in the near zone of the two parallel dipoles system at resonance with nonresonance is given in the Fig. 4(a) and (b), respectively. It should be noted that the electric ﬁelds are much stronger at resonance than those at nonresonance in the same excitation, but only accumulating in the vicinity of the dipoles. The symmetric distribution of the electric ﬁelds magnitude at resonance implies the balance of the electric ﬁeld energy and magnetic ﬁeld energy stored in the antenna open system physically. B. Test 2 Two Conducting Bodies Scattering System Consider the two perfectly conducting bodies scattering system shown in Fig. 5, which is excited by the normalized -polar plane wave , and the direction of propagation is . The sizes of two perfect conducting bodies are m, m, m, and m apart. The scattered the two conducting bodies are , in electric ﬁelds of the observation point, the near zone of the scattering bodies are chosen as the output functions. Because the scattered electric ﬁeld in the near region and , the generalized has two main components system function matrix can be constructed by MBPE technique in two directions in order to analyze the exterior resonance characteristics of near ﬁelds. The Padé rational function is and denomichosen to have the same numerator order . From Fig. 6, it can be seen that the solid line nator order calculated by MoM is mostly hidden by the MBPE curve. The zeros, poles and residues of the subsystem functions of and directions are calculated and shown in Tables II and III, respectively. From Tables II and III, we found that there exist two true and stable complex poles in the scattering system, which represent the external natural resonances and marked by asterisks respectively. Based on the theory of the complex frequency, we can easily get the exterior resonant frequencies and scattering as follows:

Fig. 9.

Frequency response of backscattered RCS.

electric ﬁelds accumulate mainly in the region between the two conducting bodies, being strong at both sides and weak at center, with stand-wave-like distribution. C. Test 3 Inﬁnitely Long Elliptical Cylinder Scattering Problem Consider an inﬁnite long perfect conducting elliptical cylinder scattered by transverse magnetic (TM) plane wave incident in the direction, as shown in Fig. 8. The cross section is an ellipse with semimajor axis 1.0 meter and semiminor axis 0.25 meter. A method of moments formulation of the EFIE is used to obtain the scattered electric ﬁeld. We have utilized pluses as expansion functions and delta functions as weighting functions. Using 180 unknowns, we numerically found two interior resonances to occur at 327.33 and 435.42 MHz within the bandwidth from 200 to 500 MHz. The radar cross section (RCS) for backscatter as a function of the frequency for the elliptical cylinder is shown in Fig. 9. It can be seen that the backscattered ﬁelds are uncorrected in the frequencies near the interior resonances. The locally magniﬁed ﬁgures show clearly the RCS has a sharp dip at the two interior resonant frequencies. We choose the frequency response of the backscattered elec. The MBPE tric ﬁeld as the generalized system function, technique is applied to the scattering problem over a frequency range 200–500 MHz, using the backscattered electric ﬁeld data

According to the values and the corresponding residues, we can estimate that the resonance phenomenon of strong peak near ﬁeld would appear at the frequency 18.169 MHz, as shown in Fig. 6. The magnitude distributions of the scattered electric ﬁeld and plane at resonance (18.169 MHz) are shown in in Fig. 7(a) and (b), respectively. It is interesting that the scattered

LI AND LIANG: GENERALIZED SYSTEM FUNCTION ANALYSIS OF EXTERIOR AND INTERIOR RESONANCES

2071

and ﬁtting to the sampling values accurately models both resonances. MoM direct solution took 4175 s to calculate the frequency response with 0.01 MHz increment from 200 to 500 MHz. The MBPE technique took a total of 32 s to generate the solutions, poles, and zeros. Table IV shows the zeros, poles, and corresponding residues of the generalized system function. By analyzing the characteristics of the poles and combining with the complex frequency theory discussed previously, we found two “true” complex poles occur in the left half of the plane but off the imaginary axis very small, which just deﬁne the interior resonant frequencies and are marked by asterisks in the Table IV. It can be seen the residues corresponding to the two poles are very small, which imply the scattered contribution from the interior resonances should become especially small. As pointed out in [8], the scattered contribution results from the numerical error in EFIE. To get the truly scattered ﬁeld, we modify the generalized system function by eliminating those poles corresponding to the interior resonances from (12). The comparison of the modiﬁed system function response with the results calculated by the combined ﬁeld integral equation (CFIE) is shown in Fig. 10. The two curves are nearly graphically indistinguishable, and the phenomena of the interior resonance are removed successfully. VI. CONCLUSION This paper has presented the generalized system function which directly associate with radiated and scattered ﬁelds to give an efﬁcient analysis of the exterior and interior resois constructed nances of antenna and scattering problems. by using the MBPE technique combined with the complex frequency theory. The behaviors of the exterior and interior resonances can be distinguished by analyzing the characteristics in a ﬁnite operational frequency band. of pole-zero of The exterior complex resonant frequencies must reside in the left half of the plane off the imaginary axis. The imaginary is related to the radiated or scattering losses, i.e., part of . The intensity of the exterior resonance can be estimated values and residues at the complex effectively in terms of resonant frequencies. The interior resonant frequencies occur on the positive imaginary axis of the plane theoretically, but due to truncation error and numerical error effects, the internal (cavity) modes are both very weakly excited and radiated very weakly. Therefore, those poles corresponding to interior resonances also locate in the left half of the plane but off imaginary axis very small. It is shown that only exterior poles can be interpreted as intrinsic to the scatterer. The truly scattered ﬁelds from a closed conducting region can be obtained simply by eliminating the poles corresponding to the interior resonances from the generalized system function. REFERENCES

[1] H. C. Pocklington, “Electrical oscillations in wires,” in Proc. Cambridge Phil. Soc., vol. 9, 1897, pp. 324–332. [2] L. Page and N. Adams, “The electrical oscillations of a prolate spheroid, paper I,” Phys. Rev., vol. 65, pp. 819–831, 1938. [3] J. A. Stratton, Electromagnetic Theory. New York, NY: McGraw-Hill, 1941.

Fig. 10. Comparison of the modiﬁed system function response with the results calculated by CFIE and EFIE: (a) Full frequency band, (b) near interior resonant frequency 1, and (c) near interior resonant frequency 2.

obtained from the MoM formulation of EFIE, which are transformed into RCS and indicated by circles on the plots contained in Fig. 9. The Padé rational function is chosen to have and a denominator order . a numerator order In this case, the MBPE-calculated curve is excellent agreement with the MoM result, including those in the vicinity of the interior resonances. It is worth pointing out that MoM requires a very larger number of evaluations in order to resolve these interior resonance behaviors. However, the MBPE with

2072

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[4] C. H. Liang and L. Li, “On the generalized resonance,” presented at the Proc. China-Japan Symp. Microwave, China, Apr. 2002. Invited Talk, Xi’an. [5] J. Mouping, “Study of generalized resonance in electromagnetic scattering by multiple conductors,” Ph.D. Dissertation, School of Elect. Eng., Xidian Univ., 2000. [6] L. Li and C. H. Liang, “Study of generalized resonance in antenna system,” in Proc. 3rd Int. Symp. Electromagnetic Compatibility, Beijing, China, May 2002, pp. 162–165. [7] J. R. Mautz and R. F. Harrington, “H-ﬁeld, E-ﬁeld, and combined ﬁeld solutions for conducting bodies of revolution,” A.E.Ü., vol. 32, no. 4, pp. 157–164, Apr. 1978. [8] F. X. Canning, “Singular value decomposition of integral equations of EM and applications to the cavity resonance problem,” IEEE Trans. Antennas Propagat., vol. AP-37, pp. 1156–1163, Sept. 1989. [9] R. Mittra and C. A. Klein, “Stability and convergence of moment method solutions,” in Numerical and Asymptotic Techniques in Electromagnetics, R. Mittra, Ed. New York: Springer Verlag, 1975. [10] T. K. Sarkar and S. M. Rao, “A simple technique for solving E-ﬁeld integral equations for conducting bodies at internal resonances,” IEEE Trans. Antennas Propagat., vol. AP-30, pp. 1250–1254, June 1982. [11] J. R. Mautz and R. F. Harrington, “A combined-source solution for radiation and scattering from a perfectly conducting body,” IEEE Trans. Antennas Propagat, vol. AP-27, pp. 445–454, Apr. 1979. [12] F. X. Canning, “Protecting EFIE-based scattering computations from effects of interior resonances,” IEEE Trans. Antennas Propagat, vol. 39, pp. 1545–1552, Nov. 1991. [13] R. E. Collin, Foundations for Microwave Engineering. New York: McGraw Hill, 1966. [14] C. H. Liang and Y. J. Xie, “The accurate variational analysis for the measurement of the complex dielectric constant of a sample rod inserted in a cavity,” Microwave Opt. Technol. Lett., vol. 5, no. 5, pp. 209–211, 1992. [15] C. J. Reddy, “Application of model based parameter estimation for RCS frequency response calculations using method of moments,” NASA/CR1998-206 951, Mar. 1998. [16] E. K. Miller and G. J. Burke, “Using model-based parameter estimation to increase the physical interpretability and numerical efﬁciency of computational electromagnetic,” Comput. Phys. Commun., vol. 68, pp. 43–75, 1991. [17] C. J. Reddy, “Application of model based parameter estimation for fast frequency response calculations of input characteristics of cavity-backed aperture antennas using hybrid FEM/MoM technique,” NASA/CR-1998-206 950, Mar. 1998. [18] D. H. Werner and R. J. Allard, “The simultaneous interpolation of antenna radiation patterns in both the spatial and frequency domains using model-based parameter estimation,” IEEE Trans. Antennas Propagat., vol. 48, pp. 383–392, Mar. 2000. [19] R. J. Allard and D. H. Werner, “The model-based parameter estimation of antenna radiation patterns using windowed interpolation and spherical harmonics,” IEEE Trans. Antennas Propagat., vol. 51, pp. 1891–1906, Aug. 2003. [20] W. Geyi, P. Jarmuszewski, and Y. Qi, “The foster reactance theorem for antennas and radiation ,” IEEE Trans. Antennas Propagat., vol. 48, pp. 401–407, Mar. 2000.

[21] S. V. Polstyanko, R. Dyczij-Edlinger, and J. F. Lee, “Fast frequency sweep technique for the efﬁcient analysis of dielectric wave-guides,” IEEE Trans. Microwave Theory Tech., vol. 45, pp. 1118–1126, June 1997. [22] P. J. Davis, Interpolation and Approximation. London, U.K.: Blaisdell, 1963. [23] A. V. Oppenheim, A. S. Willsky, and I. T. Young, Signals and Systems. Englewood Cliffs, NJ: Prentice-Hall, 1983. [24] R. F. Harrington, Field Computation by Moment Method. New York: IEEE Press, 1993. [25] R. L. Fante, “Quality factor of general ideal antennas,” IEEE Trans. Antennas Propagat., vol. AP-17, pp. 151–155, Feb. 1969. [26] J. S. McLean, “A re-examination of the fundamental limits on the radiation of electrically small antennas,” IEEE Trans. Antennas Propagat., vol. 44, pp. 672–675, May 1996. [27] R. E. Collin, “Minimum of small antennas,” J. Electromagn. Waves Applicat., vol. 12, pp. 1369–1393, 1998.

Q

Q

Long Li was born in Anshun, Guizhou, China, in January 1977. He received the B.Eng. degree in electromagnetic ﬁeld and microwave technology from Xidian University, Xi’an, China, in 1998. Since 1999, he has taken a combined Master-Doctor program and is working toward the Ph.D. degree in the National Key Laboratory of Antenna and Microwave Technology, Xidian University. His research interests include computational electromagnetics, slot antenna array, hybrid algorithms and electromagnetic compatibility.

Q

Chang-Hong Liang (M’80–SM’83) was born in Shanghai, China, in December 1943. He graduated from Xidian University (Formerly Northwest Telecommunications Institute), Xi’an, China, in 1965, and continued his graduate studies until 1967. From 1980 to 1982, he worked at Syracuse University, New York, as a Visiting Scholar. Since 1986, he has been a Professor and Ph.D. student advisor in the School of Electronic Engineering, Xidian University, where he is also a Director of the Academic Committee of National Key Lab of Antenna and Microwave Technology. He has published numerous papers and proceeding articles, is the author of ﬁve books. He is an Editor-in-Chief of the Journal of Xidian University. He has wide research interests, which include computational microwave and computational electromagnetics, microwave network theory, microwave measurement method and data processing, lossy variational electromagnetics, electromagnetic inverse scattering, electromagnetic compatibility. Prof. Liang is a Fellow of the Chinese Institute of Electronics (CIE), and has received the titles of “National Distinguished Contribution,” “National Excellent Teacher,” etc.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2073

**MIMO Wireless Communication Channel Phenomenology
**

Daniel W. Bliss, Member, IEEE, Amanda M. Chan, and Nicholas B. Chang

Abstract—Wireless communication using multiple-input multiple-output (MIMO) systems enables increased spectral efﬁciency and link reliability for a given total transmit power. Increased capacity is achieved by introducing additional spatial channels which are exploited using space-time coding. The spatial diversity improves the link reliability by reducing the adverse effects of link fading and shadowing. The choice of coding and the resulting performance improvement are dependent upon the channel phenomenology. In this paper, experimental channel-probing estimates are reported for outdoor environments near the personal communication services frequency allocation (1790 MHz). A simple channel parameterization is introduced. Channel distance metrics are introduced. Because the bandwidth of the channel-probing signal (1.3 MHz) is sufﬁcient to resolve some delays in outdoor environments, frequency-selective fading is also investigated. Channel complexity and channel stationarity are investigated. Complexity is associated with channel-matrix singular value distributions. Stationarity is associated with the stability of channel singular value and singular vector structure over time. Index Terms—Channel coding, information theory, multipath channels, multiple-input multiple-output (MIMO) systems.

I. INTRODUCTION

M

ULTIPLE-INPUT multiple-output (MIMO) systems are a natural extension of developments in antenna array communication. While the advantages of multiple receive antennas, such as gain and spatial diversity, have been known and exploited for some time [1]–[3], the use of transmit diversity has been investigated more recently [4], [5]. Finally, the advantages of MIMO communication, exploiting the physical channel between many transmit and receive antennas, are currently receiving signiﬁcant attention [6]–[8]. Because MIMO communication capacity is dependent upon channel phenomenology, studying and parameterizing this phenomenology is of signiﬁcant interest [9]–[19]. This paper makes a number of contributions to this area of study. First, while most experimental results have focused on indoor phenomenology, the phenomenology investigated here focuses on outdoor environments. Second, results for both stationary and vehicle-mounted moving transmitters are

Manuscript received March 19, 2003; revised September 27, 2003. This work was supported by the U.S. Air Force under Air Force Contract F19628-00-C0002. D. W. Bliss and A. M. Chan are with Advanced Sensor Techniques Group, MIT Lincoln Laboratory, Lexington, MA 02420-9185 USA (e-mail: bliss@ll.mit.edu, achan@ll.mit.edu. N. Chang is with the Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI 48109-2122 USA (e-mail: changn@eecs.umich.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832363

reported. Third, experimental phenomenological results are reported for both 4 4 and relatively large 8 8 MIMO systems, including channel stationarity, both in time and frequency. Fourth, two metrics of channel variation are introduced. One metric provides a measure of capacity loss assuming that receiver beamformers are constructed using incorrect channel estimates, which is useful to determine performance losses due to channel nonstationarity (either in time or frequency). The other metric is sensitive to the shape of the channel eigenvalue distribution, which is appropriate for space-time coding optimization, assuming a uniformed transmitter (UT) (that is transmitters without channel state information). Finally, a simple channel parameterization is provided which empirically matches channel eigenvalue distributions well and provides a simple approach to generate representative simulated channels for space-time coding optimization. MIMO systems provide a number of advantages over singleantenna communication. Sensitivity to fading is reduced by the spatial diversity provided by multiple spatial paths. Under certain environmental conditions, the power requirements associated with high spectral-efﬁciency communication can be signiﬁcantly reduced by avoiding the compressive region of the information theoretic capacity bound. This is done by distributing energy amongst multipath modes in the environment. Spectral efﬁciency is deﬁned as the total number of bits per second per Hz transmitted from one array to the other. Because MIMO systems use antenna arrays, interference can be mitigated naturally. In this paper, outdoor MIMO channel phenomenology near the PCS frequency allocation, 1.79 GHz, is discussed. The channel-probing signal has a bandwidth of 1.3 MHz. This bandwidth is sufﬁcient to resolve some delays, inducing frequency-selective fading in outdoor environments. In Sections II and III, information theoretic capacity of MIMO communication systems and channel estimation are reviewed. Channel difference metrics are introduced in Section IV. Performance of MIMO communication systems and optimal selection of space-time coding are dependent upon the complexity of the channel [20], [21]. This phenomenology for outdoor environments is investigated using MIMO channel-probing experiments. The results are interpreted using a simple parameterization introduced in Section V. The channel phenomenology experiments are described in Section VI, and the experimental results, reporting estimates of channel complexity and stationarity, are discussed in Section VII. II. CHANNEL CAPACITY The information theoretic capacity of MIMO systems has been discussed widely [6]–[8]. It is assumed for the sake of the

0018-926X/04$20.00 © 2004 IEEE

2074

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

following discussion that the receiver can accurately estimate a pseudostationary channel. Given this assumption, there are two types of spectral-efﬁciency bounds: informed transmitter (IT) and UT, depending on whether or not channel estimates are fed back to the transmitter. For narrowband MIMO systems, the coupling between the transmitter and receiver can be modeled using (1) is the (number of receive by transmit where antenna) channel matrix, containing the complex attenuation is an between each transmit and receive antenna, matrix containing the samples of the transmit array vector, is an matrix containing the samples of the is an matrix complex receive-array output, and containing zero-mean complex Gaussian noise. It is often useful to investigate the structure of the channel matrix and the mean-square attenuation independently. This can be achieved by studying the root-mean-square normalized channel matrix (2) (3)

If (6) is not satisﬁed for some . smaller B. UT

, it will not be satisﬁed for any

If the channel is not known at the transmitter, then the optimal transmission strategy is to transmit equal power with each antenna, , [7]. Assuming that the receiver can accurately estimate the channel, but the transmitter does not attempt to optimize its output to compensate for the channel, the maximum spectral efﬁciency is given by (7) This is a common transmit constraint as it may be difﬁcult to provide the transmitter channel estimates. Similarly to the IT case, the UT spectral-efﬁciency bound is purely a function of the channel-matrix singular values. Expressing the channel matrix with a singular vector decomposi, the capacity is a function of eigenvalues, but tion, not of the eigenvectors, of

(8) where is the mean-square transmitter-to-receiver attenuation, is the normalized channel matrix, and indicates the Frobenius norm. A. IT There are a variety of possible transmitter constraints. Here it is assumed that the fundamental limitation is the total power transmitted. The optimization of the noise-noris constrained by the malized transmit covariance matrix total noise-normalized transmit power . Allowing different transmit powers at each antenna, this constraint can be enforced . The results of the channel-specusing the form tral-efﬁciency bounds discussions presented in [8] are repeated here. The capacity can be achieved if the channel is known by both the transmitter and receiver, giving (4) where the notation indicates determinant, indicates Herindicates an identity matrix of size mitian conjugate, and . Solving for the optimal , the resulting capacity is given by (5) where of is an diagonal matrix with entries , whose values are the top eigenvalues . The values must satisfy (6) .. (10) where the singular-value entries of the diagonal matrix given by . C. Frequency-Selective Channels In environments where there is frequency-selective fading, the channel matrix is a function of frequency . As has been discussed in [22], the resulting capacity is a function of this fading structure. Exploiting the fact that frequency channels are orthogonal, the capacity in frequency-selective fading can be calculated using an extension of (5) and (7). For the UT, this leads to the frequency-selective spectral-efﬁciency bound are

(9) where the distance between frequency samples is given by , and -bin frequency-partitioned channel matrix is given by

.

The approximation is exact if the supported delay range of the channel is sampled sufﬁciently.

BLISS et al.: MIMO WIRELESS COMMUNICATION CHANNEL PHENOMENOLOGY

2075

For the IT channel capacity, power is optimally distributed amongst both spatial modes and frequency channels. The capacity can be expressed (11)

sampling, the explicit frequency-selective form can be constructed using a discrete Fourier transform (16) or equivalently

which is maximized by (5) with the appropriate substitutions for the frequency-selective channel, and diagonal entries in in (6) are selected from the eigenvalues of . Because of the block diagonal structure of , the space-frequency noise-normalized transmit covariance matrix is a block diagonal matrix, normalized so that . III. ESTIMATION The Gaussian probability density function for a multivariate, signal-in-the-mean, statistical model of the received signal is given by (12) where is the noise covariance matrix. The maximum-likelihood estimate of is given by (13) assuming that the reference signals in are known and is nonsingular. As one might intuit from the structure of (13), if a signal’s sole use is channel estimation, then the choice of , such that is proportional to the identity matrix (that is equal-power orthogonal signals) is optimal for channel probing in ﬁnite signal-to-noise ratio (SNR) environments. However, if joint channel and signal detection is used, then orthogonal signals are not necessarily optimal for link performance. The previous channel-estimation discussion explicitly assumed ﬂat fading. However, the frequency-selective channels can be estimated by ﬁrst estimating a ﬁnite impulse-response MIMO channel which can be transformed to the frequency domain. A ﬁnite impulse-response extension of (1) is given by introducing delayed copies of at delays

(17) where the -point discrete Fourier transform is represented by and the Kronecker product is represented by . IV. CHANNEL DIFFERENCE METRICS A variety of metrics are possible. Here, two metrics are discussed. Both metrics are ad hoc, but are motivated by limiting forms of the information theoretic capacity. The ﬁrst metric, discussed in Section IV-A, is sensitive to the differences in channel eigenvalue distributions. While there are an unlimited number of channel eigenvalue distributions that can provide a particular capacity, for a given mean channel attenuation and power, performance of space-time codes is sensitive to the shape of the distribution. Because the optimization of UT space-time codes depends upon the eigenvalue distribution but not the eigenvector structure, the metric introduced in Section IV-A is an appropriate metric for investigating this issue. Speciﬁcally, space-time codes must select a rate versus redundancy operating point [20], [21]. The optimal operating point is a function of the shape of the channel eigenvalue distribution. This metric is used to estimate the channel parameter introduced in Section V-C. The second metric, discussed in Section IV-B, is sensitive to differences in both the singular-value distribution and the channel eigenvector structure. In general, MIMO receivers employ some sort of beamformer to coherently combine the signals impinging upon each receive antenna. In dynamic environments (either in time or frequency) channel estimates can quickly become inaccurate. A measure of the adverse effects of using these “stale” estimates is provided by this metric. Effectively, this metric provides a measure of the fractional capacity loss in the low SNR (or equivalently low spectral-efﬁciency) limit. Because performance in the low SNR limit is not affected by interference introduced by the other transmit antennas, MIMO systems operating at higher SNR will experience greater interference and thus worse performance. Consequently, this metric is an optimistic estimate of the expected performance due to dynamic channels. A. Eigenvalue-Based Metric

. . . so that the transmit matrix has dimension resulting wideband channel matrix has the dimension

(14)

. The

(15) Using this form, an effective channel ﬁlter is associated with each transmit-to-receive antenna link. Assuming regular delay

As was mentioned in Section II, MIMO capacity is only a function of the channel singular values. Equivalently, capacity is invariant under channel-matrix transformations of the form (18)

2076

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

where and are arbitrary unitary matrices. Consequently, for some applications it is useful to employ a metric which is also invariant under this transformation. Because capacity is a function of the structure of the channel singular-value distribution, the metric should be sensitive to this structure. . A natural metric The channel capacity is a function of would employ the distance between the capacity for two channel matrices at the same average total received power, that is, the same

is the column of the channel matrix associated with where transmitter . In the low SNR limit, the optimal receive beam. If some former is given by the matched response given in other beamformer is employed, , then signal energy is lost, adversely affecting the capacity (23) One possible reason that a beamformer might use the wrong matched spatial ﬁlter is channel nonstationarity. Assuming the SNR is sufﬁciently low, the fractional capacity loss is given by

(19) However, there are two problems with this deﬁnition. First, the difference is a function of . Second, there is degeneracy in singular values that gives a particular capacity. To address the ﬁrst issue the difference can be investigated in a high SNR limit, resulting in (24) which is the power-weighted mean estimate, where is deﬁned to be the inner product between the “good” and “bad” unit-norm array responses for the th transmitter. It is generally desirable for metrics to be symmetric with respect and , thus avoiding moral attributions with regard to to channel matrices. Using the previous discussion as motivation, a symmetric form of fractional capacity loss is given by (25) where the “power-weighted” average is evaluated over transmitters. The metric presented in (25) provides an estimate of the loss in capacity if the incorrect channel is assumed in a low SNR environment. In general, the loss of capacity is much more signiﬁcant if operating in a high spectral efﬁciency, and therefore high SNR regime. If only spatial mitigation is employed (as opposed to a combination of spatial processing and multiuser detection [23], [24]), a slight channel mismatch will introduce signiﬁcant interference, and thus strongly adversely affect demodulation performance. V. CHANNEL PHENOMENOLOGY A. Singular Values The singular-value distribution of , or the related eigen, is a useful tool for understanding value distribution of the expected performance of MIMO communication systems. From the discussion in Section II, it can be seen that the channel capacity is a function of channel singular values, but not the singular-vector structure of the channel. Thus, channel phenomenology can be investigated by studying the statistics of channel singular-value distributions.

(20) indicates the th largest eigenvalue of . To inwhere crease the sensitivity to the shape of the eigenvalue distribution, the metric is deﬁned to be the Euclidean difference, assuming that each eigenvalue is associated with an orthogonal dimension, resulting in

(21)

B. Fractional Receiver-Loss Metric metric is introIn this section a power-weighted mean duced. The metric takes into account both the eigenvalue and eigenvector structure of the channels. It is motivated by the effect of receive beamformer mismatch on capacity. Starting with (7), the low SNR UT capacity approximation is given by

(22)

BLISS et al.: MIMO WIRELESS COMMUNICATION CHANNEL PHENOMENOLOGY

2077

B. Channel Parameterization A commonly employed model assumes the channel is proportional to a matrix, , where the entries are independently drawn from a unit-norm complex circular Gaussian distribution. While the distribution is convenient, it does suffer from a singular-value distribution that is overly optimistic for many environments. One solution is to introduce spatial correlations [10]–[12]. While using the transformation this approach is limited [8], it produces simply more realistic channels than the uncorrelated Gaussian model. The spatial correlation matrices can be factored so that and , where and are unitary and are positive semideﬁnite diagonal matrices, and matrices. When the arrays are located in environments that are signiﬁcantly different, then correlations seen by one array will typically be much stronger than the other, and the effect of either the left or right will dominate the shaping of the channel matrix singular-value distribution. Conversely, if the environments . In pracare similar then one would expect that tice, similar channel matrix singular-value distributions can be achieved given either assumption. However, the required values and , are, of course, different. For the experiments disof cussed in this paper, both arrays are in similar environments and a symmetric form seems a reasonable model. Assuming that the number of transmit and receive antennas are equal and have similar spatial correlation characteristics, the diagonal matrices can , producing the new random be set equal, channel matrix

Fig. 1. Ratio of bounds on mean UT capacity of = 0:2, 0.4, 0.6 to = 1.

the ratio of capacity bounds for 0.2, 0.4, and 0.6 for a 4 4 MIMO system is displayed. In practice, the ratio of bounds tends to produce slightly optimistic capacity results at values of . The essential features are accurate. Assuming that the space-time coding takes the channel statistics into account for values of 0.6 or greater, performance loss is not overwhelming. A second interesting feature is that at very high SNR the ratio of capacity slowly approaches 1. This is because at very high SNR even strongly attenuated channel modes become useful. Modeling approaches that introduce reduced “effective” numbers of antennas do not reproduce this phenomenon well. C. Channel Parameter Estimation An estimate for associated with particular transmit and receive locations is given by minimizing the mean-square metric given in (21) (29) where indicates the estimated value of . Here the expecta, indicates averaging is over an ensemble tion, denoted by of for a given and an ensemble of for given transmit and receiver sites. It is worth noting that this approach does not necessarily provide an unbiased estimate of . Estimates of , using the metric introduced in here, are dependent upon the received SNR. To reduce the bias, one can add complex Gaussian noise to to produce , mimicking the integrated SNR of the estimate of . Data presented here has sufﬁciently high SNR such that can . be estimated within VI. EXPERIMENT The experimental system employed is a slightly modiﬁed version of the Massachusetts Institute of Technology (MIT) Lincoln Laboratory system used previously [3], [23], [25]–[27]. The transmit array consists of up to eight arbitrary waveform transmitters. The transmitters can support up to a 2 MHz bandwidth. These transmitters can be used independently, as two groups of four coherent transmitters or as a single coherent group of eight transmitters. The transmit systems can be deployed in the laboratory or in vehicles. When operating coherently as a multiantenna transmit system, the individual

(26) (27) where is used to set overall scale, is given by the size of , and and indicate random unitary matrices. The form given here is somewhat arbitrary, but has the satisfying of a rank-one channel matrix is procharacteristics that as a spatially uncorrelated Gaussian matrix duced, and as is produced; thus, the parameterization can easily approximate, in a statistical sense, nearly all environments. This stochastic channel parameterization has the advantage that it is not dependent upon the particular causes of the correlation, or details of is chosen the arrays or environment. The normalization for is . so that the expected value of The model can be related to the ergodic or mean capacity [8] (averaged over an ensemble channel). Exploiting the fact that MIMO capacity is convex cap, a bound on the mean capacity is given by

(28) This bound is not necessarily tight, but is useful for illustrating the effects of channel parameter value on capacity. In Fig. 1,

2078

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

TABLE I LIST OF TRANSMIT SITES

Fig. 2. Scatter plot of mean-squared SISO link attenuation, a , versus link range for the outdoor environment near the PCS frequency allocation. The error bars indicate a range of 1 standard deviation of the estimates at a given site.

6

transmitters can send independent sequences using a common local oscillator. Synchronization between transmitters and receiver and transmitter geolocation is provided by GPS receivers in the transmitters and receivers. The MIT Lincoln Laboratory array receiver system is a highperformance 16-channel receiver system that can operate over a range of 20 MHz to 2 GHz, supporting a bandwidth of up to 8 MHz. The receiver can be deployed in the laboratory or in a stationary “bread truck.” A. MIT Campus Experiment The experiments were performed during July and August 2002 on and near the MIT campus. These outdoor experiments were performed in a frequency allocation near the PCS band (1.79 GHz). The transmitters periodically emitted 1.7 s bursts containing a combination of channel-probing and space-time-coding waveforms. A variety of coding and interference regimes were explored for both moving and stationary transmitters. The space-time-coding results are beyond the scope of this paper and are discussed elsewhere [23], [24]. Channel-probing sequences using both four and eight transmitters were used. The receive antenna array was placed on top of a tall one-story building (Brookline St. and Henry St.) surrounded by two- and three-story buildings with a parking lot along one side. Different four or eight antenna subsets of the 16-channel receiver were used to improve statistical signiﬁcance. The nearly linear receive array had a total aperture of less than 8 m, arranged as three subapertures of less than 1.5 m each. The transmit arrays were located on the top of vehicles within 2 km of the receive array. On each vehicle four antennas were approximately located at the vertices of a square, with separation of greater than two to three wavelengths. When operating as an eight-element transmitter, two adjacent parked vehicles were used, connected by a cable that distributed a local oscillator signal. The channel-probing sequence supports a bandwidth of 1.3 MHz with a length of 1.7 ms repeated ten times. All four or eight transmitters emitted nearly orthogonal signals simultaneously.

Fig. 3. CDF of channel a estimates, normalized by the mean a for each site, for SISO, 4 4 and 8 8 MIMO systems.

2

2

VII. EXPERIMENTAL RESULTS Channel-complexity and channel-stationarity performance results are presented in this section. A list of transmit sites used for these results is presented in Table I. The table includes distance between transmitter and receiver, velocity of transmitter, the number of transmit antennas, and the estimated for the transmit site. Uncertainty in is determined using the bootstrap technique [28]. Cumulative distribution functions (CDF) reported here are evaluated over appropriate entries from Table I. The systematic uncertainty in the estimation of caused by estimation bias, given the model, is less than 0.02. A. Attenuation The peak-normalized mean-squared single-input singleoutput (SISO) attenuation (path gain) averaged over transmit and receive antenna pairs for a given transmit site is displayed in Fig. 2 for the outdoor environment. The uncertainty in the estimate is evaluated using a bootstrap technique. B. Channel Complexity Channel complexity is presented using three different apestimates, eigenvalue CDFs, and proaches. Variation in estimate CDFs are presented. estimates norIn Fig. 3, CDFs of for each transmit site are displayed. CDFs malized by mean are displayed for narrowband SISO, 4 4, and 8 8 MIMO systems. As one would expect, because of the spatial diversity, the variation in mean antenna-pair received power decreases dramatically as the number of antenna pairs increases. This

BLISS et al.: MIMO WIRELESS COMMUNICATION CHANNEL PHENOMENOLOGY

2079

Fig. 4. CDF of narrowband channel eigenvalue distributions for 4 4 MIMO systems: (a) simulated Gaussian channel and (b) experimental results.

2

Fig. 5. CDF of narrowband channel eigenvalue distributions for 8 8 MIMO systems: (a) simulated Gaussian channel and (b) experimental results.

2

demonstrates one of the most important statistical effects that MIMO links exploit to improve communication link robustness. For example, if one wanted to operate with a probability of 0.9 to close the link, one would have to operate the SISO link with ) margin of over 15 dB. The MIMO an excess SISO SNR ( systems received the added beneﬁt of array gain, which is not accounted for in the ﬁgure. In Figs. 4 and 5, CDFs of eigenvalues are presented for 4 4 and 8 8 mean-squared-channel-matrix-element-normalized narrowband channel matrices, . Both simulated Gaussian channels and experimental results are displayed. Superﬁcially, the distributions of the simulated and experimental distributions are similar. However, closer inspection reveals that the experimental distributions cover a greater range of eigenvalues. This is the result of the steeper channel-eigenvalues distribution that is observed in the experimental data compared to the simulated Gaussian channel. The experimental CDFs are evaluated over all site lists. Some care must be taken in interpreting these ﬁgures because eigenvalues are not independent. Nonetheless, the steepness of the CDFs is remarkable. One might interpret this to indicate that optimized space-time codes should operate with a relatively high probability of success. The CDFs for estimates are presented in Fig. 6. The mean values of for each environment are

Fig. 6. CDF of estimates for 4

2 4 and 8 2 8 MIMO systems.

While one might expect smaller variation in the 8 8 systems because of the much larger number of paths, this effect may have been exaggerated in Fig. 6 because of the limited number of 8 8 sites available in the experiment.

The values of the channel-complexity parameter, , are, of course, dependent upon the details of the environment and the geometry of the transmit and receive arrays. As can be seen in Table I, the values of vary from one transmit site to another transmit site. Furthermore, one would anticipate in unlike environments, signiﬁcantly different values of such as the open plains of the Midwest or in highly elevated towers. The dependency upon array geometries is somewhat less clear. Because the arrays employed in this experiment are spatially undersampled, the received signal experiences signiﬁcant spatial aliasing. Increasing the array aperture may help resolve closely spaced scatterers; this occurs at the expense of folding other widely spaced scatterers back on similar array responses. Consequently, while perturbations in array geometries certainly affect particular received signals, these perturbations are not expected to affect strongly the statistical properties of the channel; thus values of are not expected to be a strong function of array geometry.

2080

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 8. Example time variation of power-weighted mean cos , (t ); (t) , for stationary and moving 4 4 MIMO systems.

fH

H g

2

Fig. 7. Eigenvalues, , of as a function of time for (a) stationary and (b) moving transmitters. The same overall attenuation, estimated at t = 0, is used for all time samples.

HH

This point can be demonstrated by constraining the choice of receive antennas used for calculating . By excluding or requiring antennas to be consecutive and calculating under these constraints, sensitivity to antenna separation can be investigated. In the following table, three sets of constraints are implemented. In the ﬁrst column, all receive antennas are used. In the second column, employed antennas are separated by at least two unused antennas. In the third column, only consecutive antennas are used.

Fig. 9. CDF of time variation of power-weighted mean cos , (t ); (t) , for stationary 4 4 MIMO system. Contours of CDF probabilities of 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9 are displayed. Because there is little variation, all curves are compressed near of 1.

fH

H g

2

There is a slight bias for greater antenna separation to produce larger values of , which is consistent with the expectation that greater antenna separation produces more random channel matrices. However, this trend is very subtle, and in all cases, the results are statistically consistent with being independent of antenna separation at these relatively large antenna separations. C. Channel Stationarity for stationary The temporal variation of eigenvalues of and moving transmitters is displayed in Fig. 7. In this ﬁgure the normalization is ﬁxed, allowing for overall shifts in attenuation. As one would expect, the eigenvalues of the moving transmitter vary signiﬁcantly more than those of the stationary environment. However, the eigenvalues of the stationary transmitter do vary somewhat. While the transmitters and receivers are physically stationary, the environment does move. This effect is particularly noticeable near busy roads. Furthermore, while the multiple antennas are driven using the same local oscillator, given

the commercial grade transmitters, there are always some small relative-frequency offsets. The example variation is given for transmit sites #7 and #14. While the moving-transmitter eigenvalues ﬂuctuate more than those of the stationary transmitter, the values are remarkably stable in time. Conversely, an example of the time metric [from variation of the power-weighted mean (25)], displayed in Fig. 8, varies signiﬁcantly for the moving transmitter within 10 ms. This indicates that the eigenvector structure varies signiﬁcantly, while the distribution of eigenvalues tends to be more stable. In the example, the stationary transmitter is located at site #7, and the moving transmitter is located at #14. Over the same period, the stationary transmitter is relatively stable. CDFs for stationary and moving transmitters are displayed in Figs. 9 and 10. In the ﬁgures, 4 4 MIMO experiment sites with a speed less than or equal to 0.2 m/s were considered to be “stationary” (sites: 7, 9, 16, and 18), and those with speeds greater than 5 m/s were considered to be “moving” (sites: 10, 12–15, 17). As was discussed in Section IV-B, the performance implications of a particular value of depend upon the operating SNR and the receiver design. At low SNR, the fractional UT capacity loss due to receiver mismatch is given directly by the value of . At high SNR, if interference mitigation is primarily achieved through spatial antenna processing, then the performance loss can be signiﬁcantly worse. This is because contamination from interfering transmit antennas is allowed to overwhelm the intended signals at the outputs of inaccurate beamformers. Furthermore, the signiﬁcant variation of the moving transmitter is an indication that implementing an IT MIMO system would be

BLISS et al.: MIMO WIRELESS COMMUNICATION CHANNEL PHENOMENOLOGY

2081

delay spread, and the resulting frequency-selective fading, is both a function of environment and link length. Consequently, some care must be taken in interpreting this result. VIII. SUMMARY In this paper, outdoor MIMO channel phenomenology was discussed. Data from an experiment performed on and near the MIT campus was used to study the phenomenology. The phenomenology was investigated from the perspective of the singular-value distributions of the channel matrices. A channel parameterization approach was introduced. Two channel-difference metrics were introduced. The ﬁrst was used to estimate the channel parameter. The second metric was employed to demonstrate signiﬁcant channel variation both as a function of time and frequency. ACKNOWLEDGMENT Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the United States Government. The authors would like to thank the excellent MIT Lincoln Laboratory staff involved in the MIMO experiment, in particular S. Tobin, J. Nowak, L. Duter, J. Mann, B. Downing, P. Priestner, B. Devine, T. Tavilla, A. McKellips and G. Hatke. The authors would also like to thank the MIT New Technology Initiative Committee for their support. The authors would also like to thank K. Forsythe, A. Yegulalp, and D. Ryan of MIT Lincoln Laboratory and V. Tarokh of Harvard University for their thoughtful comments. The authors would like to thank N. Sunkavally of MIT for his contributions to the experiment and the analysis. REFERENCES

[1] W. C. Jakes, Microwave Mobile Communications. New York: Wiley, 1974. [2] R. A. Monzingo and T. W. Miller, Introduction to Adaptive Arrays. New York: Wiley, 1980. [3] K. W. Forsythe, D. W. Bliss, and C. M. Keller, “Multichannel adaptive beamforming and interference mitigation in multiuser CDMA systems,” in Proc. 33rd Asilomar Conf. Signals, Systems & Computers, vol. 1, Paciﬁc Grove, CA, Oct. 1999, pp. 506–510. [4] A. Wittneben, “Base station modulation diversity for digital SIMULCAST,” in Proc. IEEE Vehicular Technology Conf., 1991, pp. 848–853. [5] V. Weerackody, “Diversity for direct-sequence spread spectrum using multiple transmit antennas,” in Proc. IEEE ICC, vol. 3, Geneva, 1993, pp. 1775–1779. [6] G. J. Foschini, “Layered space-time architecture for wireless communication in a fading environment when using multi-element antennas,” Bell Labs Tech. J., vol. 1, no. 2, pp. 41–59, 1996. [7] I. E. Telatar, “Capacity of multi-antenna Gaussian channels,” Eur. Trans. Telecommun., vol. 10, no. 6, pp. 585–595, Nov.-Dec. 1999. [8] D. W. Bliss, K. W. Forsythe, A. O. Hero, and A. F. Yegulalp, “Environmental issues for MIMO capacity,” IEEE Trans. Signal Processing, vol. 50, pp. 2128–2142, Sept. 2002. [9] P. Marinier, G. Y. Delisle, and C. L. Despins, “Temporal variations of the indoor wireless millimeter-wave channel,” IEEE Trans. Antennas Propagat., pp. 928–934, June 1998. [10] D. Gesbert, H. Bolcskei, D. A. Gore, and A. J. Paulraj, “Performance evaluation for scattering MIMO channel models,” in Proc. 34th Asilomar Conf. Signals, Systems & Computers, vol. 1, Paciﬁc Grove, CA, Oct. 2000, pp. 748–752. [11] D.-S. Shui, G. J. Forschini, M. J. Gans, and J. M. Kahn, “Fading correlation and its effect on the capacity of multielement antenna systems,” IEEE Trans. Commun., vol. 48, pp. 502–513, Mar. 2000.

CDF of time variation of power-weighted mean cos , for moving 4 4 MIMO system. Contours of CDF probabilities of 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9 are displayed. Fig. 10.

f

H(t ); H(t)g,

2

Fig. 11. Example of frequency-selective variation of the power-weighted (f ); (f ) . mean cos ,

fH

H g

Fig. 12. CDF of frequency-selective variation of the power-weighted mean (f ); (f ) . Contours of CDF probabilities of 0.1, 0.2, 0.3, 0.4, cos , 0.5, 0.6, 0.7, 0.8, and 0.9 are displayed.

fH

H g

very challenging for the moving transmitter, but might be viable for some stationary MIMO systems. D. Frequency-Selective Fading An example of the frequency variation of the power-weighted is displayed in Fig. 11. The variation is indicated mean using the metric presented in (25). In the example, the stationary transmitter is located at site #7. Relatively small frequency . The offsets induce signiﬁcant changes in CDF of the frequency-selective channel variation is displayed in Fig. 12 (using sites: 7, 9, 16, and 18). This sensitivity indicates that there is signiﬁcant resolved delay spread and that to safely operate using the narrowband assumption, bandwidths less than 100 kHz should be employed. It is worth noting that

2082

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[12] D. Chizhik, F. Rashid-Forrokhi, J. Ling, and A. Lozano, “Effect of antenna separation on the capacity of blast in correlated channels,” IEEE Commun. Lett., vol. 4, pp. 337–339, Nov. 2000. [13] D. Hampicke, C. Schneider, M. Landmann, A. Richter, G. Sommerkorn, and R. S. Thoma, “Measurement-based simulation of mobile radio channels with multiple antennas using a directional parametric data model,” in Proc. IEEE Vehicular Technology Conf., vol. 2, 2001, pp. 1073–1077. [14] M. Stoytchev and H. Safar, “Statistics of the MIMO radio channel in indoor environments,” in Proc. IEEE Vehicular Technology Conf., vol. 3, 2001, pp. 1804–1808. [15] D. P. McNamara, M. A. Beach, and P. N. Fletcher, “Experimental investigation of the temporal variation of MIMO channels,” in Proc. IEEE Vehicular Technology Conf., vol. 2, 2001, pp. 1063–1067. [16] C. C. Martin, N. R. Sollenberger, and J. H. Winters, “MIMO radio channel measurements: Performance comparison of antenna conﬁgurations,” in Proc. IEEE Vehicular Technology Conf., vol. 2, 2001, pp. 1225–1229. [17] J. F. Kepler, T. P. Krauss, and S. Mukthavaram, “Delay spread measurements on a wideband MIMO channel at 3.7 GHz,” in Proc. IEEE Vehicular Technology Conf., vol. 4, 2002, pp. 2498–2502. [18] T. Svantesson, “A double-bounce channel model for multi-polarized MIMO systems,” in Proc. IEEE Vehicular Technology Conf., vol. 2, 2002, pp. 691–695. [19] J. P. Kermoal, L. Schumacher, K. I. Pedersen, P. E. Mogensen, and F. Frederiksen, “A stochastic MIMO radio channel model with experimental validation,” IEEE J. Select. Areas Commun., vol. 20, pp. 1211–1226, August 2002. [20] L. Zheng and D. Tse, “Optimal diversity-multiplexing tradeoff in multiple antenna fading channels,” in Proc. 35th Asilomar Conf. Signals, Systems & Computers, Paciﬁc Grove, CA, Nov. 2001. [21] H. E. Gamal, “On the robustness of space-time coding,” IEEE Trans. Signal Processing, vol. 50, pp. 2417–2428, Oct. 2002. [22] G. G. Raleigh and J. M. Ciofﬁ, “Spatio-temporal coding for wireless communications,” IEEE Trans. Commun., vol. 46, pp. 357–366, Mar. 1998. [23] D. W. Bliss, P. H. Wu, and A. M. Chan, “Multichannel multiuser detection of space-time turbo codes: Experimental performance results,” in Proc. 36th Asilomar Conf. Signals, Systems & Computers, Paciﬁc Grove, CA, Nov. 2002. [24] D. W. Bliss, “Robust MIMO wireless communication in the presence of interference using ad hoc antenna arrays,” presented at the Proc. Military Communications Conf., MILCOM 2003, Boston, MA, Oct. 2003. , “Angle of arrival estimation in the presence of multiple access [25] interference for CDMA cellular phone systems,” in Proc. 2000 IEEE Sensor Array and Multichannel Signal Processing Workshop, Cambridge, MA, Ma. 2000. [26] C. M. Keller and D. W. Bliss, “Cellular and PCS propagation measurements and statistical models for urban multipath on an antenna array,” in Proc. 2000 IEEE Sensor Array and Multichannel Signal Processing Workshop, Cambridge, MA, Mar. 2000, pp. 32–36. [27] D. W. Bliss, “Robust MIMO wireless communication in the presence of interference using ad hoc antenna arrays,” presented at the MILCOM, Boston, MA, Oct. 2003. [28] B. Efron, The Jackknife, the Bootstrap and Other Resampling Plans. Bristol, U.K.: Society for Industrial and Applied Mathematics, 1982.

Daniel W. Bliss (M’97) received the B.S.E.E. degree in electrical engineering from Arizona State University, Tuscon, in 1989 and the M.S. and Ph.D. degrees in physics from the University of California at San Diego, in 1995 and 1997, respectively. Employed by General Dynamics from 1989 to 1991, he designed avionics for the Atlas-Centaur launch vehicle and performed research and development of fault-tolerant avionics. As a member of the Superconducting Magnet Group at General Dynamics from 1991 to 1993, he performed magnetic ﬁeld calculations and optimization for high-energy particle accelerator superconducting magnets. His doctoral work rom 1993 to 1997, was in the area of high-energy particle physics, searching for bound states of gluons, studying the two-photon production of hadronic ﬁnal states, and investigating innovative techniques for lattice gauge theory calculations. Since 1997, he has been employed by MIT Lincoln Laboratory, where he is currently a Staff Member at in the Advanced Sensor Techniques Group, where he focuses on multiantenna adaptive signal processing, primarily for communication systems, and on parameter estimation bounds, primarily for geolocation. His current research topics include algorithm development for multichannel multiuser detectors (MCMUD) and information theoretic bounds and space-time coding for MIMO communication systems.

Amanda M. Chan received the B.S.E.E. and M.S.E.E. degrees in electrical engineering from the University of Michigan, Ann Arbor, in 2000 and 2002, respectively. Currently, she is an Associate Staff Member in the Advanced Sensor Techniques Group, MIT Lincoln Laboratory, Lexington, MA. Her interests are in channel phenomenology. She has previously worked with implementation of synthetic aperture geolocation of cellular phones. Most recently, she has worked on the implementation of MIMO channel parameterization.

Nicholas B. Chang received the B.S.E. degree in electrical engineering (magna cum laude) from Princeton University, Princeton, NJ and the M.S.E. degree in electrical engineering from the University of Michigan, Ann Arbor, in 2002 and 2004, respectively. He worked for MIT Lincoln Laboratory, Lexington, MA, in 2001 and 2002, focusing on synthetic aperture geolocation of wireless systems and channel phenomenology of MIMO communications systems. He is currently a Graduate Student in the Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor. Mr. Chang is a Member of Tau Beta Pi and Sigma Xi, the Scientiﬁc Research Society.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2083

**Service Oriented Statistics of Interruption Time Due to Rainfall in Earth-Space Communication Systems
**

Emilio Matricciani

Abstract—The paper reports and discusses simulated statistics obtained by the synthetic storm technique, in the Po Valley, Northern Italy, on the interruptions (outages), due to rainfall, observed in contiguous (clock) periods of the day of duration , with 1 min 24 h. The results refer to a 32 slant path at 11.6 GHz, although the main conclusions are independent of carrier frequency and of site, and are based on a large experimental rain rate database (10.6 years of observation). The results can be used to assess the quality and unavailability of services of duration during a day in earth-space communication systems affected by rain attenuation. A distinction is made and discussed between channel unavailability and service unavailability. The numerical results and the best ﬁt and extrapolation formulas derived might provide a rough approximation to the same statistics at different elevation angles, clock intervals and carrier frequencies, for sites with the same climate of the Po Valley. Index Terms—Channel unavailability, diurnal cycles, fade duration, microwave propagation, rain attenuation, service unavailability.

I. INTRODUCTION HE design of satellite systems working in frequency bands affected by rain attenuation, (dB), are based, today, only on the long term probability distribution (i.e., fraction of time) that is exceeded in an average year (or in the worst month). In an age of a large variety of services offered, or to be offered, to users from satellites or from troposphere platforms, it would be useful to match system design to the time of the day and to the expected duration of the services offered, e.g., internet sessions, digital video and audio broadcasting. The application of forward error correction (FEC) codes, as currently done, may make rainfall attenuation a less severe problem for a satellite system in the 12-GHz band, if the system is designed and its low values (i.e., the by taking into account not range – ) and thus very short fade durations usually considered in telephony services, a concept we can call channel unavailability, but the number of interruptions of a maximum duration that a user can tolerate for a certain service, a TV or radio show, etc., a concept we can call service unavailability. At higher frequencies (e.g., 20–30 and 40–50 GHz frequency bands), coding gain may be, however, largely ineffective, both for rain attenuation and for the “quasistatic” extra attenuation due to other sources of fading, such as oxygen, water vapor and clouds. Rain attenuation, however, is likely to cause long random interruptions as the results below show.

Manuscript received March 3, 2003; revised October 27, 2003. The author is with the Dipartimento di Elettronica e Informazione, Politecnico di Milano, 20133 Milan, Italy (e-mail: matricci@elet.polimi.it). Digital Object Identiﬁer 10.1109/TAP.2004.832374

T

The necessity to distinguish between channel unavailability and service unavailability calls for statistics on the number (or probability) of outages of a given duration in speciﬁc times of the day, i.e., statistics on the interruptions of a service because of excessive rain attenuation. To author’s knowledge, however, this kind of statistics, or service grades, and a clear distinction between channel unavailability and service unavailability, have not yet been established. Once this is done, the kind of statistics and the scaling methods proposed in this paper could be used to design a satellite system based on service unavailability rather than on channel unavailability. The assessment of these statistics for a geographical area and a satellite system requires long term experiments and thus a very large economic and human effort. Few experimental data collected many years ago in relatively short observations and for long intervals of the day (e.g., four hours), are available in the open literature [1]–[5]. Reliable simulations, based on physical prediction models and a long observation, are then welcome. One such method is the “synthetic storm” technique [6] and the long observation is our database of rain-rate time series. For each rainstorm, physically described by a rain rate time series collected at a site with a rain gauge, the synthetic storm technique can generate a rain attenuation time series at any carrier frequency and polarization, for any slant path with elevation angle larger than about 10 . The synthetic storm technique was ’s successfully tested to predict conventional long term ’s [6], long term statistics of fade duration [7], long term relative to contiguous periods of the day of four hours [8]. If simulated rain attenuation time series are compared to simultaneous real measurements (e.g., as done for Spino d’Adda in a 37.8 slant path to Italsat, at 18.77 GHz [9]), the agreement is very good, especially when the rain storm motion is parallel to projection to ground of the slant path. When the motion is not parallel to projection to ground, the simulated time series are so realistic that they might as well be measured in a long observation. In fact, the synthetic storm technique yields results that are averaged over all the rain storms velocity ﬁeld. These indirect tests are of considerable importance for the present work because, in our opinion, they suggest that our simulations can provide results that remind experimental ones. An earth-space microwave radio link is also affected by fading due to clouds, water vapor and oxygen. The fading due to these phenomena can be large (e.g., see [10]), but more static than rain attenuation. They must be taken, of course, into account in a full design of the communication system. At microwaves, however, rainfall is still the major random physical cause of fading that can affect a channel for intervals of time

0018-926X/04$20.00 © 2004 IEEE

2084

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

comparable to the duration of many services, so that the service oriented statistics of interruptions reported below refer to rain attenuation only. We have applied the synthetic storm technique to simulate 11.6-GHz rain attenuation time series in a 32 slant path in Northern Italy by means of a large and reliable database of rain rate time series. As discussed above, we think that the results can be considered a good estimate of true measurements of rain attenuation in an earth-space channel in the 12 GHz band, or in higher microwave frequency bands, once the results are scaled. In [8], [11] we simulated, tested and discussed long term probability distributions concerning contiguous (clock) periods of equal duration (min) in a day, for several values of . In the present paper, we present detailed long term statistics on the number and probabilities of service interruption. The results are then useful to assess the quality and unavailability (i.e., interruptions or outages) of services of duration in a day. Next section sets the stage for our simulations, Section III reports the simulated results, Sections IV and V provide empirical formulas useful to generate predictions at different sites, frequencies, duration of contiguous periods and elevation angles, and Section VI draws some conclusions. II. RADIO LINK SIMULATIONS AND EXPERIMENTAL RAINFALL DATABASE Once the radio link geometric and radio electrical parameters have been speciﬁed and the synthetic storm technique has been applied to a rain rate time series, we get a rain attenuation time series [6]. The simulations refer to the following parameters. a) Carrier frequency at 11.6 GHz, circular polarization, although the main conclusions are independent of frequency and polarization. b) 32 slant path at Spino d’Adda (45.5 , 9.5 , 84 m above sea level), a site in a ﬂat countryside near Milan, with rain conditions typical of the Po Valley. The elevation angle and carrier frequency are those of a former ), used in radio link to geostationary satellite Sirio (15 the 1970s and 1980s for radio propagation experiments in the 12-GHz band at Spino d’Adda. c) The duration of the contiguous (clock) periods assumes the following values: 1, 5, 10, 15, 30, 60 (1 h) min, and 2, 4, 6, 12, 24 h, synchronized with 00:00 hours of the Universal Time (UT). Fig. 1 shows an example of a simulated time series, which shows a 2-h interruption between 16:00 and 18:00 h UT (17:00–19:00 local time). The type of interruption shown reminds of fade duration for a ﬁxed interval [12], but it is not exactly the same, for two reasons: (a) we have observed ﬁxed contiguous periods of the day because the services (e.g., TV, audio broadcasting) start independently of weather conditions and may be synchronized to the hour, to 15 min past the hour, etc., and (b) the fade level exceeded in these intervals can be larger than dB (threshold) at both the start and end times, while fade durations statistics always concern the same value of rain attenuation at both times. For system design the results provide data on interruptions when the built-in power margin, (outages) of duration

Fig. 1. Example of a rain attenuation time series generated by the synthetic storm technique at 11.6 GHz, circular polarization, in a 32 slant path at Spino d’Adda. For 1 dB threshold, for instance, a 2-h partial outage occurs between 14:00 and 16:00 UT (15:00 and 17:00 local time), a 2-h full outage occurs between 16:00 and 18:00 UT (17:00 and 19:00) and another (synchronized) 2-h partial outage at 18:00 UT (19:00).

(dB), is continuously less than in any clock period of dura, the results yield the long term tion . For example, for probability distribution that is exceeded in any 1-h period of the day. Now, since in a day there are 24 such contiguous periods, the statistics have been averaged over all the total number of 1-h contiguous intervals in the observation period. The results have been derived from a statistically reliable set of rain rate time series (for a deﬁnition of a rain storm and its duration, as measured with a rain gauge, see [6]) collected at Spino d’Adda, with a continuous observation from October 1979 to December 1982 (326 rain storms) and from May 1992 to December 1997 (804 rain storms), and with a random observation from 1983 to 1987 (103 rain events), 1233 rain storms in total, a very large database. Of the 1983–87 period we have estimated an equivalent continuous observation period of about 885 600 min (1.7 years). The total observation period amounts to 5 531 040 min, about 10.6 equivalent years, i.e., the order of magnitude of the life of a commercial satellite, and the total time is 225 425 min, i.e., 4.1% of an average year. As a consequence the total number of contiguous periods observed is given by 5 531 040 (min) divided by (min), e.g., 92 184 h for . III. SIMULATED RESULTS In this section we show the simulated results and discuss the and the probimportant connection between the long term ability distribution of 1-min long fades. A. Statistics for Different Values of Table I reports the overall statistics on the number of outages (interruptions) of duration in a day, for a given rain attenuation threshold (dB), for thresholds up to 10 dB. We like to show these numerical values because they can be useful to simulate systems directly, and also because future predictions derived from theoretical models could be compared to the results of this rather large database. The trends shown in Table I are physically justiﬁed: the longer is , the less likely is a full outage. For instance, for threshold

MATRICCIANI: SERVICE ORIENTED STATISTICS OF INTERRUPTION TIME DUE TO RAINFALL

2085

TABLE I NUMBER OF OUTAGES (INTERRUPTIONS) FOR CONTIGUOUS (CLOCK) PERIODS OF DURATION T (min) OF THE DAY, AS A FUNCTION OF RAIN ATTENUATION THRESHOLD A (dB) AT 11.6 GHz, IN A 32 SLANT PATH, AT SPINO d’ADDA, IN 10.6 YEARS OF OBSERVATION

the 24-h interval was never in outage conditions (not explicitly shown in Table I because we found no outages). On the contrary the two 12-h intervals experienced 14 outages (interruptions), the 1440 1-min intervals experienced 225 425 and 890 for , in 10.6 years outages for of observation. By dividing the data of Table I by 10.6 we get the average number of service interruptions in a year, given that . For the system designer, however, probabilities are more meaningful. If, for each , the data of Table I are divided by , we obtain the long term probabilities, drawn in Fig. 2 for , that a service of duration is not provided continuously to a user in any of the contiguous . As it clearly periods of duration , given that appears, the longer the service duration , the less likely it may be interrupted for the entire period. Notice, however, that these statistics, once they are complemented to unity, cannot be read of the order as availability statistics (except for values of of 1 min, see Section III-B), because the “availability” would surely include shorter intervals of partial outages (see Fig. 1 and Section IV-B). The probabilities discussed above are long term results averaged in a day, so that they do not provide hints on diurnal peri’s [11]. Figs. 3 odic cycles, as those found in the long term and and 4 show some examples of these cycles for , respectively, and for some values of threshold . Also shown, as horizontal lines, the equivalent uniform distribution of outages, i.e., the total number of outages divided by the number of contiguous periods in a day. We notice that the simulated distribution is not uniform. Only some intervals show a partial uniform distribution. The curves become smoother, of

Fig. 2. Probability that rain attenuation A (dB), predicted by the synthetic storm technique, is exceeded in an average year at 11.6 GHz, in a 32 slant path at Spino d’Adda. Observation time is 10.6 years.

course, when larger values of are considered (Fig. 4, ). These results agree with those reported in [8] and [11] and show, once more, that the more intense fades, in the Po Valley, tend to occur in the afternoon and in the evening. The consequence on system design is obvious, as service quality signiﬁcantly depends not only on service duration, but also on the time the service is started.

2086

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 3. Number of occurrences of full outages in clock intervals of 5 min along a day for thresholds 0 dB (upper ﬁgure), 1 dB (middle) and 3 dB (lower) at 11.6 GHz, 32 slant path. The local time is one hour ahead of Universal Time (UT). Notice that, to show more clearly the cycles, 24 6 repeated hours are shown. Numerical values are reported at the beginning of the interval. Observation time is 10.6 years. Horizontal dotted lines yield the equivalent uniform distributions.

+

Fig. 4. Number of occurrences of full outages in clock intervals of 15 min along a day for thresholds 0 dB (upper ﬁgure), 1 dB (middle) and 3 dB (lower) at 11.6 GHz, 32 slant path. The local time is one hour ahead of Universal Time (UT). Notice that, to show more clearly the cycles, 24 6 repeated hours are shown. Numerical values are reported at the beginning of the interval. Observation time is 10.6 years. Horizontal dotted lines yield the equivalent uniform distributions.

+

B. Relation Between 1-min Long Fades

and the Probability Distribution of

(the usual sampling time of rain attenuation stance, for measurements) we would simply obtain:

The results for shown in Table I, , once referred to the period of observation (10.6 years or 5 531 040 min), yield the long-term probability distribution (or fraction of time) that a given rain attenuation is exceeded. mentioned in the introduction and This is the long term currently used to design satellite systems, and obtained by measurements or predictions. In fact, it is very likely, from a physical point of view, that during an interval as short as 1 min, the channel is fully affected by rain attenuation, i.e., continuously, so that, for a given threshold , the number of outages in any shorter interval would be obtained by scaling the results accordingly and, as a consequence of being the 1-min interval fully attenuated, the relative frequencies would not change. For in-

(1) In our opinion this is one of the physical reasons why the predictions of the synthetic storm technique derived by means of 1-min rain rate time series yield a good estimate of the measurements [6], even if rain rate time series with smaller time resolutions can improve the estimates of the highest peaks of rain attenuation or its fastest rates of change [9]. Some experimental evidence supports this assumption. Together with full outages, we have also counted the number of partial outages, i.e., any interruption that lasts less than min. . For example, Fig. 1 shows 2 partial outages for

MATRICCIANI: SERVICE ORIENTED STATISTICS OF INTERRUPTION TIME DUE TO RAINFALL

2087

A rough estimate of the interval at which partial outages become full outages, for a given threshold exceeded, can be the (interpolated) value of for which the number of occurrences of partial outages equals the number of the full outages, i.e., partial and full outages are equally likely. Fig. 5 shows the curve , obtained for our database. We can notice that for so that intervals of about one hour are equally likely to be fully or partially attenuated. For larger than about is approximately 3 1–2 dB a clear “knee” appears so that or 4 min, regardless of the threshold . Fig. 5 predicts for and for , the latter being a threshold and a power margin very large for any practical application today. These ﬁgures are signiﬁcantly larger than 1 min, the physical resolution of the simulated rain attenuation time series, so that 1-min intervals should be very likely fully attenuated. Another physical justiﬁcation can be found in the frequency extension of the power spectrum of rain attenuation which falls off with a 20 dB/decade slope [13] up to a frequency of the order of few hundredths of hertz, i.e., periods of the order of 1 or 2 min. Beyond this frequency range, the spectrum is dominated by another physical phenomenon, namely turbulence scintillation [14], more responsible of fade durations of the order of few seconds, and here not of concern. In conclusions, with some justiﬁed conﬁdence, we can suppose that in a radio link, with a built-in power margin (dB) and affected by rain attenuation (dB), clock intervals of 1 min, or less, are very likely to be either fully unavailable (threshold is very likely to be continuously when exceeded during the interval ), or fully available when (threshold is very likely to be never exceeded during the interval ). IV. EMPIRICAL MODELS In this section we ﬁrst model the number of full interruptions, and afterwards we extend the concept of service unavailability to include some partial outages. A. Number of Full Interruptions For numerical calculations and extrapolations, the column data of Table I for can be modeled by simple mathematical functions of the power-law type, with a 0.1 dB -axis . The constants translation to include the threshold below were determined by minimizing the root mean square (rms) error. The model of the number of outages, as a function of the , is threshold (dB) and interval (min), for given by: (2) with an rms error less than 10% for any , and for thresholds up to 10 dB. The number of interruptions in an average year is given and the corresponding long term probability by distributions are obtained by dividing (2) by , Section II, i.e. (3)

Fig. 5. Interpolated value of T (min) for which the number of occurrences of partial outages equals the number of the full outages (partial and full outages are equally likely).

For (recall that means both the value of the attenuation and the threshold exceeded in the expression of any , and thus the value to insert in the case is 0) , for instance, (3) yields , i.e., 4.6%, and to be compared with the exact value 4.1% (corresponding to and , 225 425 min) of Section I. For : since there are 52 560 contiguous we get intervals of 10 min in 1 year, (3) yields interruptions in an average year (Table I, accordingly, gives ). B. Number of Synchronized Partial Interruptions Equation (2), or the data shown in Table I, can be used to extend the concept of service unavailability by taking into account synchronized partial outages, i.e., by considering the interruptions that last an interval , with the maximum continuous interruption tolerated by users (e.g., Table I). The is the meaning of a synchronized partial outage of duration following: it is a continuous interval during which the service is unavailable for a time less than the maximum tolerated by users, but always synchronized with 00:00 UT. Fig. 1, for instance, shows, for , a synchronized partial outage starting at 18:00 UT. Let us see this point by an example taken from Table I. Consider . We have 70 outages of 15 min , and 151 outages of 10 min (by reading the corresponding row of Table I). Now, since a 15-min outage includes some 10-min interruptions, namely only two out of three, there must have been synchronized partial outages of . In fact during a period of , a 10-min 10 min interval (and thus a possible 10-min interruption) can start at , , of the next 15-min period. The 10-min , however, cannot give rise interruption starting at to a synchronized 15-min interruption because it is split into

2088

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

second time with as parameter. The ratio between the latter and the former yields the scaling ratio to apply to current measurements, or predictions, for obtaining new outage statistics for , i.e. clock periods of (5) In (5) we have used the notation to refer to our reference 32 slant path at 11.6 GHz. Now the probability distribution of and clock interval , the full outages, for a given threshold is given by (6) Equations (5) and (6) may provide a good estimate of the measurements at sites other than Spino d’Adda, but with the same type of climate, because of the following argument: although (2) and (3) apply to this latter site, for a given A the ratio (5) can be more independent of site and particular rainy conditions. At another locality, the occurrences of 1-min and interruptions, for a given , can be largely different of those found at Spino d’Adda, but it is likely that both occurrences change in the same way so as to give a ratio that reminds of that given by (5). A ratio gets rid of possible common multiplying factors. Partial outages can be scaled in the same way. This provisional conclusion, however, should be tested against real data. Notice that (5) applies only to 11.6 GHz rain attenuation and to a 32 slant path. To apply it to slant paths with different radio electrical and geometric parameters we have to scale rain attenuation. To scale rain attenuation from a reference elevation angle (32 in our case) to a different elevation angle , we can suppose that (dB) is, to a ﬁrst approximation, proportional to the (average) rainy path length. Hence, for a ﬁxed rain height (as assumed in [6]), we get (7) As for frequency scaling, at microwaves we know that, for a given probability, rain attenuation (dB) exceeded at a carrier can be empirically related to rain attenuation frequency (dB) exceeded at a carrier frequency , in the same slant path, by the power law [15] (8) V. SCALING TO OTHER SITES, ELEVATION ANGLES AND FREQUENCIES The results of Section IV can be used to scale long term ’s, measured or calculated from one of the several prediction models available in the literature, to a certain with different values of , at least to sites with the same climate of the Po Valley. As discussed in Section III, we can assume that the reference time of the measured or predicted is, to a rough approximation, equal to about 1 min, and and the then apply (3) twice: the ﬁrst time with , in decibels, at carrier In conclusion, rain attenuation frequency (GHz) in a slant path with elevation angle , can be scaled according to (5)–(8), if, in (5), we insert the corresponding value of rain attenuation at 11.6 GHz for a 32 slant path, i.e., the value

Fig. 6. Model derived percentage of synchronized partial outages (duration T ), as a function of rain attenuation threshold, for several values of clock period T , for a ﬁxed ratio T =T = 0:8. Notice that when T = 1:25 min, T = 1 min. All curves intersect at A = 0:9 dB.

twointerruptions belonging to different 15-min contiguous periods. This splitting is taken into account by the presence of the , also justiﬁed by a uniform multiplying factor distribution within a short interval (Figs. 3 and 4). Of course, the results on synchronized partial outages do not include 10-min partial outages, which are not synchronized with the clock periods. They are neither present, of course, in the 10-min statistics of Table I. In conclusion, from our results we can easily estimate only the number of partial outages that start at the beginning of the time slot of duration , and at multiples of , as long as the ending time of the partial interruption is within the period . Since (2) is a function of both and , we can ﬁnd the , as number of the synchronized partial outages, (4) Fig. 6, as an example, shows the percentage of partial outages , for , for several values of . We see, for instance, that for and , the number of synchronized partial outages, of duration , is about 38% of the total number of 15-min interruptions (about 74 according to (2), or 70 according to Table I), i.e., . By (4) we can calculate all intermediate cases between 1 min and 15 min, as necessary.

(9)

MATRICCIANI: SERVICE ORIENTED STATISTICS OF INTERRUPTION TIME DUE TO RAINFALL

2089

In (9), is a constant, once and are ﬁxed. Equation (5) can to . For thus be used to ﬁnd the ratio to scale instance, if we want to scale rain attenuation measured [or pre] at in a slant dicted, we must know , so that path (zenith) to a different , from (9) we get rain attenuation to be inserted into (5) is reduced to 22%. Now, . For instance, we can calculate the scaling ratio for 10 dB in a 90 slant path at 40 GHz yields 2.2 dB at 11.6 GHz , from (5) in the reference 32 slant path. Now, if , so that, from (6), we get . From this probability (or relative frequency) it is then possible to ﬁnd the number of 10-min interruptions, or synchronized partial interruptions, in an average year, according to Section IV. VI. DISCUSSION AND CONCLUSION The simulated results obtained by the synthetic storm technique can be considered, in our opinion, as experimental ones. They provide statistics on the interruptions (outages), due to rainfall, in contiguous (clock) periods of the day of duration . The results are then useful to assess the quality and unavailability of services of duration during a day, e.g., in a satellite broadcasting system, or in a system using troposphere “geostationary” platforms 20 km aloft, in brief, in any earth-space radio link. The numerical results at 11.6 GHz in a 32 slant path (Table I) and the best ﬁt and extrapolation formulas (2)–(9) can provide a rough approximation to the same statistics for other elevation angles, clock periods and carrier frequencies, at sites with the same climate of the Po Valley. As for broadcasting system design, the data in Table I show, for instance, that a TV program that lasts 1 h is likely to be interrupted (“blocked”) in Northern Italy for times in an average year, if no built-in power margin or FEC ). For a power margin of 3 dB gain are available (case (case ), in an average year (data of Table I, divided by 10.6) there is one interruption of 30 min, 49 interruptions of 5 min, 365 interruptions of 1 min. or less Now, let us consider a probability of bit error (e.g., see [16], p. 450), and code the bit stream with one of the latest turbo codes with code rate 1/2 and 18 iterations. Assuming that in clear-sky conditions there are enough power and bandwidth for the doubled bit rate, the turbo code yields a gain of and the about 10 dB in the ratio between the energy per bit one-sided white noise power density , , e.g., see [16, p. 282, ﬁg. 7.4]. Then, in an average year, in the Milan area, by providing a system power margin (relative to clear-sky conditions) only by means of the most effective turbo code, in an average year there would be 84 interruptions of 1 min (Table I, , 890/10.6), 8.4 interruptions of 5 min (some of which can include, of course, 5 contiguous 1-min interruptions), 2.3 interruptions of 10 min, etc. These numbers provide results averaged over a day (the uniform distribution of Figs. 3 and 4), but in speciﬁc parts of the day, as Figs. 3 and 4 show, these values can be exceeded several times, especially in the afternoon and in the evening at Spino d’Adda, and a ﬁner system design should consider this fact.

The application of FEC coding schemes, as currently done, may make rainfall attenuation a less severe problem for a satellite broadcasting system, if the system is designed by taking into and account not the conventional probability distribution – ) and thus very short its low values (i.e., the range fade durations usually considered in telephony services (channel unavailability), but the number of interruptions of a maximum duration that a user can tolerate for a certain service, a TV or radio show, etc., (service unavailability). Once a clear distinction between channel unavailability and service unavailability is established, the statistics and the scaling methods proposed in this paper could be used to design an earth-space system based on service unavailability rather than on channel unavailability. At higher frequencies (e.g., 30-40-50 GHz), coding gain may be, however, largely ineffective, both for the “quasistatic” extra attenuation due to oxygen, water vapor and clouds, and for rain attenuation. For the latter, long random interruptions will be experienced, as Table I, would show, once were scaled. REFERENCES

[1] J. Goldhirsh, “Cumulative slant path rain attenuation statistics associated with the comstar beacon at 28.56 GHz for Wallops Island, VA,” IEEE Trans. Antennas Propagat., vol. 27, pp. 752–758, 1979. [2] H. W. Arnold, D. C. Cox, and A. J. Rustako, “Rain attenuation at 10–30 GHz along earth-space paths: elevation angle, frequency, seasonal, and diurnal effects,” IEEE Trans. Commun., vol. 29, pp. 716–721, 1981. [3] H. Fukuchi, T. Kozu, K. Nakamura, J. Awaka, H. Inomata, and Y. Otsu, “Centimeter wave propagation experiments using the beacon signals of CS and BSE satellites,” IEEE Trans. Antennas Propagat., vol. 31, pp. 603–613, 1983. [4] M. Mauri, “Hourly variation of attenuation statistics at 11–12 GHz as seen with SIRIO,” presented at the 32 Congr. Scientiﬁco Int. per l’Elettronica, Rome, 1985, pp. 81–85. [5] P. J. I. de Maagt, S. I. E. Touw, J. Dijk, G. Brussaard, L. J. M. Wijdemans, and J. E. Allnut, “Diurnal variations of 11.2 GHz attenuation on a satellite path in Indonesia,” Electron. Lett., vol. 29, pp. 2149–2150, 1993. [6] E. Matricciani, “Physical-mathematical model of the dynamics of rain attenuation based on rain rate time series and a two-layer vertical structure of precipitation,” Radio Science, vol. 31, pp. 281–295, 1996. , “Prediction of fade durations due to rain in satellite communica[7] tion systems,” Radio Science, vol. 32, pp. 935–941, 1997. [8] , “Diurnal distribution of rain attenuation in communication and broadcasting satellite systems at 11.6 GHz in Italy,” IEEE Trans. Broadcasting, vol. 44, pp. 250–258, 1998. [9] E. Matricciani, M. Mauri, and C. Riva, “A rain rate data base useful to simulate reliable rain attenuation time series for applications to satellite and tropospheric communication systems,” in Proc. Eur. Conf. Wireless Technology (ECWT 2002), Milan, Sept. 26–27, 2002, pp. 265–268. [10] S. Ventouras, I. Otung, and C. Wrench, “Simulation of satellite systems operating at Ka-band and above using experimental time series of tropospheric attenuation,” in Proc. Inst. Elect. Eng. Colloquium on Simulation and Modeling of Satellite Systems, London, U.K., Apr. 2002, pp. 11/1–11/5. [11] E. Matricciani, “An assessment of rain attenuation impact on satellite communication: matching service quality and system design to the time of the day,” Space Commun., vol. 16, pp. 195–205, 2000. [12] A. Safaai-Jazi, H. Ajaz, and W. L. Stutzman, “Empirical models for rain fade time on Ku- and Ka-band satellite links,” IEEE Trans. Antennas Propagat., vol. 43, pp. 1411–1415, 1995. [13] E. Matricciani, “Physical-mathematical model of dynamics of rain attenuation with application to power spectrum,” Electron. Lett., vol. 30, pp. 522–524, 1994. [14] E. Matricciani, M. Mauri, and C. Riva, “Scintillation and simultaneous rain attenuation at 12.5 GHz to satellite Olympus,” Radio Science, vol. 22, pp. 1861–1866, 1997. [15] G. Drufuca, “Rain attenuation statistics for frequencies above 10 GHz from raingauges observations,” J. Rech. Atmosperique, pp. 399–411, 1974. [16] T. Pratt, C. W. Bostian, and J. E. Allnutt, Satellite Communications, 2nd ed. New York: Wiley, 2003.

2090

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Emilio Matricciani was born in Italy in 1952. After serving in the army, he received the Laurea degree in electronics engineering from Politecnico di Milano, Milan, Italy, in 1978. He joined Politecnico di Milano in 1978 as a recipient of a research scholarship in satellite communications, and in 1981, he became an Assistant Professor of electrical communications. In 1987, he joined the Università di Padova, Padua, Italy, as an Associate professor of microwaves. Since 1991, he has been with Politecnico di Milano as an Associate Professor of electrical communications. In the year 2001, he qualiﬁed as a Full Professor of telecommunications. He has been involved in the experiments conducted with the Italian satellite Sirio in the 1970s in the 12–14 and 18 GHz bands, and afterwards, in the planning and conducting the experiments with Italsat in the 20–30 and 40–50 GHz bands, in the 1980s and 1990s. His actual research interests include satellite communications for ﬁxed and mobile systems, radio wave propagation, history of science and technology. In addition to the institutional activities, he teaches the fundamental aspects of communicating scientiﬁc and technical information to undergraduate, graduate, master, and doctorate students.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2091

**Full-Wave Analysis of Dielectric Frequency-Selective Surfaces Using a Vectorial Modal Method
**

Angela Coves, Benito Gimeno, Jordi Gil, Miguel V. Andrés, Member, IEEE, Angel A. San Blas, and Vicente E. Boria, Senior Member, IEEE

Abstract—A novel vectorial modal method is presented for studying guidance and scattering of frequency-selective structures based on lossy all-dielectric multilayered waveguide gratings for both TE and TM polarizations. The wave equation for the transverse magnetic ﬁeld is written in terms of a linear differential operator satisfying an eigenvalue equation. The deﬁnition of an auxiliary problem whose eigenvectors satisfy an orthogonality relationship allows for a matrix representation of the eigenvalue equation. Our proposed technique has been applied to the study of lossy all-dielectric periodic guiding media with periodicity in one dimension. This method yields the propagation constants and ﬁeld distributions in such media. The reﬂection and transmission coefﬁcients of a single layer under a plane-wave excitation can be obtained by imposing the boundary conditions. Study of the scattering parameters of the whole multilayered structure is accomplished by the cascade connection of components as characterized by their scattering parameters. Results obtained with this method for the propagation characteristics of a one-dimensional periodic dielectric medium are compared with those presented by other authors, and results for the scattering of several dielectric frequency-selective surfaces (DFSS) are compared with both theoretical and experimental results presented in the literature, ﬁnding a very good agreement. A symmetrical band-stop ﬁlter with a single waveguide grating is also demonstrated theoretically. Index Terms—Band-stop ﬁlter, dielectric grating, frequencyselective surfaces, Galerkin method, orthogonality relationship, vectorial modal method.

I. INTRODUCTION

T

HE frequency-selective characteristics of multilayered periodic structures, both dielectric [1]–[5] and metallic [6]–[8] are of considerable importance in electromagnetics. Periodic screens have been used in the last decades in many applications as ﬁlters, radomes and polarizers of electromagnetic waves. A dielectric layer with periodically varying dielectric constant can be used as an alternative way to obtain a frequency selective surface. Dielectric waveguide gratings can be used in the microwave-frequency band as dichroic subreﬂectors in

Manuscript received September 25, 2002; revised February 4, 2003. This work was supported by the Ministerio de Ciencia y Tecnología, Spanish Government, under Research Projects TIC2000-0591-C03-01 and TIC2000-0591C03-03. A. Coves and A. A. San Blas are with the Departamento de Física y Arquitectura de Computadores, Universidad Miguel Hernández de Elche, 03202 Elche (Alicante), Spain. B. Gimeno and M. V. Andrés are with the Departamento de Física Aplicada y Electromagnetismo- I.C.M.U.V., Universidad de Valencia, 46100 Burjasot (Valencia), Spain. J. Gil and V. E. Boria are with Departamento de Comunicaciones, Universidad Politécnica de Valencia, 46022 Valencia, Spain. Digital Object Identiﬁer 10.1109/TAP.2004.832507

large reﬂecting antennas [9]. Thin-ﬁlm dielectric structures containing a periodic variation along the ﬁlm have recently been of considerable interest in integrated optics [10], because of the important role they play in applications such as beam-to-surface-wave couplers, ﬁlters, distributed feedback ampliﬁers and lasers, nonlinear generation of second harmonics, and beam reﬂection on steering devices of the Bragg type. Eielectric frequency-selective surfaces (DFFS) have been analyzed with a variety of different methods, both analytical and numerical. Rigorous analytical methods [1], [11] for studying dielectric periodic structures are limited however to gratings with special simple groove shapes. Numerical methods [12], on the other hand, have made possible the analysis of periodic structures of signiﬁcant geometry/material complexity in the periodic cell. The numerical method presented in this paper is based on a novel vectorial modal method [13], [14] for studying guidance and scattering by lossy all-dielectric guiding periodic structures. For the structures considered here, each layer is either a periodic dielectric grating, formed by any number of lossy dielectric slabs, or a uniform dielectric slab, and all grating layers have the same periodicity. This formulation has been applied to the accurate analysis of the modal spectrum (propagation constants and ﬁelds distribution) of dielectric periodic structures for both TE and TM polarizations, and results are successfully compared with those presented in the technical literature. The reﬂection and transmission coefﬁcients of the structures under a plane-wave excitation are obtained with the generalized scattering matrix (GSM) theory [15], [16] and successfully compared with both theoretical and experimental results obtained by other authors. It must be emphasized that the complexity of the new method proposed in this paper does not increase with the number of dielectric slabs present in the unit cells, such as it usually happens with other classical analysis techniques (e.g., the TRM technique [17]). Furthermore, the presence of losses in the dielectric slabs can be easily considered by simply introducing a complex permittivity in the formulation derived. The formalism used to obtain the modal spectrum of lossy dielectric periodic structures and the orthogonality relations of the modes is presented in Section II. In Section III-A, numerical results obtained using this formalism for the propagation constants and the ﬁelds distribution in the grating region are compared with those presented in the bibliography. The spectral response of several DFSS are obtained and compared with both numerical and experimental results obtained by other authors in Section III-B, and a symmetrical band-stop ﬁlter with a single waveguide grating is also proved theoretically.

0018-926X/04$20.00 © 2004 IEEE

2092

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 1. Characteristic unit cell of a periodic dielectric medium.

II. THEORY The ﬁrst objective of this section is to obtain the modal spectrum of a general lossy all-dielectric periodic guiding axis. Next, making use of this medium uniform in the multimodal characterization, the boundary conditions between adjacent layers will be imposed by means of the GSM approach in order to obtain the scattering parameters of a one-dimensional (1-D) dielectric FSS. Next we present the theoretical bases of the method for direction studying a dielectric 1-D periodic medium in the with translational invariance along the axis (see Fig. 1). We assume that the electric and magnetic ﬁelds in this medium can be expressed as a linear superposition of ﬁelds with explicit harmonic dependence on the coordinate (we assume that the time dependence is always implicit and has an harmonic form for all vector ﬁelds) (1) (2) where is the propagation constant, and , represent the transverse components to the direction of the electric and magnetic ﬁelds, respectively, and , are the components in the direction. The transverse components of the ﬁelds, when no sources are present, satisfy the vector wave equations [13], [14] (3) (4) the relative complex permittivity of the pebeing riodic medium, whereas is the free-space wavenumber . In these equations, we can identify in square brackets the differential operators governing the transverse components evolution along the axis. For our purposes, it is more interesting to rewrite (3) and (4) as a pure 2-D eigenvalue problem for the differential evolution operators and (5) operators, where and are the eigenvectors of the and respectively; these operators are nonself-adjoint operators. The

eigenvectors of nonself-adjoint operators do not satisfy a standard orthogonality condition; this prevents the possibility of expanding arbitrary functions in terms of the operator eigenvectors, which is in fact one of our objectives: obtain a modal basis to represent the electromagnetic ﬁeld. But we can still represent (3) and (4) in a matrix form if such equations are expressed in an auxiliary system which can be used to derive the matrix form of the eigenvalue problem, whose eigenvectors satisfy an orthogonality relationship of the form (6) For the auxiliary system we have used the modes corresponding to a homogeneous medium of relative dielectric permittivity as the auxiliary basis. The geometry of the problem is shown in Fig. 1; the periodicity D of the structure has been chosen direction, being the problem uniform in the axis. in the and periodic The medium is assumed as nonmagnetic direction, so it is fully characterized by its complex in the relative permittivity in the unit cell deﬁned as

(7) where is the Heaviside function, is the number of lossy dielectric slabs inside the periodic cell (which can be an arbitrary number), and the i-th dielectric slab is centered at point and its width is , as it is shown in Fig. 1. The more adequate auxiliary basis functions in this case are the well-known Floquet harmonics, i.e., the eigenfunctions of the evolution operator corresponding to a periodic structure immersed in a homoge[6]. For a medium inneous medium of relative permittivity variant with the coordinate , the auxiliary basis of the Floquet modes is formed by the set of TE and TM family modes [6] (8) (9)

COVES et al.: FULL-WAVE ANALYSIS OF DFSS USING VECTORIAL MODAL METHOD

2093

where (10) and is the angle of incidence of the fundamental harmonic associated with the structure excitation. These modes satisfy the orthogonality relationship (6) (11) Thus the modes of the real problem can be expanded in terms of the auxiliary system as (12) where and are the complex coefﬁcients of the modal expansion for the transverse magnetic and electric ﬁelds of the th mode, respectively. We are certainly concerned with the matrix representation of the linear operator of the real problem . The basis will then matrix elements of the operator in the be easily obtained by applying the standard Galerkin moment method [18]. By inserting the ﬁrst equation of (12) into the ﬁrst equation of (5), and applying the linear properties of , we ﬁnd (13) The next step in the application of the Galerkin procedure is , and to take the to choose a set of weighting functions yielding the following linear matrix inner product for each eigenvalue problem

function, which is a discontinuous function, and also its transverse gradient. Some confusion may exist in the evaluation of these integrals due to the discontinuous nature of the dielectric function and its transverse gradient. In [19] it is described the way to solve these integrals correctly for the case of a rectangular waveguide with a relative permittivity function deﬁned as a sum of lossy dielectric slabs. For the particular case of a 1-D periodic dielectric medium, these integrals have been analytically calculated for each polarization obtaining

(18)

(19)

where the summation includes all the lossy dielectric slabs inside the periodic cell. As a consequence, only a diagonalization process for each polarization has to be performed numerically for obtaining the propagation constants and the magnetic ﬁelds of the periodic medium at each frequency point. At this point, it is important to remark that we have transformed the differential eigenvalue (5) for the operator , which is responsible of the evolution of the transverse magnetic ﬁeld, into a linear matrix eigenvalue problem. An analog equation for the operator , responsible of the evolution of the transverse electric ﬁeld, can be derived (20) Thus, the information contained in the above matrix (14) and (20) is the same as in the differential equations shown in (5). Diagonalization of the matrix yields the squared of the th mode propagation constant (the th eigenvalue of ), through the and also its transverse magnetic amplitude knowledge of the th eigenvector . It is important to note not only provides us with the that the diagonalization of propagation constants (eigenvalues) and transverse magnetic amplitudes (eigenvectors) of the modes, but also with the transverse electric ﬁelds of the modes, which are related to through constraints derived from Maxwell’s equations [20]. This fact is very important from a computational point-of view, matrix is only because the diagonalization process for the requested in the numerical implementation of this method. Therefore, it is not necessary to implement the numerical diagonalization of the matrix corresponding to the electric . operator However, the matrix has an inﬁnite number of elements. In order to develop a realistic method, we have to work with a ﬁnite set of well-known auxiliary ﬁelds. Unfortunately, there are no general conditions that guarantee the convergence of the expansions. This convergence will depend on both the nature of

were the matrix elements of the follows:

(14) operator are obtained as

(15) For practical purposes, it is convenient to introduce the difference operator , resulting (16) Thus, the matrix elements of the operator tained by means of (15) are ob(17) where the ﬁrst term is diagonal because the operator is expressed in its own orthogonal basis. However, it can be easily shown that there is no coupling among TE and TM modes when evaluating the integrals of the matrix elements of the operator. Thus, the problem can be easily separated into two TE and TM polarizations. The integrals involved in the evaluation of the matrix elements of contain the relative dielectric permittivity

2094

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

the operator and the auxiliary problem chosen to deﬁne the orthogonal basis. In general, we observe that the modes are better described by increasing the number of auxiliary modes. In the same way, auxiliary bases that can represent the most relevant features of the real problem produce faster convergence. In any case, numerical convergence tests must be done by sweeping the number of auxiliary modes over meaningful ranges and studying the stability of solutions. Finally, the transverse electric ﬁeld of the th mode is related through constraints given by Maxwell’s equations, which to in this case can be expressed in terms of the modal characteristic impedances [6] (21) (22) where these modal characteristic impedances are given by (23) (24) Note that the TM characteristic impedance varies with the coordinate, that is, the characteristic impedance is not a constant because the relative permittivity is a function of the coordinate. Once obtained the ﬁelds and propagation constants in all regions of the structure, both homogeneous and periodic, the problem is reduced to obtain the scattering parameters of the structure. To this end, the boundary conditions between adjacent layers will be imposed, obtaining the GSM at each interface between adjacent layers of the structure, i.e., the amplitudes of reﬂected and transmitted modes. For a propagation distance through a layer with propagation constants , the multimode scattering matrix is deﬁned as the scattering matrix of a uniform transmission line of length [16]. Then, we construct the GSM of the global structure by means of the cascaded connection of the individual GSMs of the interfaces and the scattering matrices corresponding to the propagation through the layers, following the technique described in [16]. The global GSM yields the amplitudes of the scattered modes, which are reﬂected and transmitted by the structure, considering an incident plane wave with a unit amplitude. Finally, the reﬂectance and transmittance of the structure are easily obtained. In the described theory, when all modes are included, the matrix is inﬁnitely-dimensional. In order to develop a realistic method, we have to work with a ﬁnite set of well-known auxiliary ﬁelds to expand the modes of each periodic dielectric layer in terms of the modes of the auxiliary basis functions. On the other hand, also the multimode scattering matrix is inﬁnitely-dimensional. The most straightforward way to reduce the scattering problem to a computationally tractable form is to truncate the individual layer multimode scattering matrices at a ﬁnite size which is large enough to allow for accurate calculation of the scattered modes, reﬂected and transmitted, which are signiﬁcant in the overall solution but small enough to be tractable for numerical calculation. Then, for each particular

case, a study of convergence must be performed in order to reach an accurate solution for both the propagation characteristics in each periodic layer, and the scattering parameters of the overall structure. III. NUMERICAL AND EXPERIMENTAL RESULTS A. Accuracy and Convergence Properties of the Modal Expansion in Periodic Dielectric Media In order to compare our results for the propagation characteristics of an inﬁnite periodic dielectric structure with those obtained previously by other authors, we ﬁrst examine an inﬁnite periodic structure (see Fig. 1) with ﬁve dielectric slabs . The paramewithin the unit cell of period , , , , ters are , , , , , , and . In the calculations the auxiliary system used was an inﬁnite homogewith the same period neous dielectric medium with of the real problem. Results of the convergence study of the solutions of this inﬁnite periodic medium are shown in Fig. (2a) and (b). In these ﬁgures it is represented, for TE [Fig. 2(a)] and TM [Fig. 2(b)] polarization, the normalized propagation conof the ﬁrst and third mode (left-hand axis), and stant the sixth mode (right-hand axis) as a function of the number of auxiliary modes, which in this case are the Floquet modes of an inﬁnite homogeneous medium with relative permittivity . Results are obtained for a frequency of 10 GHz and for a normalized Floquet wavenumber . For TE polarization it can be seen that for all modes represented the convergence is reached with only 15 modes of the auxiliary basis, taking 0.01 s per frequency point for obtaining the ﬁrst ﬁfteen modes (in a Pentium II@350 MHz processor). Nevertheless, a higher number of auxiliary modes is needed for the case of TM polarization in order to reach the convergence of the solutions, auxiliary Floquet modes, which which in this case is takes again only 0.01 s per point. For the same problem considered before, Fig. 3 shows the as a funccurves of the normalized propagation constant tion of the normalized Floquet wavenumber of the ﬁrst and second mode for both TE and TM polarizations. In this ﬁgure the results calculated in [21] with the transverse resonant method (TRM) are also shown, and an excellent agreement is observed. The present method also allows the calculation of the ﬁeld patterns. The electric ﬁeld distribution of the ﬁrst and second mode along the characteristic cell for the geometry of Fig. 2 is shown in Fig. 4. Fig. 4(a) shows the behavior of the electric ﬁeld normalized with respect to its maximum value for the ﬁrst and second TE modes as a function of the normalized cofor values of (straight line) and 1 ordinate (dashed line). In this ﬁgure we have also represented the results and calculated in [21] for the ﬁrst TE mode at , showing an excellent agreement with our results. The normalized magnetic ﬁeld for the ﬁrst and second TM for values of mode is shown in Fig. 4(b) as a function of (straight line) and 1 (dashed line). We have also represented the results calculated in [22] for the ﬁrst TM mode at

COVES et al.: FULL-WAVE ANALYSIS OF DFSS USING VECTORIAL MODAL METHOD

2095

Fig. 2. Study of convergence of the normalized propagation constant j=k j as a function of the number N of auxiliary modes for the ﬁrst and third mode (left-hand axis), and sixth mode (right-hand axis) for both (a) TE and (b) TM polarizations of an inﬁnite periodic dielectric medium with ﬁve dielectric slabs within ~ the unit cell. Parameters: D = 17:99 mm, " = 1:0, " = 1:28, " = 2:56, " = 1:28, " = 1:0, " = 1:0, l = 5:105 mm, l = 2:09 mm, l = 3:60 mm, l = 2:09 mm and l = 5:105 mm. Results obtained for a normalized Floquet wavenumber (k D= ) = 0 and a working frequency of 10 GHz.

Fig. 3. Normalized propagation constant j=k j as a function of the normalized Floquet wavenumber (k D= ) of the ﬁrst and second mode for both TE and TM polarizations of the geometry detailed in Fig. 2. Comparison between our results (lines) and results obtained with the TRM method [21] (dots).

, and for the second TM mode at showing also a good agreement with our results. B. Plane-Wave Scattering by a DFSS

,

Following the analysis of Section II, we have carried out numerical studies of the reﬂection and transmission coefﬁcients of several dielectric FSSs in comparison with theoretical and experimental results obtained by other authors. The ﬁrst case analyzed is a periodic dielectric grating formed by two dielectric and thickslabs within the unit cell of period immersed in air under a TE plane wave ness ; we have set , incidence at , and . Results of the convergence study of the reﬂection coefﬁcient of this structure are shown in Fig. 5. In this ﬁgure it is represented the reﬂection coefﬁcient at three different frequencies as a function of the number of

included in the construction of the GSMs. The remodes ﬂectance for lower frequency is represented in the right-hand axis. The second and third frequency, represented in the left-hand axis, are resonant frequencies of the grating. For all the cases considered, the reﬂected and transmitted ﬁelds were , that is, the found to conserve power to within one part in total active reﬂected and transmitted power coefﬁcients related . to the propagation modes differ from unity by less than For the numerical results shown in the rest of this section, it was modes in the construction of the mulsufﬁcient to take timode scattering matrices. The frequency dependence of the reﬂectance of the grating is shown in Fig. 6; our results are compared with those calculated in [17] with the TRM. For this case we have chosen and , taking 0.01 s per frequency point, including the CPU time for the computation of the modes and the scattering analysis. The agreement between both methods is excellent. The second case analyzed is the transmittance of a periodic dielectric grating immersed in air for a TM-polarized wave at normal incidence. The structure is characterized by a periodic , with two dielectric slabs with cell of period parameters , (plexiglas), and (air); the width of the grating is (see . In the calculations a loss tangent of Fig. 7) and (lossy dielectric) was used, and we have considered an angle of incidence on the structure of to take into account the possible asymmetry of the experimental setup and , as explained in [23]. We have chosen taking 0.33 s per frequency point. Fig. 7 compares the calculated response using this method with the experimentally measured response, ﬁnding a good agreement. The third case analyzed is a multilayered periodic structure with two grating layers and six homogeneous layers [see Fig. 8(a)] between air and a substrate. The parameters of this , , , , structure are , , and . The and grating thicknesses are

2096

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 4. Field distributions along the unit cell of the ﬁrst and second mode for the geometry of Fig. 2. (a) Distribution of the normalized electric ﬁeld for the ﬁrst and second TE mode at (k D= ) = 0 (straight line) and 1 (dashed line). Results calculated in [21] are shown with dots. (b) Distribution of the normalized magnetic ﬁeld for the ﬁrst and second TM mode at (k D= ) = 0 (straight line) and 1 (dashed line). Results calculated in [22] are also shown with dots.

Fig. 5. Convergence study of the reﬂectance of a periodic dielectric grating in air as a function of the number of modes M included in the construction of the GSMs. The results are presented for three different frequencies under a TE plane wave incidence at = 45 . The grating has a period D = 1:0 mm and thickness h = 1:713 mm with parameters: l = l = 0:5 mm, " = 2:56, " = 1:44, " = 1:44. ~

Fig. 7. Calculated and measured transmittance of a periodic dielectric grating immersed in air for a TM-polarized incident wave. The grating period is D = 30:0 mm and the thickness is h = 8:7 mm. Parameters: l = l = 15:0 mm, " = 2:59, " = 1:0, " = 1:0. Comparison between our results (lines) ~ and the experimental ones (dots) presented in [23].

Fig. 6. Frequency dependence of the reﬂectance of the periodic dielectric grating detailed in Fig. 5 under a TE plane wave incidence at = 45 . Comparison between our results (line) and those calculated in [17] with the TRM (dots).

. In Fig. 8(b) we compare our results for the transmittance of the structure for a normal

TE-polarized incident wave with those obtained in [24] with the rigorous coupled-wave theory, showing a good concordance. and , taking 0.11 s per For this case we chose frequency point. Finally, a novel design of a reﬂection (band-stop) ﬁlter for TE-polarization has been performed using the presented theory. The design of a reﬂection ﬁlter involves the selection of the ﬁlter parameters such as the thickness of each layer either homogeneous or periodic and the dielectric permittivity distribution within the unit cell of the periodic layers, in order to achieve symmetrical line shapes and reduced reﬂectance around the central wavelength, for a particular polarization and angle of the incident wave. A simple reﬂection ﬁlter with a single-layer waveguide grating has been designed for a normal TE-polarized incident wave, containing a rectangular grating composed of two dielectric materials whose reﬂectance is shown in Fig. 9, with the structure illustrated in the inset. The spectral response shown in Fig. 9 can be predicted using classical unmodulated slab waveguide theory. For the greater part of the spectrum, the grating has the reﬂectance of a thin ﬁlm with a dielectric constant equal to the average dielectric constant of the grating. The grating thickness has been chosen

COVES et al.: FULL-WAVE ANALYSIS OF DFSS USING VECTORIAL MODAL METHOD

2097

Fig. 8. Transmittance of a multilayered periodic structure with two grating layers (layers 1 and 7) and six homogeneous layers between air and a substrate. Comparison between our results (line) and those calculated in [24] with the rigorous coupled-wave theory (dots).

ﬁlter with small sideband reﬂection. To maximize the efﬁciency of this device, a subwavelength grating period is chosen in order to permit that only the zero order Floquet mode to propagates in free space. The ﬁlter has been designed in the spectral range of 13–17 GHz, for the peak frequency centered , at 15 GHz, having the following parameters: , (E-glass), (silica), and . The spectral response of the in the range of 13–17 GHz ﬁlter in Fig. 9 shows that in the (except around the band-stop frequency), and range of 14.88–15.12 GHz, thus being the bandwidth (full width at half maximum of the reﬂected power) (1.6%). In the calculations we have chosen and , taking 0.03 s per frequency point.

Fig. 9. Reﬂection ﬁlter spectral response with the structure illustrated in the inset. The peak frequency is centered at 15 GHz. The ﬁlter parameters are: D = 11:28 mm, h = 4:37 mm, " = 6:13 (E-glass), " = 3:7 (silica), l = l = D=2.

IV. CONCLUSION P A vectorial modal method has been applied to analyze 1-D periodic dielectric media with any number of lossy dielectric slabs within the unit cell for both TE and TM polarizations. This formulation allows the study of electromagnetic scattering from multilayered periodic structures by means of the GSM theory. The study of a wide variety of conﬁgurations with small computational cost has been performed. Furthermore, we have tested the theory by comparison with theoretical and experimental results found in the technical literature, showing a good agreement. A reﬂection ﬁlter using practical materials has been designed for a central frequency of 15 GHz, showing low symmetrical sidebands in the range of 13–17 GHz. New techniques for ﬁlter design are opened applying the ﬁlter synthesis theory with transmission lines for band-stop, bandpass and low-pass ﬁlters with periodic structures. In future, we are also interested in the analysis of the 3-D oblique incidence upon the dielectric grating. ACKNOWLEDGMENT The authors would like to thank the Reviewers for their useful comments and suggestions.

to be half-wavelength for the central wavelength of the ﬁlter, so the spectral response shows a reduced reﬂectance around this wavelength. But at a speciﬁc frequency (resonance frequency) the diffractive character of the grating enables the incident wave to excite a leaky mode supportable by the grating, resulting in a transmission null, as shown in Fig. 9 [25]. The approximate value for the resonance frequency location can be predicted imposing the phase-match condition for the equivalent unmodulated slab , where is the propagation waveguide constant of the unmodulated waveguide in the y-direction, is the wavevector provided by the grating. At this and frequency the waveguide mode is excited, and this mode will reradiate plane waves into the air regions above and below the layer through the same space harmonic, thereby acting as a leaky wave. The reradiated waves interfere with the directly reﬂected and transmitted ﬁelds, and when the two components above the layer add in phase, strong reﬂection takes place. For a given thickness and dielectric materials of the grating, the resonance frequency can be adjusted choosing an adequate value of the grating period such that it coincides with the central frequency of the ﬁlter, resulting in a symmetrical reﬂection

2098

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

REFERENCES

[1] T. K. Gaylord and M. G. Moharam, “Analysis and applications of optical diffraction by gratings,” Proc. IEEE, vol. 73, pp. 894–938, May 1985. [2] M. K. Moaveni, “Plane wave diffraction by dielectric gratings, ﬁnitedifference formulation,” IEEE Trans. Antennas Propagat., vol. 37, pp. 1026–1031, Aug. 1989. [3] J. C. W. A. Costa and A. J. Giarola, “Analysis of the selective behavior of multilayer structures with a dielectric grating,” IEEE Trans. Antennas Propagat., vol. 43, pp. 529–533, May 1995. [4] C. Zuffada and T. Cwik, “Synthesis of novel all-dielectric grating ﬁlters using genetic algorithms,” IEEE Trans. Antennas Propagat., vol. 46, pp. 657–663, May 1998. [5] A. Coves, B. Gimeno, D. Camilleri, M. V. Andres, A. San Blas, and V. E. Boria, “Scattering by dielectric frequency-selective surfaces using a vectorial modal method,” in IEEE AP-S Int. Symp. and URSI National Radio Sci. Meeting, San Antonio, TX, June 16–21, 2002, pp. 580–583. [6] R. Mittra, C. H. Chan, and T. Cwik, “Techniques for analyzing frequency selective surfaces—a review,” Proc. IEEE, vol. 76, pp. 1593–1615, Dec. 1988. [7] K. Kobayashi, “Diffraction of a plane wave by a thick strip grating,” IEEE Trans. Antennas Propagat., vol. 37, pp. 459–470, Apr. 1989. [8] T. Cwik and R. Mittra, “The cascade connection of planar periodic surfaces and lossy dielectric layers to form an arbitrary periodic screen,” Proc. IEEE, vol. 76, pp. 1593–1615, Dec. 1988. [9] V. D. Agrawal and W. A. Imbriale, “Design of a dichroic Cassegrain subreﬂector,” IEEE Trans. Antennas Propagat., vol. 27, pp. 466–473, 1979. [10] S. S. Wang and R. Magnusson, “Design of waveguide-grating ﬁlters with symmetrical line shapes and low sidebands,” Opt. Lett., vol. 19, no. 12, pp. 919–921, Jun. 1994. [11] S. T. Peng, T. Tamir, and H. L. Bertoni, “Theory of periodic dielectric waveguides,” IEEE Trans. Microwave Theory Techniques, vol. 23, pp. 123–133, Jan. 1975. [12] W. P. Pinello, R. Lee, and A. C. Cangellaris, “Finite element modeling of electro-magnetic wave interactions with periodic dielectric structures,” IEEE Trans. Microwave Theory Tech., vol. 42, pp. 2294–2301, Dec. 1994. [13] E. Silvestre, M. V. Andres, and P. Andres, “Biorthonormal-basis method for the vector description of optical-ﬁber modes,” J. Lightwave Technol., vol. 16, no. 5, pp. 923–928, May 1998. [14] E. Silvestre, M. A. Abian, B. Gimeno, A. Ferrando, M. V. Andres, and V. E. Boria, “Analysis of inhomogeneously ﬁlled waveguides using a bi-orthonormal-basis method,” IEEE Trans. Microwave Theory Tech., vol. 48, pp. 589–596, Apr. 2000. [15] R. C. Hall, R. Mittra, and K. M. Mitzner, “Analysis of multilayered periodic structures using generalized scattering matrix theory,” IEEE Trans. Antennas Propagat., vol. 36, pp. 511–517, Apr. 1988. [16] T. S. Chu and T. Itoh, “Generalized scattering matrix method for analysis of cascaded and offset microstrip step discontinuities,” IEEE Trans. Microwave Theory Tech., vol. 34, pp. 280–284, Feb. 1986. [17] H. L. Bertoni, L. H. S. Cheo, and T. Tamir, “Frequency-selective reﬂection and transmission by a periodic dielectric layer,” IEEE Trans. Antennas Propagat., vol. 37, pp. 78–83, Jan. 1989. [18] D. G. Dudley, Mathematical Foundations for Electromagnetic Theory, 1st ed. Piscataway, NJ: IEEE Press, 1994. [19] A. Coves, B. Gimeno, M. V. Andres, J. A. Montsoriu, and E. Silvestre, “Evaluation of Discontinuities in Modal Vectorial Methods,” , submitted for publication. [20] R. E. Collin, Field Theory of Guided Waves, 2nd ed. New York: IEEE Press, 1991. [21] J. C. W. A. Costa and A. J. Giarola, “Electromagnetic wave propagation in multilayer dielectric periodic structures,” IEEE Trans. Antennas Propagat., vol. 41, pp. 1432–1438, Oct. 1993. , “Wave propagation in multilayer dielectric periodic structures,” [22] in IEEE AP-S Int. Symp. Digest, vol. 4, Ann Arbor, MI, 1993, pp. 1964–1967. [23] S. Tibuleac, R. Magnusson, T. A. Maldonado, P. P. Young, and T. R. Holzheimer, “Dielectric frequency-selective structures incorporating waveguide gratings,” IEEE Trans. Microwave Theory Tech., vol. 48, pp. 553–561, Apr. 2000. [24] R. Magnusson and S. S. Wang, “Transmission bandpass guided-mode resonance ﬁlters,” Appl. Opt., vol. 34, no. 35, pp. 8106–8109, Dec. 1995. [25] T. Tamir and F. Y. Kou, “Varieties of leaky waves and their excitation,” IEEE J. Quant. Electron., vol. 22, pp. 544–551, 1986.

Angela Coves was born in Elche, Spain, on May 20, 1976. She received the Licenciada degree in physics from the Universidad de Valencia, Valencia, Spain, in 1999, where she is currently working toward the Ph.D. She is currently with the Departamento de Física Aplicada y Electromagnetismo, Universidad de Valencia. Since 2001, she has been an Assistant Professor with the Departamento de Física y Arquitectura de Computadores, Universidad Miguel Hernández de Elche, Elche (Alicante), Spain. Her current research interests include numerical methods in computer-aided techniques for the analysis of microwave passive components such as waveguide structures and multilayered periodic structures including dielectric media.

Benito Gimeno was born in Valencia, Spain, on January 29, 1964. He received the Licenciado degree in Physics in 1987 and the Ph.D. degree in 1992, both from the Universidad de Valencia, Spain. He was a Fellow at the Universidad de Valencia from 1987 to 1990. Since 1990, he served as Assistant Professor in the Departmento de Física Aplicada y Electromagnetismo at the Universidad de Valencia, where he became Associate Professor in 1997. He was working at European Space Research and Technology Centre of the European Space Agency (ESTEC) as a Research Fellow during 1994 and 1995. In 2003, he obtained a fellowship from the Spanish Government for a short stay (three months) at the Universita degli Studi di Pavia, Pavia, Italy, as a Visiting Scientiﬁc. He is currently with the Departamento de Física Aplicada y ElectromagnetismoI.C.M.U.V., Universidad de Valencia, Burjasot (Valencia), Spain. His current research interests include the areas of computer-aided techniques for analysis of microwave passive components, waveguide and cavities structures including dielectric resonators and photonic band-gap crystals.

Jordi Gil was born in Valencia, Spain, on April 27, 1977. He received the Licenciado degree in physics from the Universidad de Valencia, in 2000, where he is currently working toward the Ph.D. He is currently working at the Ingegneria dei Sistemi IDS-S.p.A., Pisa, Italy. Since 2001, he has been a young researcher in the frame of the EC project MMCODEF “Millimeter-wave and Microwave Components Design Framework for Ground and Space Multimedia Network” in collaboration with the European Space Agency (ESA). His current research interests include numerical methods in computer-aided techniques for the analysis of microwave passive components such as waveguide structures and multilayered periodic structures including dielectric media.

Miguel V. Andrés (M’91) was born in Valencia, Spain, in 1957. He received the Licenciado en Física degree in 1979 and the Doctor en Física (Ph.D.) degree in 1985, both from the Universidad de Valencia. From 1983, he served successively as Assistant Professor and Lecturer in the Departamento de Física Aplicada, Universidad de Valencia. From 1984 to 1987, he was visiting for several periods the Department of Physics, University of Surrey, U.K., as a Research Fellow. Until 1984, he was engaged in research on microwave surface waveguides. His current research interests are optical ﬁber devices and systems for signal processing and sensor applications, and electromagnetic wave propagation in microwave and optical waveguides and devices. Dr. M. V. Andrés is a Member of the Institute of Physics (IOP).

COVES et al.: FULL-WAVE ANALYSIS OF DFSS USING VECTORIAL MODAL METHOD

2099

Angel A. San Blas was born in Fortaleny (Valencia), Spain, on September 20, 1976. He received the Ingeniero de Telecomunicación degree from the Universidad Politécnica de Valencia, in November 2000. In 2001, he was awarded a researcher position in the Departamento de Comunicaciones, Universidad Politécnica de Valencia, where he worked for two years. In 2003, he joined the Departamento de Física y Arquitectura de Computadores, Universidad Miguel Hernández de Elche, Spain, where he is currently an Assistant Professor. His current research interests include the analysis of discontinuities in waveguide structures, design of microwave ﬁlters, and coaxial excitation of microwave devices.

Vicente E. Boria (SM’02) was born in Valencia, Spain, on May 18, 1970. He received the Ingeniero de Telecomunicación degree with ﬁrst-class honors and the “Doctor Ingeniero de Telecomunicación” degree from the Universidad Politécnica de Valencia, Valencia, Spain, in 1993 and 1997, respectively. In 1993, he joined the Departamento de Comunicaciones, Universidad Politécnica de Valencia, where he was an Assistant Lecturer from 1993 to 1995, a Lecturer from 1996 to 1997, an Associate Professor from 1998 to 2002, and became Full Professor in 2003. In 1995 and 1996, he was awarded a Spanish Trainee position at ESTEC-ESA, Noordwijk, the Netherlands, where he worked in the area of electromagnetic analysis and design of waveguide devices. He has authored or coauthored over 20 papers in refereed international technical journals and over 70 papers in international conference proceedings in his areas of research interest. His current research interests include numerical methods for the analysis of waveguide and scattering structures, automated design of waveguide components, radiating systems, measurement techniques and power effects in passive systems. In 1993, Dr. Boria received from the Spanish “Ministerio de Educación y Ciencia” the First National Prize of Telecommunication Engineering Studies for his outstanding student record. In 2001, he received from the Social Council of Universidad Politécnica de Valencia the First Research Prize for his outstanding activity during the period 1995 to 2000. Since 1992, he has been a member of the IEEE Microwave Theory and Techniques Society (IEEE MTT-S) and of the IEEE Antennas and Propagation Society (IEEE AP-S). Since 2003, he has been a member of the Technical Committee of the IEEE-MTT International Microwave Symposium and of the European Microwave Conference.

2100

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**On the Interaction Between Electric and Magnetic Currents in Stratiﬁed Media
**

Daniel Llorens del Río and Juan R. Mosig, Fellow, IEEE

Abstract—The presence of both electric and magnetic current elements embedded in stratiﬁed media is necessary to model many problems of interest in current integrated circuit and printed antenna technology. The integral equation formulation as applied to these problems is reviewed. Special attention is given to the Green’s dyadic for the electric ﬁeld generated by a magnetic current element. The fact that spectral-domain transmission line Green’s functions for a multilayered structure are closed form integrable along the axis normal to the stratiﬁcation is exploited to greatly improve the efﬁciency and the accuracy of the method. Theory and implementation are demonstrated in two practical problems: 1) arbitrarily shaped apertures in thick conducting screens, and 2) a metallic airbridge over a slot line. Index Terms—Integral equations, slot antennas, apertures in thick screens, airbridges.

Fig. 1. Aperture in thick screen: original problem. Above and below the screen lie arbitrary, laterally open multilayered media.

I. INTRODUCTION

R

ECENT progress in printed circuit technology has made pervasive the use of multiple ground-plane, multiple-via connected circuits [1]. Coplanar waveguide (CPW) has ﬁnally become the transmission line of choice for most millimeter-wave printed circuit applications [2], for example when integrated with slot antennas. In most cases, the CPW is used together with airbridges or combined with via-holes interconnects. In all these structures, the most efﬁcient approach is to model slots and apertures with equivalent magnetic currents and metallic interconnections (vias, airbridges) with equivalent electric currents. The scattered ﬁelds would then be obtained, invoking linearity, as convolutions of the equivalent currents with the adequate Green’s functions (1a) (1b) A particularly interesting situation is encountered when modeling slots in thick metallic planes. An approximate method to deal with this problem, termed the “Delta” approach, has recently been presented in [3]. In its ﬁnal formulation, this method only provides a ﬁrst order approximation to the effect of ﬁnite thickness and neglects the effect of the walls, particularly the possible coupling between the walls at both sides of the slot. Other methods [4] only consider the walls approximately. The

Manuscript received June 2, 2003; revisedAugust 22, 2003. This work was supported by the ESA/ESTEC under Contract 14062/00/NL/GD. The authors are with the Laboratory of Electromagnetics and Acoustics (LEMA), Swiss Federal Institute of Technology, EPFL-Ecublens Lausanne, CH-1015, Switzerland (e-mail: daniel.llorensdelrio@epﬂ.ch). Digital Object Identiﬁer 10.1109/TAP.2004.832329

thickness effect is particularly important in millimetric CPW circuits that use compact ﬁlters [5] where very narrow slots are common. As it will be shown in this paper, an efﬁcient and general modeling approach is to use vertical electric currents in the lateral metallic walls and horizontal magnetic currents in the apertures themselves. Therefore, we have again a mixed situation where the four Green’s functions in (1) are needed. and are usuIn the case of multilayered media, and , ally written in a mixed potential form [6]. For it is more advantageous, if electric and magnetic cells do not overlap, to keep the ﬁeld formulation. The strong singularities of these functions will be reduced to absolutely integrable singularities with a preprocessing of the vertical coordinate in the spectral domain, which was pioneered in [7] and extended in scope in [8] and [9]. This preprocessing greatly increases the efﬁciency and accuracy of the method, but requires that all unknowns be exactly perpendicular or parallel to the axis of stratiﬁcation (a “2.5D” geometry). Most printed technology products match this requirement. II. MODELING APERTURES IN THICK SCREENS To illustrate the general strategy proposed in this paper to deal interactions, we start by considering the “thick slot” with problem of Fig. 1, where a stratiﬁed medium includes a thick in which a slot has been created. This ground plane problem is usually solved by using equivalence principles to reduce it to a more tractable conﬁguration. Fig. 2 shows the standard strategy, which divides the problem in three parts linked by magnetic currents. In the layered media above and below the thick slot, standard MoM integral equation approaches may be used. The medium representing the slot is now a cavity (or a waveguide closed at both ends); here, apropiate modal expansions or cavity/waveguide Green’s functions should be used.

0018-926X/04$20.00 © 2004 IEEE

LLORENS DEL RÍO AND MOSIG: ON THE INTERACTION BETWEEN ELECTRIC AND MAGNETIC CURRENTS IN STRATIFIED MEDIA

2101

we assume that incident excitation waves exist outside the slot, on both sides. Therefore, the only unknowns will be magnetic above and below the slot, and electric current currents on the vertical walls of the slot. Accordingly, the equivalent thick slot problem of Fig. 3 yields three coupled integral equations. The ﬁrst two impose continuity of electric and magnetic tangential ﬁeld at lower and upper apertures:

Fig. 2. Equivalent problem with cavity or closed waveguide.

(2a) (2b) The third integral equation forces the electric tangential ﬁeld to vanish on electric walls (2c) In these equations, the currents in the upper and lower sides of and the superindices of the Green’s functions the slot refer to the regions of the problem as indicated (Fig. 3). This formulation can be easily generalized to a more complicated geometry. Any external media can be theoretically handled by and/or by providing appropiate Green’s functions including additional unknown currents as required. The magnetic and electric equivalent currents are then expanded as linear combinations of rooftop basis functions and and tested with a Galerkin approach. This results in a moment block matrix of the form (3)

Fig. 3. Equivalent problem with laterally open parallel plate medium and wall electric current unknowns. Numbers 1–5 refer to regions where equivalent currents J or M do exist.

This method has two main inconvenients. Firstly, the modes of the waveguide must be computed, which is difﬁcult to do for a waveguide of arbitrary section. Secondly, the modal expansion must be matched at the apertures to the multilayered media above and below, which requires the computation of extra coupling integrals. In this paper, we propose a new strategy for creating an equivalent problem. The geometry is again divided into three regions, and the media above and below are the same. However, in the medium representing the thick slot (the “internal” medium), the equivalent is now a parallel plate waveguide (PPWG). To account for the metallic walls, we add unknown vertical electric currents embedded in the PPWG. With this strategy, we need to compute interactions between horizontal magnetic currents and vertical electric currents. As this situation arises in many other conﬁgurations (for instance, airbridges over slot lines) this is the main concern of this paper. A. Formulation It is apparent that in either of these strategies, the computational burden will at least double when compared to a zero-thickness analysis, owing to the slot aperture being meshed twice. For a general solution to the problem, this is impossible to avoid. Still, a blind attack on the problem would not be excessively time-consuming for this reason, but for the complex, nonanalytical dependence of the space domain , and PPWG’s Green’s function on the three coordinates (Fig. 3). For the sake of simplicity, we shall assume that the media above and below the slot are devoid of scatterers which could support additional induced electric and magnetic currents. Also

The symmetry of , and submatrices stems, respectively, from the following reciprocity relationships between reaction terms [10] (4a) (4b) (4c) The matrix of (3) is not, however, symmetric. If, instead of a is used, strict Galerkin approach, a set of test functions a symmetric matrix results, as noticed by Harrington [6]. This is equivalent to multiplying the ﬁrst two (block) rows of system . (3) by The submatrices contain terms of the type . They are computed with a dual mixed potential expression involving potential Green’s functions , which can . It is remarkable present an integrable singularity of type in (2b) will diverge that each individual Green’s function as , so a double aperture formulation such as the one used here will not work with vanishingly small screen thickness. Even with moderately small thickness, it will be advantageous to extract not only the singular part of each reaction term

2102

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

, but also the contribution of the closest images. This is clearly a limitation of the double-aperture formulation, as has already been pointed [11]. The “Delta” formulation [3] is speciﬁcally designed to handle the limiting case. . The submatrix involves terms of the form These are computed by putting in full mixed potential form, according to the formulation in [12]. Rectangular rooftops are used to expand in an orthogonal mesh.1 The classical is used, as in [8], but the computaSommerfeld form for [13] is never necessary, because it can be substition of tuted by proper application of reciprocity relation (4c). We note that the line integrals inherent in our choice of the Sommerfeld do not appear in the particular case of a homogeform of neously ﬁlled PPWG. Otherwise, the spectral integration technique, which is detailed in the next section for the elements of , is also applied to the computation of the elements of . The integrated functions to be tabulated present at most a logarithmic singularity. B. Submatrices

correspond to spatial domain rotational symmetry (8a) (8b) (8c) Making use of the identities (14) in Appendix I, at most four Sommerfeld transform operations are needed to obtain the trans, and only one to obtain verse components (vertical magnetic currents are not considered). These operations are (9a) (9b) (9c) (9d) (9e) Therefore, the azimuthal dependence of is evaluated (inexpensively) during matrix ﬁll. All off-diagonal components are singular at the origin as , with bounded, when . This can be (with gathered from the behavior of the free space dyadic )

These contain elements of the form , which are handled with a pure ﬁeld formulation. For a generalized PPWG (which may be ﬁlled with a number of stratiﬁed dielectric layers), the spectral Green’s dyadic has the form

(5) (10) In principle, as discussed in previous sections, every Green’s , and . function will depend on three space coordinates, However, the vertical regularity of the structure makes it possible to take this dependence into account in the spectral domain. For it is well known that the existence of a TL model for the stratiﬁed medium ensures that the dependence with respect to the vertical coordinates will be of the form

The various components are obtained from the well known spectral TL model, where the stratiﬁed medium is represented by two (TE and TM) equivalent transmission line networks, and all ﬁeld quantities can be expressed in terms of voltages and or curcurrents in this network, under either voltage rent excitation [14]. We have cast them in the following form, ready for numerical implementation: (6a) (6b) (6c) (6d) (6e) (6f) (6g) whose symmetry properties (7a) (7b) (7c)

shall use the conditions that ^ ^ • e = ee(z; ) = ee (z )e (); (separability) ^ ^ z or e z (orthogonality). • e A mesh of the kind needed is pictured in Fig. 8.

1We

(11) where parameters do not depend on the vertical coordinates once source and observer layers are ﬁxed. Formulae of this type, that explicitly reveal the exponential dependence on both vertical coordinates and for any transmission line parameter ( or ) have been presented in [8] and, more recently, again in [9]. The transverse Fourier transform does not affect the vertical can be cardependence; thus, any spatial integration along ried over in the spectral domain, analytically. This effectively eliminates the spatial Green’s functions dependence on these coordinates, avoiding cumbersome three-dimensional spatial interpolation during matrix-of-moments ﬁll. In the particular case of elements of the type is ﬁxed because it corresponds to a horizontal

?

k

LLORENS DEL RÍO AND MOSIG: ON THE INTERACTION BETWEEN ELECTRIC AND MAGNETIC CURRENTS IN STRATIFIED MEDIA

2103

TABLE I GREEN’S FUNCTIONS FOR THE COMPUTATION OF ELEMENTS AND THEIR SINGULARITIES AT THE ORIGIN

OF

H

magnetic current cell. Therefore, only the -dependence has to be treated in the spectral domain. With the same requisites of orthogonality of the mesh for the vertical walls, we can write

• Integrated Green’s functions have better spectral behavior and are thus easier to evaluate in the spatial domain. • Spatial integration becomes simpler, not only because surface integrals are reduced to line integrals, but also because they exhibit weaker singularities at the origin than their nonintegrated counterparts. • Tables may be reused for differently shaped slots, as long as the vertical structure (on which spectral integration depends) remains unchanged. • Since vertical integrations are analytical, accuracy is increased. It is important to make a note about the method used to perform the Sommerfeld inversion. An elliptical integration path circumventing the spectral Green’s function singularities [16], combined with the weighted averages algorithm [17] for the tail of the integral along the real axis, is used throughout. The different Green’s functions to be tabulated are classiﬁed in vectors according to their singularity type, and transformed in block; therefore, for each of these vectors, the spectral TL model is solved only once. This classiﬁcation is necessary because the weighted averages algorithm is an extrapolation method that needs information about the asymptotic behavior to perform optimally. of the integrand as D. Examples Two different examples, a rectangular slot and a “dogbone” slot antenna, have been analyzed, built and measured. A series of breadboards has been built for each, where slot thickness varies from 35 m (printed slot, essentially a zero thickness case) to several millimeter. The rectangular slot antenna is fed by coupling to a 50 microstrip line. Its dimensions are 25 mm 5 mm and it is found mm, substrate, to resonate at 6.32 GHz on when metal thickness equals 35 m. The series of measured breadboards included screens with thicknesses of , and mm. Comparison between different numerical models and measurements is given in Figs. 4–6 for conciseness, comparison for the 3 and 6 mm cases has been omitted, but it is reported in [18]. Three different theoretical predictions are shown as follows: • the “Delta” function approach [3]; • a rigorous mode-matching cavity model [19]; • the technique described in this paper. Our technique always follows closely the cavity results (while -thick slots, and avoiding the cavity formulation) up to that both agree very well with measurements. As expected, the “Delta” approach is good only for thin slots (up to ). , It can be seen that even for slot thickness as small as the effect on resonant frequency is clearly nonnegligible (a % displacement) which stresses the interest and need of this analysis. The second model is a “dogbone”-shaped slot antenna, also fed by coupling to a 50 microstrip line (Fig. 7).

(12) The last right-hand term in the series of equalities of (12) is the key for a successful numerical integration. The new “intehas reduced-order singularities. The grated” dyadic exact type of these singularities can be obtained either from the “integrated” spatial static Green dyadic, or from the behavior of . The ﬁrst approach is more practhe spectral dyadic as tical because it allows to obtain closed analytical expressions for the singular part of the reaction term, in a manner similar to what is done in [15] for potential integrals. This development is detailed in Appendix II. The resulting singularity extraction procedure is necessary for an efﬁcient and accurate computawhen electric and magnetic cells share tion of an edge. Thanks to the spectral integration in (12), the strongest ’s elements is absolutely integrable, which singularity of makes the technique much easier to apply. With the spectral integration technique, six Sommerfeld transforms are required per vertical cell level-magnetic layer pair: four for the horizontally oriented half-rooftops and one for either increasing or decreasing vertically oriented half-rooftops. This is summarized in Table I. For example, in the mesh of Fig. 8, which has two vertical cell levels and two magnetic layers, a functions of this type must be tabulated. total of C. Discussion In addition to avoidance of three-dimensional interpolation, the spectral preprocessing technique has four advantages. These are not restricted to the interactions between current elements of different type, but in that case they are particularly important, because a ﬁeld formulation is employed.

2104

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 4.

Rectangular slot antenna: thickness

35 m =1400 at 6 GHz.

Fig. 6.

Rectangular slot antenna: thickness

10 mm =5 at 6 GHz.

Fig. 5. Rectangular slot antenna: thickness

1 mm =50 at 6 GHz.

Fig. 7. ”Dogbone” slot in thick screen, fed by microstripline. Dimensions in : ;e : ;l ;w ;l ;w : ;p millimeters: r : : . : . Substrate: h mm,

95

= 2 5 = 2 5 = 34 = 5 = 47 = 0 76 = 2 485

= 2 164 =

The mesh used in the MoM analysis is shown in Fig. 8. The freely available mesh generator, TRIANGLE,2 was used. The choice of a triangular mesh for the magnetic unknowns is justiﬁed by the fact that, opposite to the vertical walls, slot geometry can be rather complex. It is worth pointing out that a shape like the “dogbone” does not add any complexity to our approach, while the cavity approach would become very cumbersome and, indeed, it has not been considered here. mm, mm, and The breadboard series included screens mm thick. They were made of brass and gold-plated for best contact with the printed antenna substrate; their size was 14 cm 14 cm, large enough to avoid ﬁnite ground plane size inﬂuencing input impedance. Simulations and measurements are compared in Figs. 9 to 11. It can be seen that the numerical analysis follows closely

Fig. 8. Mesh for the “dogbone” model with mask of thickness mm, with two vertical electric cell levels between the parallel plates. Electric cells in white (microstripline sides of dogbone slot), magnetic cells in gray (top and bottom of dogbone).

=3

+

2http://www-2.cs.cmu.edu/~quake/triangle.html

the behavior indicated by the measurements, as thickness increases. The opening of the resonance loop in the Smith chart is characteristic.

LLORENS DEL RÍO AND MOSIG: ON THE INTERACTION BETWEEN ELECTRIC AND MAGNETIC CURRENTS IN STRATIFIED MEDIA

2105

Fig. 9. ”Dogbone” antenna: thickness 35 m

=1700 at 5 GHz.

Fig. 11.

”Dogbone” antenna: thickness 5 mm

=12 at 5 GHz.

Despite its very different nature, the IE formulation of the airbridge/slot problem is basically equivalent to that of the interaction plays an essential thick slot and again the role. Due to the presence of horizontal electric cells, a new appears, that contains interactions of submatrix type . These are computed again with a the form mixed potential formulation, where the corresponding Green’s functions have been integrated in the spectral domain along the source coordinate. In addition, line integral terms appear because of the airbridge corners [21]. A. Examples As a test case, a slot loop antenna that resonates at 3 GHz has been selected. This antenna (Fig. 12) was a model for studying the radiating element of open structure, integrated receiver front-ends for submillimeter-wave operation [22]. In that application, the diode can be connected in series to the feed line; then, the airbridge (in the symmetrical position) is used to provide a dc return path. The airbridge should not affect the RF characteristics of the slot loop, so it should be comparatively long. In the asymmetrical position, the airbridge can be used to connect to ground a diode that has been placed at the center of the loop. The role of the airbridge is wholly different because now it must conduct RF signal. The slot loop, which is about one wavelength at RF, presents a magnetic current null at that position, which allows for a very short airbridge. For the ﬁrst case, a series of airbridges of different form factors, listed in Fig. 13, was built, and antenna impedance was measured at the end of the CPW feeding line (Fig. 14). An excellent agreement is observed for all cases, as compared to simulations (Fig. 15). It can be appreciated that for the

Fig. 10.

”Dogbone” antenna: thickness 1 mm

=60 at 5 GHz.

III. AIRBRIDGE MODELING CPW circuits usually require airbridges for proper operation, to eliminate the unwanted slotline mode [2]. In many circuits, such as the dual-mode ﬁlter introduced in [20], or in MEMS delay lines, the response of the circuit is highly dependent on airbridge dimensions.

2106

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 12. Slot antenna propotype. Dimensions in millimeters: r = 15:92; w = 2; l = 19:89; w = 1:140; s = 0:930; w = 0:126; s = 2:748. Substrate: h = 1:57 mm, = 2:33.

Fig. 13.

Dimensions (in millimeters) for each case of Figs. 14, and 15.

Fig. 14. Measured jS j for the slot loop antenna of Fig. 12 with the airbridges of Fig. 13, with airbridges on the symmetrical position.

longest airbridge (case #2) the main resonance of the bare loop is least affected. The asymmetrical placement of the airbridge produces in turn a very complex behavior, with multiple resonances, due to the introduction of asymmetric modes in the loop. The main resonance of the bare loop remains unaffected however, because at that frequency the condition imposed by the airbridge matches the natural symmetric conﬁguration of magnetic currents. Again, comparison of measurement and simulation (Fig. 16) shows good agreement. IV. CONCLUSION A technique to compute interactions between electric and magnetic currents embedded in multilayered media has been described, as applied to the analysis of 1) slots in metallization screens of ﬁnite thickness, and 2) airbridges in slot/CPW circuits.

For the “thick slot” problem, this is done by solving for magnetic currents at both sides of the slot and for electric currents at the walls on its contour. The problem posed by these vertical walls inside a parallel-plate medium is ideally suited to a mixed spectral-spatial formulation. A pure mixed potential formulation has been kept for the electric elements. For the parallel plate medium, this choice leads to less integral types and consequently to reduced memory requirements and faster computations. For electric-magnetic interactions, a ﬁeld formulation is used. The singularity of the mixed-type Green’s functions is integrated over the source cell (always a magnetic current cell) with a closed formula. The method presented here has the ﬂexibility required to go beyond regular slot shapes, where the use of the cavity approach would be much more involved. This ability has been demonstrated with the analysis of a “dogbone”-shaped aperture.

2107

(14a) (14b) (14c) (14d) (14e)

APPENDIX II SINGULARITY EXTRACTION When in (12) is a triangular RWG rooftop, the inner integral in (12) is proportional to (substituting the static part of (10) ) for (15a)

Fig. 15. Computed jS j for the slot loop antenna of Fig. 12 with the airbridges of Fig. 13, with airbridges on the symmetrical position.

(15b) where (15a) is for vertically oriented and (15b) for horizontally oriented , i.e., . These integrals are both analytical. The second one results in an integral over which has the same mild-type singularity of a regular potential Green’s function [15]. For (15a), if is a constant and (16) is substituted in (15a), we obtain (17) This is absolutely integrable, as can be shown by considering (18)

Fig. 16. Computed and measured jS j for the slot loop antenna of Fig. 12 with an airbridge (case #4 of Fig. 13) on the asymmetrical position.

(because that

). Noting

Quite remarkably, the same technique has been applied to a very different problem, namely, the analysis of the effect of an airbridge short-circuiting a slot loop antenna at different positions. Again, comparison with measurements has shown the validity of the method. APPENDIX I TRANSFORM RELATIONS

(19)

(13a)

(13b)

2108

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

we may write, separating from an inﬁnitesimal region around , say, a circular sector of radius

(20) , and the integrand of The second term becomes zero as the third term is bounded, so it also vanishes as . There remains (21) which is immediate. If is a linear function of , the resulting integral can be seen to be a combination of (17) and a term of the same type as (15b). ACKNOWLEDGMENT The authors would like to thank Dr. Lin of ESTEC (Earth Observation Programs) for his support of this work. REFERENCES

[1] H. Chen, Q. Li, L. Tsang, C.-C. Huang, and V. Jandhyala, “Analysis of a large number of vias and differential signaling in multilayered structures,” IEEE Trans. Microwave Theory Tech., vol. 51, pp. 818–829, Mar. 2003. [2] R. Simons, Coplanar Waveguide Circuits, Components and Systems, 1st ed. New York: Wiley, 2001. [3] J. R. Mosig, Scattering by arbitrarily-shaped slots in thick conducting screens: An approximate solution, in IEEE Trans. on Antennas Prop.. Accepted for publication to. [4] D. T. Auckland and R. F. Harrington, “Electromagnetic transmission through a ﬁlled slit in a conducting screen of ﬁnite thickness, TE case,” IEEE Trans. Microwave Theory Tech., vol. 26, pp. 499–505, July 1978. [5] J. Sor, Y. Qian, and T. Itoh, “Miniature low-loss CPW periodic structures for ﬁlter applications,” IEEE Trans. Microwave Theory Tech., vol. 49, pp. 2336–2341, Dec. 2001. [6] R. F. Harrington, Field Computation by Moment Methods, 1st ed. New York: Macmillan, 1968. [7] M.-J. Tsai, C. Chen, N. G. Alexopoulos, and T.-S. Horng, “Multiple arbitrary shape via-hole and air-bridge transitions in multilayered structures,” IEEE Trans. Microwave Theory Tech., vol. 44, pp. 2504–2511, Dec. 1996. [8] N. Kınayman and M. I. Aksun, “Efﬁcient use of closed-form Green’s functions for the analysis of planar geometries with vertical connections,” IEEE Trans. Microwave Theory Tech., vol. 45, pp. 593–603, May 1997. [9] M. Vrancken and G. A. E. Vandenbosch, “Hybrid dyadic-mixed potential and combined spectral-space domain integral-equation analysis of quasi 3-D structures in stratiﬁed media,” IEEE Trans. Microwave Theory Tech., vol. 51, pp. 216–225, Jan. 2003. [10] V. H. Rumsey, “Reaction concept in electromagnetic theory,” Physical Rev., vol. 94, no. 6, pp. 1483–1491, June 1954. [11] R. F. Harrington and D. T. Auckland, “Electromagnetic transmission through narrow slots in thick conducting screens,” IEEE Trans. Antennas Propagat., vol. 28, pp. 616–622, Sept. 1980.

[12] J. Chen, A. A. Kishk, and A. W. Glisson, “Application of a new MPIE formulation to the analysis of a dielectric resonator embedded in a multilayered medium coupled to a microstrip circuit,” IEEE Trans. Microwave Theory Tech., vol. 49, pp. 263–279, Feb. 2001. [13] K. A. Michalski and D. Zheng, “Electromagnetic scattering and radation by surfaces of arbitrary shape in layered media, part II: Implementation and results for contiguous half-spaces,” IEEE Trans. Antennas Propagat., vol. 38, pp. 345–352, Mar. 1990. [14] K. A. Michalski and J. R. Mosig, “Multilayered media Green’s functions in integral equation formulations,” IEEE Trans. Antennas Propagat., vol. 45, pp. 508–519, Mar. 1997. [15] D. R. Wilton, S. Rao, A. W. Glisson, D. H. Schaubert, O. Al-Bundak, and C. M. Butler, “Potential integrals for uniform and linear source distributions on polygonal and polyhedral domains,” IEEE Trans. Antennas Propagat., vol. 32, pp. 276–281, Mar. 1984. [16] P. Gay-Balmaz and J. R. Mosig, “Three dimensional planar radiating structures in stratiﬁed media,” Int. J. Microwave and Millimeter-Wave CAE, vol. 3, no. 5, pp. 330–343, 1997. [17] J. R. Mosig, “Integral equation technique,” in Numerical Techniques for Microwave and Millimeter-Wave Passive Structures, 1st ed, T. Itoh, Ed. New York: Wiley, 1989, ch. 3. ´ [18] D. Llorens del Río, I. Stevanovic, and J. R. Mosig, “Analysis of printed structures including thick slots,” presented at the Proc. COST-284 Meeting, Budapest, Apr. 2003. [19] A. Álvarez Melcón, “Applications of the integral equation technique to the analysis and synthesis of multilayered printed shielded microwave ﬁlters and cavity backed antennas,” Ph.D. dissertation, Ecole Polytechnique Fédérale de Lausanne, LEMA-DE, 1998. [20] L. Cohen, H. Baudrand, D. Bajon, and J. Puech, “Full wave analysis of coplanar four-poles resonators using odd and even modes,” in Proc. Int. Workshop on Microwave Filters, Toulouse, France, June 2002. [21] T. M. Grzegorczyk and J. R. Mosig, “Line charge distributions arising in the integral equation treatment of bent scatterers in stratiﬁed media,” in Proc. Inst. Elect. Eng. Microw. Antennas Propag., vol. 148, Dec. 2001, pp. 365–368. [22] P. Otero, G. V. Eleftheriades, and J. R. Mosig, “Integrated modiﬁed rectangular loop slot antenna on substrate lenses for millimeter- and submillimeter-wave frequencies mixer applications,” IEEE Trans. Antennas Propagat., vol. 46, pp. 1489–1497, Oct. 1998.

Daniel Llorens del Río received the Electrical Engineer degree from the University of Málaga, Málaga, Spain, in 2000. He is currently working toward the Ph.D. degree at the Laboratory of Electromagnetics and Acoustics, Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland. His research interests are numerical methods and antenna modeling.

Juan R. Mosig (S’76–M’87–SM’94–F’99) was born in Cadiz, Spain. He received the Electrical Engineer degree in 1973 from the Universidad Politecnica de Madrid, Spain. In 1976, he joined the Laboratory of Electromagnetics and Acoustics (LEMA), Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, from which he obtained the Ph.D. degree in 1983. Since 1991, he has been a Professor at EPFL and since 2000, the Head of the EPFL Laboratory of Electromagnetics and Acoustics. In 1984, he was a Visiting Research Associate at Rochester Institute of Technology, Rochester, NY. He has also held scientiﬁc appointments at universities of Rennes (France), Nice (France), Technical University of Danemark and the University of Colorado at Boulder. He is the author of four chapters in books on microstrip antennas and circuits and more than 100 reviewed papers. He is co-organizer and lecturer of yearly short courses in numerical electromagnetics (Europe and USA). He is the Chairman of a European COST project on antennas and is responsible for several research projects for the European Space Agency. His research interests include electromagnetic theory, numerical methods and planar antennas. Dr. Mosig is a Member of the Swiss Federal Commission for Space Applications.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2109

**Scattering by Arbitrarily-Shaped Slots in Thick Conducting Screens: An Approximate Solution
**

Juan R. Mosig, Fellow, IEEE

Abstract—In this paper, the integral equation formulation of the thick aperture problem is reviewed and then modiﬁed to make it continuously valid for any aperture thickness. Hence, the new proposed thick aperture formulation is free from the difﬁculties usually encountered when applying it to a vanishing thickness slot. Afterwards, a simpliﬁcation of the formulation is proposed, which reduces dramatically the computational burden while providing valid results for apertures whose thicknesses remain small compared with their linear transverse dimensions but having otherwise arbitrary shapes and sizes. Preliminary numerical results conﬁrm the validity of the proposed technique and show clearly its advantages. Index Terms—Thick slots, apertures, conducting screens, integral equations, Green’s functions.

Fig. 1. General geometry for an arbitrarily shaped aperture in a conducting curved screen of variable thickness.

I. INTRODUCTION

A

CLASSICAL problem in EM-theory is the scattering of an electromagnetic wave by an aperture in a conducting thick screen. This problem has countless applications in modern technology, ranging from waveguide ﬁlters using interconnecting wall holes and irises to cavity-backed slot-fed antennas and passing through many problems of ﬁeld penetration through slits and holes, of paramount relevance in electromagnetic compatibility. In a general case (Fig. 1), the screen may be curved and have a nonzero thickness, the aperture will have arbitrary shape and dimensions and even the lateral metallic walls associated to the aperture rim may have an irregular proﬁle, thus leading to a truly three-dimensional (3-D) problem. In this work, we will concentrate in the case where the thick conducting screen is bound by two parallel surfaces and is locally ﬂat. Even with this simpliﬁcation, the problem remains 3-D and for analysis purposes, a reduction to two dimensions has been traditionally obtained in two ways. With reference to Fig. 1, either the screen thickness is neglected and then we formulate the problem in two coordinates locally tangential to the screen, or a translational symmetry along one tangential coordinate is assumed, and then we work in a 2-D cut of the problem deﬁned by its proﬁle in the coordinates. Historically, the ﬁrst model analyzed was the zero-thickness screen (frequently but improperly called the zero-thickness slot geometry). This problem can be traced back to Lord Rayleigh

Manuscript received June 23, 2003; revised September 1, 2003. This work was supported by the Swiss “Ofﬁce Fédérale de l’Education et de la Science” under Grant European COST-284 Action. The author is with the Laboratory of Electromagnetics and Acoustics (LEMA), Swiss Federal Institute of Technology, EPFL-Ecublens Lausanne, CH-1015, Switzerland (e-mail: juan.mosig@epﬂ.ch). Digital Object Identiﬁer 10.1109/TAP.2004.832325

in 1897 [1] and was extensively analyzed in a series of classical papers (see [2]–[5], to mention but a few) mainly dealing with electrically small slots. Consequently, quasistatic or low frequency approximations were used. On the other hand, speciﬁc techniques were also developed for large apertures, using geometrical [6] and spectral [7] theories of diffraction. The rigorous formulation of a zero-thickness aperture with arbitrary size and shape is made through the use of the equivalence theorem and equivalent magnetic currents. This leads to an integral equation problem solved with the use of dyadic Green’s functions [8]. This nowadays classical formulation was extensively discussed in an excellent review paper [9] and is summarized in advanced textbooks in electromagnetics [10]. The second 2-D model, assuming translational invariance and valid for long, thin apertures (slits) was solved by using asymptotic Wiener-Hopf techniques [11] or coupled integral equations [12]. These works deal essentially with thick slots having rectangular proﬁles in the plane. The integral equation approach was extended to arbitrary proﬁles [13] and was also combined with ﬁnite elements to cope with more general conﬁgurations possibly including inhomogeneous media [14], [15]. Back to the general 3-D aperture problem of Fig. 1, it can be formally solved by using equivalence principles leading to a set of coupled equations. Typically, the two outer problems (outside the thick slot) will be formulated as integral equations and the inner problem (inside the thick slot) as a cavity problem where the Helmholtz equation is to be satisﬁed. In practice, the numerical implementation will be a difﬁcult task, asking in the external regions for complicated Green’s functions and 2-D-boundary elements, which must be coupled to 3-D-ﬁnite elements inside the slot. A clever simpliﬁed implementation, based on the reciprocity principle [16] has been used to analyze microstrip antennas fed through reasonably thick rectangular slots [17]. Finally, it must be mentioned that the circular aperture case is of particular relevance in optics, and that the thick case has

0018-926X/04$20.00 © 2004 IEEE

2110

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

the theory which follows is also formally valid for the more general geometry of Fig. 1. Following the standard procedure, we replace the two openings of the thick aperture by zero-thickness conducting surfaces. and The two sides of the surface separating regions will be denoted and , while the two sides of the surface and will be denoted and . separating regions Now, according to the equivalence theorem, we deﬁne unknown equivalent magnetic surface currents in the following way [Fig. 2(a)] (1) Since surface magnetic currents are cross products of unit normal vectors and electric ﬁelds, the continuity of the tangential electric ﬁeld is automatically fulﬁlled in the interfaces between our three regions. The introduction of the conducting surfaces allows the consideration of three formally independent problems, one for each region, that are indirectly coupled through the equivalent magnetic currents. In particular the becomes a cavity fully bounded by conducting walls. region We use now the well known concept of “short-circuited exci[19], deﬁned as the ﬁelds created by tation ﬁelds” the impressed sources in the region where they exist (here ) but with the aperture opening covered by the conducting surface. With the introduction of the scattered ﬁelds existing in each region, the boundary conditions imposing the continuity of the tangential components of the total magnetic ﬁeld across the two interfaces are written as

Fig. 2. Two arbitrary regions connected through a slot on a conducting screen of ﬁnite (a) and zero (b) thickness.

been solved by Roberts in an optical context [18], emphasizing the determination of plane wave reﬂection and transmission coefﬁcients. In this paper, the integral equation formulation of the thick aperture problem is reviewed and then modiﬁed to make it continuously valid for any aperture thickness. Hence, the new proposed thick aperture formulation is free from the difﬁculties usually encountered when applying it to a vanishing thickness slot. Afterwards, a simpliﬁcation of the formulation is proposed, which reduces dramatically the computational burden while providing valid results for apertures whose thicknesses remain small compared with their linear transverse dimensions (or with the square root of their surface) but having otherwise arbitrary shapes and sizes.

(2) where, to keep the notation simple, we have avoided to show the cross product with the normal unit vector , but it is understood compofrom now on that we only consider the tangential nents of the ﬁelds. The transposition of these boundary conditions into integral equations should be straightforward. Invoking linearity and superposition, we can write the scattered ﬁelds due to any induced or equivalent source as a convolution of the source with the pertinent dyadic Green’s functions over the source’s domain of existence . For instance, the magnetic ﬁeld of a magnetic current is (3) where we have introduced the convolution notation . To develop the ﬁrst boundary condition in (2), we remark that the scat, tered magnetic ﬁeld in the region is that created by the ﬁelds are due to and to while in region . When we consider the ﬁelds at the interface, the three above mentioned currents acts through convolution with, respectively, the three Green’s functions

II. THEORETICAL BACKGROUND The procedure leading to the coupled integral equations which solve the problem of a thick slot is well known [10], [12], [14], [15]. We will brieﬂy recall it here for the sake of completeness and for introducing the notation used throughout this paper. Consider the generic problem of Fig. 2(a), in which and are originally two arbitrary inhomogeneous regions also separated by a thick conducting wall. The region (sources). A portion includes a set of impressed currents of the screen is suppressed, leaving a 3-D-hole, which deﬁnes a , connecting and [Fig. 2(a)]. As stated in new region the introduction, in most problems of practical interest the conducting screen is limited by two parallel surfaces and is locally is usually a cylindrical volume with ﬂat. Also, the region arbitrary but constant cross-section in the -plane and with its axis parallel to the screen’s normal coordinate . Nevertheless,

MOSIG: SCATTERING BY ARBITRARILY-SHAPED SLOTS IN THICK CONDUCTING SCREENS

2111

that we abridge as, respectively, , and . These Green’s functions are also “short-circuited,” i.e., they are the Green’s functions associated to the respective regions when they are isolated (decoupled) from each other by conducting zero thickness walls placed in the thick aperture surfaces. Applying an identical reasoning to the second boundary condition, we can now translate directly the set (2) into the following system of two coupled integral equations:

(4) and The system of equations (4) for the unknowns fully deﬁnes the thick slot problem. Although in theory they can be used for the general problem of Fig. 1, the calculation of Green’s functions would be too much involved. Hence, we will restrict from now on our analysis to the simpler geometries of the kind illustrated in Fig. 2(a), leaving the general problem to numerically intensive techniques like ﬁnite elements or ﬁnite differences. III. CAVITY GREEN’S FUNCTIONS AND THE ZERO THICKNESS SLOT If we start from the very beginning considering a zero thickshrinks to a null volume and we ness slot, the cavity region only need to consider two regions and separated by an and we deﬁne equivalent surface interface in whose sides and [Fig. 2(b)]. The single integral magnetic currents equation is now (5) Therefore, if we solve the system of (4) associated to the thick slot problem in the limiting case of a vanishing slot thickness , we should end up with the result , which is the solution of the integral equation (5). Unfortunately, this is not the case in practice, as the cavity Green’s functions show a divergent behavior when the cavity thickness vanishes. This fact deserves further consideration and will be investigated now. correspond to the The four cavity Green’s functions four interactions shown in Fig. 3(a). Electromagnetic reciprocity ensures that we must have (6) where the index reminds us that this Green’s functions correspond to a “mutual” interaction between the two parallel surfaces bounding the slot and hence the cavity. With this simpliﬁed notation, the set of coupled integral equations (4) can be cast into a convenient matricial form

Fig. 3. (a) Four cavity Green’s functions, (b) a generic situation, and (c) its solution by images.

if the interior of the cavity is homogeneous or it is symmetrically ﬁlled with dielectric media, because then we would have by symmetry (8) where the index reminds us that this Green’s functions correspond to a “self” interaction of one of the surfaces bounding the slot and hence the cavity with itself. correIn all cases, the four cavity Green’s functions spond all to particular cases of the situation depicted in Fig. 3(b). Formally, we can solve this problem by transforming the cavity into an inﬁnite waveguide. This is achieved by taking images of the source respect to both the lower and upper cavity walls as in Fig. 3(c). But in this situation, it is well-known that all the images will keep the sign of the original magnetic source. Green’s functions (which are of the Therefore, all the HM-type) will diverge in the limiting case, as all the images coalesce into a single source of inﬁnite intensity. This heuristic conclusion will be conﬁrmed later on by a rigorous analytic development in a more speciﬁc geometry. At this time, let us simply point out the evident consequence: in its current formulation (7), the thick aperture problem cannot be solved numerically in the limiting case of a zero thickness slot, since all the elements in the Green’s function matrix would diverge. Indeed, numerical difﬁculties should be expected when

(7) where electromagnetic reciprocity ensures the symmetry of the Green’s functions matrix. A further simpliﬁcation can be used

2112

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

trying to solve (7) for small values of the thickness , and alternate forms must be investigated to provide a smooth transition to the zero-thickness case. IV. AN ALGEBRAIC INTERLUDE To throw out some extra light in the problems revealed in the previous section, let us consider the algebraic counterpart of the integral equation system (7), namely the linear system (9) where the coefﬁcients play the role of the potentially di. The formal verging cavity Green’s functions and solution of this system is

sides of the slot). Therefore, the logical thought is to replace the original equations by their sum and difference

(15) We also have two close unknowns (the values of the magnetic currents in both sides of the slot). So, the meaningful quantities are their average and their deviation from average. Therefore, we replace also the unknowns by their half-sum and half-difference (16) with the result

(17) (10) Now, we can easily see that if under a certain condition all the coefﬁcients diverge but in such a way that the following conditions are satisﬁed: (11) then we obtain the limit solution (12) which is indeed the solution of the algebraic equation equivalent to the zero-thickness slot integral equation (5). The conclusion is that the thick slot equation (7) contains as a particular case the zero-thickness slot solution, if the cavity Green’s functions fulﬁll conditions equivalent to (11). These conditions will be checked in a coming section. But even with these conditions satisﬁed, the presence of the convolution operator prevents the use of the equation (7) in situations approaching the zero thickness case and an improved formulation of the thick slot problem must be sought after. To get some hints about what must be done, let us progress a further step in the simpliﬁcation of our problem and move from algebra to arithmetic by introducing a set of numerical values for the coefﬁcients, namely We have here ﬁnally uncovered the clue for a successful attack to problem. The combination in the ﬁrst equation of (17) includes both a small coefﬁcient and a small unknown and hence can be safely neglected. Therefore by starting with , the ﬁrst equation provides directly the initial guess for the average value . This is already an excellent estimation of the true solutions of the original system (14), namely . If we need a better estimation providing different values for the unknowns, we just replace in the second equation and obtain directly , and and . If still better accutherefore racy is needed, the cyclic iteration can be pursued indeﬁnitely. Now, coming back to formal algebra, let us symbolize our linear system (14) by the matrix equation (18) It is easy to show that replacing the original individual equations by their sum and difference, is equivalent to premultiplication by a matrix and the linear system (15) corresponds to the matrix equation (19) By the same token, replacing the original unknowns by their half-sum and half-difference can be also related to this matrix since (20) and therefore the ﬁnal transformed problem (17), easily amenable to an iterative solution, is formally given by (21) (14) A close look to this system with engineer eyes reveals two very similar equations (the information about the ﬁeld values in both But is just a scaled version of the unitary 45 rotation matrix (22)

(13) which reproduce quite faithfully the numerical conditions arising in a typical thin-slot situation. The corresponding linear system is

MOSIG: SCATTERING BY ARBITRARILY-SHAPED SLOTS IN THICK CONDUCTING SCREENS

2113

Hence, we conclude that the potentially useful transformation of our linear system is just achieved by pre- and post-multiplying by a 45 rotation matrix. V. THICK SLOT INTEGRAL EQUATIONS AND ROTATION MATRICES Let us apply to our thick slot matrix integral equation (7) the pre- and post-multiplications by the rotation matrix as indicated in (21). The ﬁnal result is given in (23), at the bottom of the page, where we have introduced the “average” and “deviation” values of the magnetic currents in the slot (24) The matrix equation (23) looks much more complicated than the original one (7) and it could be feared that we have worsened our chances. But, as in the numerical example of the previous section, the ﬁrst line in the system (23) is the clue, since none of its elements will diverge when the slot thickness vanishes, if conditions (11) are fulﬁlled. We can therefore start with the and solve the ﬁrst equation in the system assumption . It is remarkable indeed (23) to obtain a ﬁrst estimation of that if media and in both sides of the slot are identical (for instance, free space), then we have and then the ﬁrst equation in (23) becomes uncoupled, directly providing the exact value of . To clarify these ideas, let us fully develop the proposed procedure in the case of a slot ﬁlled by an homogeneous or symmetrically disposed dielectric medium, and therefore satisfying the symmetry condition (8). In this case, the notation can be greatly simpliﬁed by introducing the combinations (25) that we can call the “sigma” and “delta” cavity Green’s functions ( and ). For a vanishing thickness slot, the sigma Green’s function will diverge but the delta one will vanish. With this notation, it is a straightforward matter to show that the matrix equation (23) is equivalent to

. Therefore, the second line in (26) automatigent term is , and the ﬁrst line reduces to the cally gives the result zero-thickness slot equation. Hence, we can set up the following procedure for thin slots: Step 1) assume Step 2) solve a modiﬁed zero-thickness slot equation to obtain a ﬁrst estimate of (27) Step 3) estimate by solving the equation

Step 4) improve, if necessary, the estimation of solving

by

Step 5) go to Step 3). It is worth mentioning that all the above steps are single uncoupled integral equations. In most cases, stopping after the Step 2) will be enough to predict the ﬁrst order deviation from the zero-thickness case introduced by a reasonable slot thickness. In fact, Step 2) is identical to the zero-thickness slot integral equation (5), but with the Green’s function kernel cor. Therefore, if the “delta” cavity rected by an additive term Green’s functions could be approximated by an easily computable expression, the Step 2) would provide ﬁrst corrections for thick slots with no increase in the computational complexity. The next section proposes some reasonable expressions for the “delta” Green’s function. VI. APPROXIMATIONS FOR DELTA AND SIGMA GREEN’S FUNCTIONS First of all we move from ﬁelds to potentials and introduce the convenient formalism of the “Mixed Potential Integral Equation” [20], [21]. Until now, all the Green’s functions referred in previous section are of the HM-type (magnetic ﬁeld due to a magnetic current). Therefore any generic convolution in the previous sections can be expanded in terms of potentials

(26) This is a great improvement with respect to the original matrix equation (7)! When the slot thickness vanishes, the only diver-

(28) where and are the vector and scalar potentials associated and is the equivalent with transverse magnetic currents magnetic charge.

(23)

2114

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

As it is well known, in free space we have for the mixed potential Green’s functions the values (29) with the free space scalar Green’s function given by (30) The question is how to compute these quantities in the cavity geometry. The problem is not trivial and will depend obviously on the cavity’s shape and on the medium ﬁlling it. In general, for arbitrary shaped slots, the answer can be obtained only by intensive numerical procedures. But we may try to introduce a powerful approximation, which should lead to reasonable results if the slot’s transverse dimensions aren’t smaller than the slot thickness: we just neglect the lateral conducting walls of the cavity. Although the validity of this assumption can only be judged a posteriori, its appeal is enormous. First, the cavity delta and sigma Green’s functions will have “universal” expressions independent of the slot/cavity shape. And secondly, these expressions will be reasonably simple. Fig. 4 shows the parallel plate waveguide conﬁguration which remains when we neglect the lateral walls. In this case, relations (29) are still valid, but the scalar Green’s function is no longer the free-space one (30). Its calculation is easily performed in the spectral domain [20]. For a source located on the lower wall we get the result

Fig. 4. Approximating arbitrarily shaped cylindrical cavities by a parallel plate waveguide.

Fig. 5. Modulus of the normalized delta Green’s function for several slot thicknesses: t = =1000 (dashed line), t = =100 (dotted line), and t = =10 (dash-dotted line). The straight solid line is the free space Green’s function.

(31) A partial check of the above result is provided by the fact that if we let go to inﬁnity (the parallel plate waveguide reduces , which is to its lower plate), we obtain the expected result, twice the free space value. Keeping now ﬁnite and particularizing to the values and , we get the potential versions of our cavity “self” and “mutual” Green’s functions (6), (8) and making sums and differences with them we get the potential versions of our cavity “sigma” and “delta” Green’s functions (25) (32) We have here a clear conﬁrmation of our theoretical predictions. While the “self,” “mutual,” and “sigma” cavity Green’s functions diverge for a vanishing slot thickness , the “delta” function goes to zero. Moreover, it is straightforward to show that these Green’s functions fulﬁll the conditions equivalent to (11). Moving from the spectral domain to the space domain, we can write the “delta” potential Green’s function as a Sommerfeld integral (33)

where is the radial source-observer distance. A series expansion of the hyperbolic tangent in the above equation will result in a series expression for the delta Green’s function. The amazing result is that the delta Green’s function can be expressed as an alternating-sign inﬁnite series identical to the scalar potential of an electric point charge when both source and observer are in the mid-plane of the parallel plate waveguide. To obtain speciﬁc information about the near ﬁeld (quasistatic) behavior, we look . Since in this at the asymptotical spectral behavior for case the hyperbolic tangent becomes unity, the delta Green’s function corresponds in the near ﬁeld to twice the free space Green’s function . This behavior is conﬁrmed by the numerical evaluation of the Sommerfeld integral (33) using well tested algorithms [22], [23]. Fig. 5 shows the normalized potential delta Green’s function for three slot thicknesses of 0.001, 0.01, and 0.1 free space wavelengths. It is evident at a glance how in the near ﬁeld behaves as , since the diagonal line in Fig. 5 is . As a rule of thumb, we could infer from Fig. 5 that the delta Green’s functions remain close to twice the free space Green’s function while the radial distance is smaller than the ). But for greater radial distances, slot thickness (say the values of the delta Green’s function decay very fast and it should be possible to neglect it. To put these results in perspective, let us consider a slot in a thick conducting screen separating two semi-inﬁnite free

MOSIG: SCATTERING BY ARBITRARILY-SHAPED SLOTS IN THICK CONDUCTING SCREENS

2115

Fig. 6. Thick slot of thickness =10 illuminated by normally incident plane wave having its electric ﬁeld along the y -coordinate.

2

spaces. The equation to be solved is (27) or rather its mixed potential MPIE form. Hence, applying (28) to (27) we will for get for instance a combination the scalar potential. We can easily demonstrate using image and , theory that the potential Green’s functions associated with the seminﬁnite media, are both given by twice the free space Green’s function . Therefore, is just . But the correction term also behaves in the near ﬁeld as and therefore the total kernel is expected to . It could be objected have a quasistatic behavior of type that an additive “correction” identical to the corrected term shouldn’t be called a correction, being much more than this. But this is only the limiting near-ﬁeld situation, when the source-observer distance is smaller than the slot thickness. For larger radial distances, the delta Green’s function decays very fast (Fig. 5) and so does its “correcting” effect.

Fig. 7. Normalized x-component of magnetic current along the line y = =2 aperture of thickness =10. Normal incidence plane over a square wave illumination. —M ; —M ; —M ; —zero-thickness slot. Solid line—real part, dashed line—imaginary part.

2

5

3

VII. PRELIMINARY RESULTS To check the validity of our assumptions and of our proposed equations, several very simple numerical experiments have been performed on a rather thick square slot (transverse dimensions and thickness ) (Fig. 6). The slot has been made in a screen separating two semiinﬁnite free space regions and it is excited with a normally incident plane wave having its electric ﬁeld along the -coordinate. The main and more interesting component of the magnetic current is then along . We have considered this component along the two medians of the square and the transverse one slot, a “longitudinal” one (Fig. 6). The problem has been ﬁrst solved with a rigorous treatment, where the set of equations (4) is used, together with exact expressions for the Green’s functions in the cavity. This “full wave cavity” model gives then the most accuand in both sides of the rate expressions for the currents slot, represented by circles and squares in Figs. 7 and 8. They slot. But it must show the expected behavior from a be pointed out that the full wave cavity approach is a very time consuming method, mainly due to the bad convergence of cavity Green’s functions and their lack of translational symmetry. And the situation will be much worse, not to say untractable, for an arbitrarily shaped slot. Even disregarding the cavity problem,

Fig. 8. Normalized x-component of magnetic current along the line x = =2 over a square aperture of thickness =10. Normal incidence plane wave illumination. —M ; —M ; —M ; —zero-thickness slot. Solid line—real part, dashed line—imaginary part.

2

5

3

we should expect an important slowdown with respect to the zero thickness case, since we have twice more unknowns. The snag with the zero-thickness formulation (5) is that it gives unsatisfactory results, since we get a unique current (stars in Figs. 7–8) that only matches the true values in one side of the aperture (in this case, the excitation side). Using our cor, which rected equation (27), we obtain a ﬁrst estimation for happens to be an almost perfect average value (diamonds in Figs. 7 and 8). This clearly indicates that the iterative process of Section V will converge very quickly. Results of these iterations will be reported in a coming paper. Here, we will rather explore how good are the results obtained with (27), that doesn’t introduce any numerical overload with respect to the zero thickness case. To this end, we have consquare slot, thick, excited this sidered the same

2116

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 9. Radiation pattern in E -plane. Full-wave cavity approach (solid line), “Delta approach” (dashed line), Zero-thickness approximation (dash-dotted line). aperture of =10 thickness. Plane wave impinging from bottom with 45 incidence and H -ﬁeld polarized in x-direction. Square

2

time with a plane wave incident at an angle of 45 . Fig. 9 shows the scattered ﬁeld radiation pattern obtained with the rigorous “full-wave cavity model,” assuming a zero-thickness slot and with our “delta approach.” Even without iterations, the delta approach already provides a much better result than the zerothickness approach and with no increase in computational complexity. Moreover, accuracy can be increased if needed. This would be essential when looking for precise predictions of near ﬁeld quantities. VIII. CONCLUSION In this paper, we present a rigorous integral equation formulation of the thick aperture problem providing a smooth transition to the zero-thickness case, inspired by an analogy with an algebraic problem. The full usefulness of the new formulation is only evident if the cavity Green’s functions can be easily calculated or at least efﬁciently approximated. In this paper, we propose to use as starting point the zero-thickness case. Consequently, a logical approximation is to neglect the internal lateral walls of the slot and to assume that the equivalent cavity is a parallel plate waveguide. The ﬁnal result is a new integral equation whose unknown is the average value of the magnetic currents in both sides of the thick slot. And this new equation has exactly the same degree of complexity as the zero thickness slot equation, since the only modiﬁcation is the addition of a correcting “delta” term for the Green’s functions, that can be analytically approximated and that disappears naturally in the zero-thickness limiting case. It is hard to obtain an a priori estimate of the accuracy of this approximate technique. It should logically work better for large shallow slots and deteriorate as the slot thickness increases. But the analysis of rather electrically thick slots should still lead to reasonable results if the slot’s transverse dimensions remain large compared with its thickness. In any case, thicknesses of the order of a tenth of wavelength should not pose any real challenge as the results in this paper demonstrate for the scattered ﬁeld. If the knowledge of the average value of the currents is not sufﬁcient, the technique can always be improved by looping across an iterative process, which provides improved

values of the half-difference and half-sum of the magnetic currents in both sides of the slots. Also, other approximations of the cavity Green’s functions could be explored, like the use of images respect to the lateral walls or a modal waveguide expansion, that should be excellent for very deep and narrow slots. The formulation presented is this paper is very ﬂexible and combines naturally well with the integral equation based models currently used for cavity backed antennas, thick irises in waveguide ﬁlters, slot-fed patches and thick coplanar lines. These conﬁgurations and many related ones are of paramount relevance in innovative and emerging applications, where conducting wall thickness cannot be any more neglected, because of the technology (self-supporting metallic plates rather than printed sheets), the frequency (mm- and sub mm-waves) or both. An intensive numerical exploration of these geometries, including predictions of very sensitive near-ﬁeld quantities like multiport scattering parameters, should provide a more detailed appraisal of the scope of this theory and of its accuracy. ACKNOWLEDGMENT Thanks are given to Dr. E. Suter, McKinsey Consultants, ´ Geneva, Switzerland, and to Mr. I. Stevanovic, LEMA-EPFL, Switzerland, for helpful discussions and numerical checking of the ideas developed in this paper. REFERENCES

[1] L. Rayleigh, “On the incidence of aerial and electric waves upon small obstacles in the form of ellipsoids or elliptic cylinders, and on the passage of electric waves through a circular aperture in a conducting screen,” Phil. Mag., vol. 44, pp. 28–52, July 1897. [2] H. A. Bethe, “Theory of diffraction by small holes,” Phys. Rev., vol. 66, pp. 163–182, Oct. 1944. [3] H. Levine and J. Schwinger, “On the theory of difraction by an aperture in an inﬁnite plane screen I,” Phys. Rev., vol. 74, pp. 958–974, 1948. [4] , “On the theory of difraction by an aperture in an inﬁnite plane screen II,” Phys. Rev., vol. 75, pp. 1423–1432, 1949. [5] Y. Rahmat-Samii and R. Mittra, “Electromagnetic coupling through small apertures in a conducting screen,” IEEE Trans. Antennas Propagat., vol. AP-25, pp. 180–187, Mar. 1977. [6] J. B. Keller, “Geometrical theory of diffraction,” J. Appl. Phys., vol. 28, pp. 426–444, 1962. [7] R. Mittra, Y. Rahmat-Samii, and W. Ko, “Spectral theory of diffraction,” Appl. Phys., vol. 10, pp. 1–13, 1976.

MOSIG: SCATTERING BY ARBITRARILY-SHAPED SLOTS IN THICK CONDUCTING SCREENS

2117

[8] C. T. Tai, Generalized Vector and Dyadic Analysis. New York: IEEE Press, 1992. [9] C. M. Butler, Y. Rahmat-Samii, and R. Mittra, “Electromagnetic penetration through apertures in conducting surfaces,” IEEE Trans. Antennas Propagat., vol. AP-26, pp. 82–93, Jan. 1978. [10] C. Balanis, Advanced Engineering Electromagnetics. New York: Wiley, 1989. [11] S. C. Kashyap and M. A. K. Hamid, “Diffraction characteristics of a slit in a thick conducting screen,” IEEE Trans. Antennas Propagat., vol. AP-19, pp. 499–507, July 1971. [12] F. L. Neerhoff and G. Mur, “Diffraction of a plane electromagnetic wave by a slit in a thick screen placed between two different media,” Appl. Sci. Res., vol. 28, pp. 73–88, July 1973. [13] D. T. Auckland and R. F. Harrington, “A nonmodal formulation for electromagnetic transmission through a ﬁlled slot of arbitrary cross-section in a thick conducting screen,” IEEE Trans. Microwave Theory Tech., vol. MTT-28, pp. 548–555, June 1980. [14] J. Jin and J. Volakis, “TM scattering by an inhomogeneously ﬁlled aperture in a thick conducting plane,” in Proc. Inst. Elect. Eng., vol. 137, June 1990, pp. 153–159. [15] S. Gedney and R. Mittra, “Electromagnetic transmission through inhomogeneously ﬁlled slots in a thick conducting plane—Arbitrary incidence,” IEEE Trans. Electromagn. Compat., vol. 34, pp. 404–415, Nov. 1992. [16] D. M. Pozar, “A microstrip antenna aperture coupled to a microstrip line,” Electron. Lett., vol. 21, pp. 49–50, 1985. [17] P. Haddad and D. Pozar, “Characterization of aperture coupled microstrip patch antenna with thick ground plane,” Electron. Lett., vol. 30, pp. 1106–1107, July 1994. [18] A. Roberts, “Electromagnetic theory of diffraction by a circular aperture in a thick, perfectly conducting screen,” J. Opt. Soc. Amer. A, Opt. Image Sci., vol. 4, pp. 1970–1983, Oct. 1987. [19] J. Van Bladel, Electromagnetic Fields. New York: McGraw-Hill, 1964. [20] J. R. Mosig, “Integral equation technique,” in Numerical Techniques for Microwave and Milimeter Wave Passive Structures. New York: Wiley, 1989, ch. 3.

[21] K. Michalski and J. R. Mosig, “Multilayered media Green’s functions in integral equation formulations,” IEEE Trans. Antennas Propagat., vol. 45, pp. 508–519, Mar. 1997. [22] J. R. Mosig, R. C. Hall, and F. E. Gardiol, “Numerical analysis of microstrip patch antennas,” in Handbook of Microstrip Antennas, London, U.K.: Peregrinus, 1989, ch. 8. [23] K. Michalski, “Extrapolation methods for sommerfeld integral tails,” IEEE Trans. Antennas Propagat., vol. 46, pp. 1405–1418, Oct. 1998.

Juan R. Mosig (S’76–M’87–SM’94–F’99) was born in Cadiz, Spain. He received the Electrical Engineer degree in 1973 from the Universidad Politecnica de Madrid, Spain. In 1976, he joined the Laboratory of Electromagnetics and Acoustics (LEMA), Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, from which he obtained the Ph.D. degree in 1983. Since 1991, he has been a Professor at EPFL and since 2000, the Head of the EPFL Laboratory of Electromagnetics and Acoustics. In 1984, he was a Visiting Research Associate at Rochester Institute of Technology, Rochester, NY. He has also held scientiﬁc appointments at universities of Rennes (France), Nice (France), Technical University of Danemark and the University of Colorado at Boulder. He is the author of four chapters in books on microstrip antennas and circuits and more than 100 reviewed papers. He is co-organizer and lecturer of yearly short courses in numerical electromagnetics (Europe and USA). He is the Chairman of a European COST project on antennas and is responsible for several research projects for the European Space Agency. His research interests include electromagnetic theory, numerical methods and planar antennas. Dr. Mosig is a Member of the Swiss Federal Commission for Space Applications.

2118

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Double Higher Order Method of Moments for Surface Integral Equation Modeling of Metallic and Dielectric Antennas and Scatterers

´ Miroslav Djordjevic, Member, IEEE, and Branislav M. Notaroˇ, Senior Member, IEEE s

Abstract—A novel double higher order Galerkin-type method of moments based on higher order geometrical modeling and higher order current modeling is proposed for surface integral equation analysis of combined metallic and dielectric antennas and scatterers of arbitrary shapes. The technique employs generalized curvilinear quadrilaterals of arbitrary geometrical orders for the approximation of geometry (metallic and dielectric surfaces) and hierarchical divergence-conforming polynomial vector basis functions of arbitrary orders for the approximation of electric and magnetic surface currents within the elements. The geometrical orders and current-approximation orders of the elements are entirely independent from each other, and can be combined independently for the best overall performance of the method in different applications. The results obtained by the higher order technique are validated against the analytical solutions and the numerical results obtained by low-order moment-method techniques from literature. The examples show excellent accuracy, ﬂexibility, and efﬁciency of the new technique at modeling of both current variation and curvature, and demonstrate advantages of large-domain models using curved quadrilaterals of high geometrical orders with basis functions of high current-approximation orders over commonly used small-domain models and low-order techniques. The reduction in the number of unknowns is by an order of magnitude when compared to low-order solutions. Index Terms—Electromagnetic analysis, electromagnetic scattering, higher order modeling, integral equations, method of moments (MoM).

I. INTRODUCTION

A

NTENNAS involved in modern wireless systems are often composed of metallic and dielectric parts of arbitrary shapes and with arbitrary curvature. There is a clear need for advanced analysis and design tools for predicting the performance and optimizing the parameters of such antennas prior to costly prototype development. These tools have to be based on general computational electromagnetic methods for modeling of arbitrary three-dimensional (3-D) combined metallic and dielectric structures. In addition, antenna designers demand that the simulation tools be accurate, fast, reliable, and run on relatively small computing platforms, such as standard desktop PCs. One of the most general approaches to the analysis of metallic and dielectric structures is the surface integral equation (SIE)

Manuscript received February 14, 2003; revised August 4, 2003. This work was supported by the National Science Foundation under Grant ECS-0115756. The authors are with the Department of Electrical and Computer Engineering, University of Massachusetts Dartmouth, Dartmouth, MA 02747-2300 USA (e-mail: miroslav@ieee.org; bnotaros@umassd.edu). Digital Object Identiﬁer 10.1109/TAP.2004.833175

approach, where both electric and magnetic surface currents are introduced over boundary surfaces between homogeneous parts of the structure, and surface integral equations based on boundary conditions for both electric and magnetic ﬁeld intensity vectors are solved with current densities as unknowns. The SIEs are discretized by the method of moments (MoM) [1], which gives rise to MoM-SIE modeling techniques [2]–[4]. Overall, the MoM-SIE method is an extremely powerful and versatile numerical methodology for electromagnetic-ﬁeld simulation in antenna and scattering applications that involve perfectly conducting and penetrable (dielectric and linear magnetic) materials. However, practically all the existing 3-D MoM-SIE simulation tools for metallic/dielectric structures are low-order or small-domain (subdomain) techniques—the structure is modeled by surface geometrical elements (boundary elements) that are electrically very small and the electric and magnetic currents over the elements are approximated by low-order (zeroth-order and ﬁrst-order) basis functions. More precisely, the boundary in each dimension, elements (patches) are on the order of being the wavelength in the medium. This results in a very large number of unknowns (unknown current-distribution coefﬁcients) needed to obtain results of satisfactory accuracy, with all the associated problems and enormous requirements in computational resources. In addition, commonly used boundary elements are in the form of ﬂat triangular and quadrilateral patches, and thus they do not provide enough ﬂexibility and efﬁciency in modeling of structures with pronounced curvature. An alternative which can greatly reduce the number of unknowns for a given problem and enhance further the accuracy and efﬁciency of the MoM-SIE analysis in antenna/scattering applications is the higher order or large-domain computational approach. According to this approach, a structure is approximated by a number of as large as possible geometrical elements, and the approximation of current components within individual elements is in the form of a single (two-fold) functional series of sufﬁciently high order. Only relatively recently the computational electromagnetics (CEM) community has started to extensively investigate and employ higher order surface and volume elements and higher order basis functions in the frame of MoM, including both the SIE formulation [5]–[9] and volume integral equation (VIE) formulation [10]–[15], and the ﬁnite element method (FEM) [6], [16]–[20]. For MoM-SIE modeling of general structures that may possess arbitrary curvature, it is essential to have both higher order geometrical ﬂexibility for curvature modeling and higher order

0018-926X/04$20.00 © 2004 IEEE

´ ˇ DJORDJEVIC AND NOTAROS: DOUBLE HIGHER ORDER MoM FOR SIE MODELING

2119

current-approximation ﬂexibility for current modeling in the same method. In other words, if higher order (large-domain) basis functions for currents are used on ﬂat patches, many small patches may be required for the geometrical precision of the model, and then higher order basis functions actually reduce to low-order functions (on small patches). On the other hand, geometrical ﬂexibility of curved patches can be fully exploited only if they can be made electrically large, which implies the use of higher order current expansions within the elements as well. Finally, in order to make the modeling of realistic structures optimal, it is convenient to have elements of different orders and sizes combined together in the same model. If all of these requirements are to be satisﬁed, implementation of hierarchical-type higher order polynomial basis functions for the approximation of electric and magnetic surface currents over curved boundary elements seems to be the right choice. This paper proposes a novel higher order (large-domain) PC-oriented Galerkin-type MoM-SIE technique for 3-D electromagnetics based on higher order geometrical modeling and higher order current modeling, which we refer to as a double higher order method. The surface elements proposed for the approximation of geometry (metallic and dielectric surfaces) are generalized curvilinear quadrilaterals of arbitrary geometrical orders. The basis functions proposed for the approximation of currents within the elements are hierarchical divergence-conforming polynomial vector basis functions of arbitrary orders. The proposed technique represents a generalization of the MoM-SIE technique [9], where bilinear quadrilateral surface elements (boundary elements of the ﬁrst geometrical order) are used with higher order polynomial current expansions. The new method enables excellent curvature modeling (e.g., a sphere is practically perfectly modeled by only six curved quadrilateral boundary elements of the fourth geometrical order) and excellent current-distribution modeling (e.g., using the eighth-order polynomial current-approximation in the two parametric coordinates on a quadrilateral boundary element). This enables using large curved MoM quadrilaterals that are on ) in each dimension as building the order of (e.g., blocks for modeling of the electromagnetic structure (i.e., the boundary elements can be by two orders of magnitude larger in area than traditional low-order boundary elements). Element orders in the model, however, can also be low, so that the lower order modeling approach is actually included in the higher order modeling. The geometrical orders and current-approximation orders of the elements are entirely independent from each other, and the two sets of parameters of the double higher order model can be combined independently for the best overall performance of the method. Because the proposed basis functions are hierarchical, a whole spectrum of element sizes with the corresponding current-approximation orders can be used at the same time in a single simulation model of a complex structure. Additionally, each individual element can have drastically different edge lengths, enabling a whole range of “regular” and “irregular” element shapes (e.g., square-shaped, rectangular, strip-like, trapezoidal, triangle-like, etc.) to be used in a simulation model as well. Some preliminary results

of double-higher order MoM modeling of purely metallic structures are presented in [21]. This paper is organized as follows. Section II presents the mathematical development of the proposed boundary elements and describes numerical components of the new double higher order MoM-SIE technique. This includes the derivation of surface integral equations for electric and magnetic surface current density vectors as unknown quantities, development of generalized Galerkin impedances (the system matrix elements) for arbitrary boundary elements (i.e., for any choice of surface elements for geometrical modeling and any choice of divergence-conforming basis functions for current modeling), generation of generalized curvilinear quadrilateral elements for higher order modeling of geometry, implementation of hierarchical polynomial vector basis functions for higher order modeling of currents over the quadrilaterals, and evaluation of generalized Galerkin impedances for the new proposed double higher order quadrilateral elements. In Section III, the accuracy, convergence, and efﬁciency of the new MoM-SIE technique are evaluated and discussed in several characteristic examples. The results obtained by the higher order MoM are compared with the analytical solutions and the numerical results obtained by loworder MoM techniques from literature. Numerical examples include a dihedral corner reﬂector, a metallic spherical scatterer (analyzed using six different higher order models), a dielectric spherical scatterer (analyzed using ﬁve different higher order models), and a circular cylinder of ﬁnite length with attached wire monopoles. The examples show excellent ﬂexibility and efﬁciency of the new technique at modeling of both current variation and curvature, and demonstrate its advantages over low-order MoM techniques.

II. NOVEL DOUBLE HIGHER ORDER MOM FOR ELECTROMAGNETIC MODELING A. Surface Integral Equation Formulation Consider an electromagnetic system consisting of arbitrarily shaped metallic and dielectric bodies. Let the system be excited by a time-harmonic electromagnetic ﬁeld of complex ﬁeld-inand , and angular frequency . This ﬁeld may tensities be a combination of incident plane waves or the impressed ﬁeld of one or more concentrated generators. According to the surface equivalence principle (generalized Huygens’ principle), we can break the entire system into subsystems, each representing one of the dielectric regions (domains), together with the belonging metallic surfaces, with the remaining space being ﬁlled with the same medium. One of the domains is the external space surrounding the structure. The scattered electric and magnetic ﬁelds, and , in each domain can be expressed in terms of the equivalent (artiﬁcial) surface electric current, of density , and equivalent (artiﬁcial) surface magnetic currents, of density , which are placed on the boundary surface of the domain, with the objective to produce a zero total ﬁeld in the surrounding space. On the metallic surfaces, only the surface electric currents exist (these are actual currents) and .

2120

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

The boundary conditions for the tangential components of the total (incident plus scattered) electric and magnetic ﬁeld vectors on the boundary surface between any two adjacent dielectric domains (domains 1 and 2) yield

B. Generalized Galerkin Impedances for Arbitrary Surface Elements Assume ﬁrst that all the surfaces (metallic and dielectric) in the system are approximated by a number of arbitrary surface elements. Let us approximate the surface electric and magand , over every element netic current density vectors, in the model by a convenient set of basis functions with unknown complex current-distribution coefﬁcients. In order to determine these coefﬁcients, the EFIE/MFIE system in (1)–(3) is tested by means of the Galerkin method, i.e., using the same functions used for current expansion. The four types of generalized Galerkin impedances (the system matrix elements) corresponding to the four combinations of electric- and magnetic-curand deﬁned on the th surface rent testing functions and the electric- and magnetic-current basis funcelement and deﬁned on the th element in the model are tions given by (15) (16) (17) (18) The generalized voltages (the excitation column-matrix elements) are evaluated as (19) (20)

(1) (2) where we assume that the incident (impressed) ﬁeld is present only in domain 1. On the conducting bodies, the boundary cononly, so for metallic ditions (1) and (2) reduce to surfaces in domain 1 we have (3) The scattered electric ﬁeld in the region of complex permittivity and complex permeability is expressed in terms of the electric and magnetic current densities as follows: (4) (5) (6) while the scattered magnetic ﬁeld is obtained as (7) (8) (9) In the above expressions, and are the magnetic and electric vector potentials, and and are the electric and magnetic scalar potentials, respectively. The potentials are given by (10) (11) (12) (13) where is the boundary surface of the considered domain, and the Green’s function for the unbounded homogeneous medium of parameters and (14) being the propagation coefﬁcient in the medium and the distance of the ﬁeld point from the source point. Having in mind the integral expressions for ﬁelds and in (4)–13, (1)–(3) represent a set of coupled electric/magnetic ﬁeld and as unknowns, integral equations (EFIE/MFIE) for which can be discretized and solved using the MoM.

, and apSubstituting (5) into (15), expanding plying the surface divergence theorem leads to the following expressions for electric/electric Galerkin impedances:

(21) is the outward normal to the boundary contour where of the surface . When the divergence-conforming current expansion on boundary elements is used, the last term in (21) is identically equal to zero, because the normal components of are either zero at the element edges or the testing functions two contributions of the elements sharing an edge exactly cancel out in the ﬁnal expressions for generalized impedances. Finally,

´ ˇ DJORDJEVIC AND NOTAROS: DOUBLE HIGHER ORDER MoM FOR SIE MODELING

2121

expressing the potentials in (21) in terms of the electric-current over the th surface element , we obtain basis function

(22)

Similarly, starting with (6) and (11), expanding , and performing a cyclic permutation of the scalar triple product, the expression for electric/magnetic generalized impedances in (16) can be transformed to read

**Generalized parametric quadrilateral of geometrical orders K and 1). M = (K + 1)(K + 1) is the total number of interpolation points.
**

K

Fig. 1.

(K

;K

the same within an element). The quadrilateral can be described analytically as

(23)

By duality, the magnetic/electric and magnetic/magnetic generalized Galerkin impedances in (17) and (18) have the same respective forms as those in (23) and (22), and are given by

(24)

(25)

Equations (22)–(25) provide general expressions for MoM generalized impedances for solving the EFIE/MFIE in (1)–(3) using any type of surface discretization and any adopted set of divergence-conforming basis functions in the context of the Galerkin method. In what follows, we shall restrict our attention to the speciﬁc higher order MoM technique proposed for analysis of electromagnetic radiation and scattering in this paper. C. Higher Order Geometrical Modeling As basic building blocks for geometrical modeling of 3-D electromagnetic structures of arbitrary shape and material composition, we propose generalized curved parametric quadrilaterals of higher (theoretically arbitrary) geometrical orders (Fig. 1). A generalized quadrilateral is determined by points (interpolation nodes) arbitrarily and ( , ) are positioned in space, where geometrical orders of the element along - and - parametric coordinates, respectively (note that the orders do not need to be

(26) are the position vectors of the interpowhere are Lagrange-type interpolation polynolation nodes, mials satisfying the Kronecker delta relation , and representing the parametric coordinates of the with th node, and are constant vector coefﬁcients related to . For more details on geometrical properties of parametric elements (in the context of FEM) the reader is referred to [22], [23]. In this paper, we use the equidistant distribution of interpolation nodes along each coordinate in the parametric space, while the use of speciﬁc nonequidistant node distributions, which would provide additional modeling ﬂexibility and accuracy in some applications, is possible as well. In addition, any other choice of higher order surface expansions for geometrical modeling that can be represented as a double sum of 2-D power (e.g., parametric quadrilaterals using spline functions functions for describing the geometry) can also readily be implemented in our method for electromagnetic analysis. Note that, in general, the surface tangent is discontinuous on the boundary of two attached curved generalized parametric quadrilateral elements deﬁned by (26), regardless of the geoand of the quadrilaterals. However, this metrical orders geometrical discontinuity across the boundaries of adjacent elements becomes less pronounced as the elements of higher geometrical orders are used. For instance, when approximating a circular cylinder using 32 interpolation points along its circumference and three different geometrical models constructed elements; (B) 16 from: (A) 32 ﬁrst-order second-order elements; and (C) eight fourthelements per cylinder circumference, order the angles between the surface tangents of the neighboring elements at the junctions in models (A), (B), and (C) are 168.750 , 179.787 , and 180.011 , respectively, compared to the exact 180 . If a more accurate model is needed, one can increase the total number and/or geometrical orders of patches. Note also that this geometrical problem is not present if the geometry is

2122

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

described in terms of spline functions, which can provide continuous surface tangents across the edges shared by curved elements (e.g., third-order splines used to solve scattering from bodies of revolution in [24]). All the geometries considered as examples in this paper are modeled using specialized geometrical preprocessor codes, and no general meshers are employed. Development and discussions of general geometrical preprocessors for mesh generation for an arbitrary geometry using higher order surface elements is beyond the scope of this paper. D. Higher Order Basis Functions for Electric and Magnetic Currents Electric and magnetic surface current density vectors over every generalized quadrilateral in the model are represented as

to be a preferable choice for modeling of surface currents in all applications. It enables considerable reductions in the overall number of unknowns, at no expense in terms of the accuracy of current and charge modeling over surfaces. An excellent theoretical elaboration of this approach (in the context of FEM) can be found in [25]. E. Generalized Galerkin Impedances for Higher Order Quadrilateral Elements and in (27) and (28) are The unknown coefﬁcients determined by solving the EFIE/MFIE system with the generalized Galerkin impedances given in (22)–(25), which we now specialize for the implementation of generalized curved quadrilateral elements of arbitrary geometrical orders, (26), and hierarchical divergence-conforming polynomial vector basis functions of arbitrary current-approximation orders, (29). Without the loss of generality, we consider only the -components of basis and testing functions. Furthermore, we consider the functions in the following simpliﬁed form: (32) where are the simple 2-D power functions (33)

(27)

(28) are divergence-conforming hierarchical-type vector where basis functions deﬁned as

, even , odd (29) Parameters and are the adopted degrees of the polynomial current approximation, which are entirely independent and ), and , from the element geometrical orders ( , , and are unknown current-distribution coefﬁand in (29) are obtained as cients. The unitary vectors (30) with given in (26), and formation is the Jacobian of the covariant trans(31) Note, that the sum limits in (27) and (28) that correspond to the variations of a current density vector component in the direction across that component are by one smaller than the orders corresponding to the variations in the other parametric coordinate. This mixed-order arrangement, which ensures equal approximation orders for surface charge densities corresponding to the - and -directed current basis functions, has been found

The generalized Galerkin impedances corresponding to the complete basis functions in (29) can be obtained as a linear combination of those corresponding to the simpliﬁed functions in (32) and (33). In addition, the impedances for any higher order set of basis functions of divergence-conforming polynomial type can also be constructed as a linear combination of the impedances for the simple power functions in (32) and (33). A notable example may be higher order hierarchical basis functions with improved orthogonality properties constructed from ultraspherical and Chebyshev polynomials [26], [27] (note that the technique presented in [26], [27] is restricted to bilinear ) only, as well as quadrilaterals (elements with that these basis functions, being more complicated than the regular polynomials, require larger MoM matrix ﬁlling times, and are therefore impractical when iterative solvers are not used). Upon substituting (32) into (22), the electric/electric impedances corresponding to the testing function deﬁned by indexes and on the th quadrilateral and the basis function deﬁned by indexes and on the th quadrilateral become

(34)

´ ˇ DJORDJEVIC AND NOTAROS: DOUBLE HIGHER ORDER MoM FOR SIE MODELING

2123

where and are the current-approximation orders of the th quadrilateral along the - and -coordinate, respectively, and are the corresponding orders for the th quadrilateral, and the integration limits in both quadrilaterals are and . The source-to-ﬁeld distance is computed as (35) Taking into account the parametric representation of the quadrilateral surface element, (26), then leads to the ﬁnal expression: where is the basic Galerkin ﬁeld integral evaluated as (39)

(40) Note that only two types of scalar basic Galerkin integrals, and in (37) and (40), are needed for the entire Galerkin impedance matrix. Moreover, only -integrals are sufﬁcient for purely metallic structures. These integrals are evaluated and , of quadrilateral elements in only once for any pair, the model. Rapid and accurate combined numerical/analytical methods are developed for the integration over curved higher order generalized quadrilateral surfaces, for the - and -integrals. When the distance in (35) is relatively small (or zero), the procedure of extracting the (quasi)singularity is performed [28]. As can be expected, the problems with the (quasi)singular integration are more pronounced with the ﬁeld integrals. Efﬁcient algorithms for recursive construction of the generalized Galerkin impedances and the EFIE/MFIE system matrix are used in order to avoid redundant operations related to the summation indexes in the Gauss–Legendre integration formulas, as well as the indexes and for current expansions and and for geometrical representations within the impedances. Starting with the generalized voltages given in (19) and (20), several models of lumped and distributed excitations and loads [29] are included in the proposed MoM technique (loads are introduced using the concept of a compensating electric ﬁeld). The resulting system of linear algebraic equations with complex unand is solved classically, by the Gaussian elimknowns ination. By postprocessing of these coefﬁcients, the currents and over any generalized quadrilateral in the model and ﬁelds and in any dielectric region (including the far ﬁeld) are obtained. III. NUMERICAL RESULTS AND DISCUSSION A. Dihedral Corner Reﬂector As an example of structures with ﬂat surfaces, consider the scattering from a metallic 90 dihedral corner reﬂector. The two large, are modeled by a plates, each being bilinear quadrilateral elements total of , which in this case reduce to squares, with the polynomial degrees in all of the elements. Without the use unknowns. Fig. 2 of symmetry, this results in shows the radar cross-section (RCS) of the reﬂector in the full azimuthal (horizontal) plane for the vertical polarization of the incident plane wave. The results obtained by the higher order

(36) where and are the geometrical orders along the and -coordinate, respectively, and are the geometrical vector coefﬁcients in the polynomial expansion of the th quadrilateral, , , and are the corresponding parameters for the th quadrilateral in the model, and is the basic Galerkin potential integral given by

(37) Similarly, using (32) and expanding the gradient of Green’s function, the electric/magnetic impedances in (23) are transformed to

(38) Using (26) then yields

2124

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Radar cross-section of a 90 dihedral corner reﬂector, in the full horizontal plane, for the vertical polarization of the incident plane wave, obtained by the higher order MoM and by the low-order MoM from [30].

Fig. 4. Four geometrical models of a spherical scatterer constructed from (a) = = 1). 96, (b) 216, (c) 384, and (d) 600 bilinear quadrilaterals (

K

K

of unknowns, based on a topological analysis, for a common low-order MoM solution with the reﬂector subdivided into triangular patches with Rao–Wilton–Glisson (RWG) basis functions [31] is around 12000, which is about 10 times the number of un. knowns required by the higher order MoM and B. Metallic Spherical Scatterer

Fig. 3. Radar cross-section of a dihedral corner reﬂector for four different orders (2, 4, 6, and 8) of the polynomial approximation of currents in the higher order MoM.

MoM are compared with the low-order MoM results from [30] (the number of unknowns is not speciﬁed in [30]), and an excellent agreement is observed. Note that the quadrilaterals in the on a side. higher order model are The convergence analysis of the higher order current approximation is performed for this example. Four different levels of the polynomial approximation of currents are adopted: (1) ; (2) ; ; and (4) (3) . The corresponding RCS results are shown in Fig. 3. We observe excellent convergence properties of the polynomial basis functions, the RCS prediction average absolute differences between levels (1) and (2), (2) and (3), and (3) and (4), being 7.6, 3.4, and 0.3 dB, respectively. In speciﬁc, note that even the second-order current approximation yields accurate result for the lobes at the directions perpendicular to the dihedral sides. Additionally, with the fourth-order basis functions, the dominant double-reﬂected ﬁelds in the forward region of the reﬂector are also predicted reasonably accurately. Finally, the sixth-order (or higher) current-approximation model adds the accuracy in the computation of ﬁelds in the back region of the reﬂector as well. Note also that the estimated number

As an example of curved metallic structures, consider a spherilluminated by an inciical metallic scatterer of radius dent plane electromagnetic wave in the frequency range 10–600 MHz. In the ﬁrst set of experiments, the ﬁrst-order geometin all elements). rical modeling is employed ( Four different geometrical models constructed from (1) bilinear quadrilaterals [Fig. 4(a)], (2) bilinear quadrilaterals [Fig. 4(b)], (3) bilinear quadrilaterals [Fig. 4(c)], and (4) bilinear quadrilaterals [Fig. 4(d)] are implemented, with the second-order current approximation in every element in all of the four models. The total numbers of unknowns without the use of symmetry 768, 1728, in models (1), (2), (3), and (4) amount to 3072, and 4800, respectively. Shown in Fig. 5 is the RCS of the sphere, normalized to the . The results obsphere cross-section area, as a function of tained by the higher order MoM are compared with the analytical solution in the form of Mie’s series. An excellent agreement between the numerical results obtained with the model (4) and analytical results is observed with the average absolute RCS prediction error less than 3%, while models (1), (2), and (3) provide acceptable results only up to the frequency at which 0.53, 1, and 1.6, respectively [the results obtained by the model (1) are not shown in Fig. 5]. Note that an increase in the curand in models (1)–(3) does rent-approximation orders not yield better results at higher frequencies, meaning that the

´ ˇ DJORDJEVIC AND NOTAROS: DOUBLE HIGHER ORDER MoM FOR SIE MODELING

2125

Fig. 7. Two geometrical models of a spherical scatterer constructed from (a) six and (b) 24 generalized quadrilaterals of the fourth geometrical order (K = K = 4). Fig. 5. Normalized radar cross-section [RCS=(a )] of a metallic sphere, for three higher order MoM models employing the ﬁrst-order geometrical modeling in Fig. 4(b)–(d), respectively, along with the exact solution (Mie’s series).

Fig. 8. Normalized radar cross-section [RCS=(a )] of a metallic sphere, for two higher order MoM models employing the fourth-order geometrical modeling in Fig. 7(a) and (b), respectively, along with the exact solution (Mie’s series).

Fig. 6. Induced electric surface current over the surface of the model in Fig. 4(b) at two frequencies, corresponding to (a) a= = 0:6 and (b) a= = 1:2.

errors in the RCS prediction using these models are a consequence of the inaccuracy in geometrical modeling of the sphere surface. Note also that, even though this is an almost small-domain application of the proposed large-domain method, where a large number (600) of elements (with relatively low current approximation orders) is needed for the sphere surface to be geometrically accurately represented by parametric surfaces of the ﬁrst geometrical order, the largest quadrilateral elements in the on a side at the highest frequency considmodel (4) are ered, which is still considerably above the usual small-domain . limit of For an additional insight into the correlation of errors in modeling of geometry and errors in modeling of currents, Fig. 6 shows the induced electric surface current over the surface of the model (2) at two frequencies, corresponding to and (b) . We observe that, while the (a) mutual orientation of quadrilateral elements in the model at the frequency (a) does not inﬂuence the surface current distribution over the sphere surface, the interconnections and surface-tangent discontinuities between quadrilaterals at the frequency (b) act like wedges, and a nonphysical current distribution is

obtained that follows the geometry of the quadrilateral mesh, where the variations of the current density magnitude clearly indicate the boundaries of the quadrilaterals constituting the model. These variations, of course, do not exist on the surface of the actual spherical scatterer. In other words, the error in modeling of curvature expressed in terms of the wavelength is negligible at the frequency (a), while at the frequency (b), it can not be ignored. The same conclusion is then translated from the current distribution consideration to the far ﬁeld and RCS computation at frequencies (a) and (b), as can be observed from Fig. 5. In the second set of experiments, the fourth-order geometrical 4 in all elements). The modeling is employed ( 6 fourth-order sphere surface is ﬁrst approximated by (A) quadrilaterals [Fig. 7(a)] in conjunction with the eighth-order 8) in each element and current approximation ( 24 fourth-order quadrilaterals then by (B) [Fig. 7(b)] with the sixth-order current approximation ( 6) in each element. This results in a total of 768 and 1728 unknowns in models (A) and (B), respectively, with no symmetry used. Fig. 8 shows the simulated RCS of the sphere obtained by the two geometrically higher order MoM models, as compared with the exact solution calculated in terms of Mie’s series. We observe an excellent agreement between the numerical results

2126

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 9.

("

Normalized radar cross-section [RCS=(a )] of a dielectric = 2:25) sphere, for three higher order MoM models employing the

Fig. 10. Normalized radar cross-section [RCS=(a )] of a dielectric (" = 2:25) sphere, for two higher order MoM models employing the fourth-order geometrical modeling in Fig. 7(a) and (b), respectively, along with the exact solution (Mie’s series).

ﬁrst-order geometrical modeling in Fig. 4(a)–(c), respectively, along with the exact solution (Mie’s series).

obtained with the model (A) and analytical results up to the freand the curved quadrilateral elequency at which across. In particular, ments in the model are approximately the maximum absolute RCS prediction error is less than 1% for (quadrilaterals are maximally across), and then increases slightly for . With the model (B), an excellent agreement with the exact solution is obtained in the entire frequency range considered, with the maximum aband solute RCS prediction error less than 0.5% for less than 3% for . Note that all the results for scattering from metallic spheres presented in this subsection are obtained by solving the EFIE (3) and no treatment of internal resonances is applied. The new double higher order method appears to yield equally accurate results at the internal resonances of the sphere, even though the condition number of the MoM matrix is very large at these frequencies. The RCS solution is sensitive to internal resonances only when the current approximation orders are not sufﬁcient, which is also in agreement with the previous results [32]. C. Dielectric Spherical Scatterer As an example of curved dielectric structures, consider a spherical dielectric scatterer 1 m in radius in the frequency range 10–600 MHz. The relative permittivity of the dielectric (polyethylene). Shown in Fig. 9 is the RCS of the is sphere calculated using the ﬁrst-order geometrical modeling , with the sphere surface being approximated by means of (1) bilinear quadrilaterals [Fig. 4(a)], (2) bilinear quadrilaterals [Fig. 4(b)], and (3) bilinear quadrilaterals [Fig. 4(c)], along with the analytical solution in the form of Mie’s series. The adopted electric and magnetic current approximation orders in models (1), (2), and 4, 2, and 2 and the resulting total numbers (3) are 6144, 3456, and 6144, respectively. We of unknowns observe that the RCS predictions are slightly shifted toward higher frequencies with all the three models, the frequency shift being the most pronounced with the model (1) at higher frequencies. The fact that the geometrical models are inscribed into the sphere certainly contributes to this shift of the results.

Note, however, that a very good agreement can be observed between the numerical results obtained by the model (3) and the analytical results in the entire frequency range considered. Note also that the numerical results in Fig. 9 obtained by any of the three models in Fig. 4(a)–(c) are signiﬁcantly more accurate than the corresponding numerical results obtained with the same models for the metallic sphere (Fig. 5), which can be attributed to the fact that inaccuracies in modeling of surfaces of penetrable (dielectric) bodies do not degrade the overall analysis results as signiﬁcantly as in the case of nonpenetrable (metallic) bodies. Fig. 10 shows the RCS of the dielectric sphere evaluated using geometrical models the two fourth-order shown in Fig. 7. In the model (A), the adopted electric and magnetic current approximation orders are , while in the model (B), these orders are set . We observe that, as to be compared to the exact solution (Mie’s series), the model (A) performs well up to the frequency at which and the curved quadrilateral elements in the model are about or across . Furthermore, the maximum absolute RCS prediction error is , with the maximum length less than 2% for of curved quadrilateral elements not exceeding . The model (B) provides an accurate RCS prediction in the entire frequency range considered (quadrilaterals are across at the highest frequency), with the maximum absolute error less than 1% for (maximum ) and a slightly side dimension of quadrilaterals is about increased error in the rest of the frequency range considered due to a minimal frequency shift of the results. D. Wire Monopoles Attached to a Metallic Cylinder As an example of antennas with curved surfaces, consider a system of wire monopoles attached to a metallic cylinder. The radius of the cylinder is 10 cm and its height 22 cm. The system is analyzed in two conﬁgurations: (1) with a single 12-cm monopole antenna attached to the cylinder and (2) with an 8-cm driven monopole and 44-cm parasitic monopole attached to the

´ ˇ DJORDJEVIC AND NOTAROS: DOUBLE HIGHER ORDER MoM FOR SIE MODELING

2127

Fig. 11. Circular cylinder of ﬁnite length with attached wire monopoles, = = 2) quadrilaterals and two wires. modeled by 32 biquadratic (

K

K

Fig. 12. Normalized far ﬁeld pattern and the antenna input impedance of the antenna system in Fig. 11 with only one monopole antenna present [conﬁguration (1)], obtained by the higher order MoM and by the low-order MoM from [33].

cylinder, as indicated in Fig. 11. The radii of the driven and passive monopoles are 1 and 2 mm, respectively. The antenna system is analyzed at the frequency of 833 MHz. Shown in Fig. 11 is the simulated geometrical model of the structure. The cylinder is modeled using 28 and 32 second-order quadrilateral surface elements in conﬁgurations (1) and (2), respectively. Each monopole is modeled by a single straight wire segment. The driven monopole is fed by a point delta generator at its base. Note that the triangle-like curved quadrilaterals are used around the wire-to-surface connections in order to easily enable current continuity across junctions. Note also that the ﬂexibility of the generalized quadrilaterals at approximating both the curvature of the surface and the curvature of the edges of the cylinder, along with their ﬂexibility to accommodate for degenerate quadrilateral shapes, enable the effective modeling of the cylinder with two junctions by means of only 32 surface elements. Note ﬁnally that neither the fact that the two adjacent outer edges of the quadrilaterals approximating the bases of the cylinder form an angle of 180 at the quadrilateral vertex they share nor the fact that the quadrilateral edges in the wire-to-surface junctions are extremely short (on the order of the wire radius) as compared to the other three edges of the quadrilateral do not deteriorate the accuracy of the current modeling and the overall accuracy of the analysis. The results for the radiated far ﬁeld obtained by the higher order MoM are compared with the results obtained by the low-order MoM from [33]. The patterns are shown in Fig. 12 for the conﬁguration (1) and Fig. 13 for the conﬁguration (2). The two-fold symmetry is used in both MoM approaches and a very good agreement of the two sets of results is observed. The discrepancy between the results is less than 3.5% in the entire pattern range in Fig. 13 and is practically nonexistent in Fig. 12. The simulation results for the monopole antenna impedance for the two conﬁgurations are given in Figs. 12 and 13 as well. We observe that the impedances computed by the two methods also agree very well. Note that the numbers of unknowns required by the higher order MoM, 49 for the conﬁguration (1) and 62 for the conﬁguration (2), are considerably smaller than the corresponding numbers of unknowns required by the low-order MoM [33], 936 and 986.

Fig. 13. Normalized far ﬁeld pattern and the antenna input impedance of the antenna system in Fig. 11 with both a driven monopole and a parasitic monopole present [conﬁguration (2)], obtained by the higher order MoM and by the low-order MoM from [33].

IV. CONCLUSION This paper has proposed a highly efﬁcient and accurate double higher order PC-oriented Galerkin-type MoM for modeling of arbitrary metallic and dielectric antennas and scatterers. The method is based on higher order geometrical modeling and higher order current modeling in the context of the SIE formulation for combined metallic (perfectly conducting) and dielectric (penetrable) structures. It employs generalized curvilinear quadrilaterals of arbitrary geometrical orders for the approximation of geometry (metallic and dielectric surfaces) and hierarchical divergence-conforming polynomial vector basis functions of arbitrary orders for the approximation of electric and magnetic surface currents within the elements. The geometrical orders and current-approximation orders of the elements are entirely independent from each other, and can be combined independently for the best overall performance of the method in different applications. The paper has presented the mathematical and computational development of the new MoM-SIE technique, including the evaluation of generalized

2128

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Galerkin impedances (MoM matrix elements) for double higher order quadrilateral boundary elements. The accuracy, convergence, and efﬁciency of the new MoM-SIE technique have been demonstrated in several characteristic examples. The results obtained by the higher order MoM have been validated against the analytical solutions and the numerical results obtained by low-order MoM techniques from literature. The ﬂexibility of the new technique has allowed for a very effective modeling of a dihedral corner reﬂector, a metallic spherical scatterer, a dielectric spherical scatterer, and a circular cylinder of ﬁnite length with attached wire monopoles by means of only a few large ﬂat and curved quadrilateral boundary elements and a minimal number of unknowns. All the examples have shown excellent ﬂexibility and efﬁciency of the new technique at modeling of both current variation and curvature. The examples have demonstrated advantages of large-domain models using curved quadrilaterals of high geometrical orders with basis functions of high current-approximation orders over commonly used small-domain models and existing low-order techniques from literature (the reduction in the number of unknowns is by an order of magnitude when compared to low-order solutions), but also over almost small-domain models that represent lower order versions of the proposed large-domain, high-order (more precisely, low-to-high order) technique. Finally, it has been demonstrated that both components of the double higher order modeling, i.e., higher order geometrical modeling and higher order current modeling, are essential for accurate and efﬁcient MoM-SIE analysis of general antenna (scattering) structures.

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22] [23]

REFERENCES

[1] R. F. Harrington, “Field computation by moment methods,” in IEEE PRESS Series on Electromagnetic Waves Piscataway, NJ, 1993. [2] S. M. Rao, C. C. Cha, R. L. Cravey, and D. L. Wilkes, “Electromagnetic scattering from arbitrary shaped conducting bodies coated with lossy materials of arbitrary thickness,” IEEE Trans. Antennas Propagat., vol. 39, pp. 627–631, May 1991. [3] M. Analoui, M. Kagawa, and Y. Kagawa, “Electromagnetic scattering from conductor-coated material bodies,” Int. J. Numerical Modeling: Electronic Networks, Devices and Fields, vol. 4, no. 4, pp. 287–299, Dec. 1991. [4] S. M. Rao, T. K. Sarkar, P. Midya, and A. R. Djordjevic, “Electromagnetic radiation and scattering from ﬁnite conducting and dielectric structures: surface/surface formulation,” IEEE Trans. Antennas Propagat., vol. 39, pp. 1034–1037, July 1991. [5] J. M. Song and W. C. Chew, “Moment method solutions using parametric geometry,” J. Electromagn. Waves and Applicat., vol. 9, no. 1/2, pp. 71–83, Jan.–Feb. 1995. [6] R. D. Graglia, D. R. Wilton, and A. F. Peterson, “Higher order interpolatory vector bases for computational electromagnetics,” IEEE Trans. Antennas Propagat., vol. 45, pp. 329–342, Mar. 1997. [7] G. Kang, J. Song, W. C. Chew, K. C. Donepudi, and J. M. Jin, “A novel grid-robust higher order vector basis function for the method of moments,” IEEE Trans. Antennas Propagat., vol. 49, pp. 908–915, June 2001. [8] S. Y. Chen, W. C. Chew, J. M. Song, and J. S. Zhao, “Analysis of low frequency scattering from penetrable scatterers,” IEEE Trans. Geoscience Remote Sensing, vol. 39, pp. 726–735, Apr. 2001. ´ [9] B. M. Notaroˇ, B. D. Popovic, J. Peeters Weem, R. A. Brown, and Z. s ´ Popovic, “Efﬁcient large-domain MoM solutions to electrically large practical EM problems,” IEEE Trans. Microwave Theory Tech., vol. 49, pp. 151–159, Jan. 2001. ´ [10] B. M. Notaroˇ and B. D. Popovic, “General entire-domain method for s analysis of dielectric scatterers,” in Proc. Inst. Elect. Eng.—Microwaves, Antennas and Propagation, vol. 143, Dec. 1996, pp. 498–504. [24]

[25] [26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

, “General entire-domain Galerkin method for analysis of wire antennas in the presence of dielectric bodies,” in Proc. Inst. Elect. Eng.—Microwaves, Antennas and Propagation, vol. 145, Feb. 1998, pp. 13–18. , “Large-domain integral-equation method for analysis of general 3D electromagnetic structures,” Proc. Inst. Elect. Eng.—Microwaves, Antennas and Propagation, vol. 145, no. 6, pp. 491–495, Dec. 1998. P. De Doncker, “A potential integral equations method for electromagnetic scattering by penetrable bodies,” IEEE Trans. Antennas Propagat., vol. 49, pp. 1037–1042, July 2001. K. Sertel and J. L. Volakis, “Method of moments solution of volume integral equations using parametric geometry modeling,” Radio Sci., vol. 37, no. 1, pp. 10–17, Jan.–Feb. 2002. R. D. Graglia, “The use of parametric elements in the moment method solution of static and dynamic volume integral equations,” IEEE Trans. Antennas Propagat., vol. AP-36, pp. 636–646, May 1988. J. S. Savage and A. F. Peterson, “Higher-order vector ﬁnite elements for tetrahedral cells,” IEEE Trans. Microwave Theory Tech., vol. 44, pp. 874–879, June 1996. L. S. Andersen and J. L. Volakis, “Development and application of a novel class of hierarchical tangential vector ﬁnite elements for electromagnetics,” IEEE Trans. Antennas Propagat., vol. 47, pp. 112–120, Jan. 1999. J. P. Webb, “Hierarchical vector basis functions of arbitrary order for triangular and tetrahedral ﬁnite elements,” IEEE Trans. Antennas Propagat., vol. 47, pp. 1244–1253, Aug. 1999. J. Liu and J.-M. Jin, “A novel hybridization of higher order ﬁnite element and boundary integral methods for electromagnetic scattering and radiation problems,” IEEE Trans. Antennas Propagat., vol. 49, pp. 1794–1806, Dec. 2001. ´ s M. M. Ilic and B. M. Notaroˇ, “Higher order hierarchical curved hexahedral vector ﬁnite elements for electromagnetic modeling,” IEEE Trans. Microwave Theory Techniques, vol. 51, pp. 1026–1033, Mar. 2003. ´ s M. Djordjevic and B. M. Notaroˇ, “Higher-order moment-method modeling of curved metallic antennas and scatterers,” in Proc. IEEE Antennas and Propagation Society Int. Symp. Dig., vol. 4, Columbus, OH, June 22–27, 2003, pp. 94–97. P. P. Silvester and R. L. Ferrari, Finite Elements for Electrical Engineers. Cambridge, U.K.: Cambridge Univ. Press, 1996. M. Salazar-Palma, T. P. Srkar, L. E. Garcia-Castillo, T. Roy, and A. ´ Djordjevic, Iterative and Self-Adaptive Finite Elements in Electromagnetic Modeling. Boston, MA: Artech House, 1998. R. D. Graglia, P. L. E. Uslenghi, R. Vitiello, and U. D’Elia, “Electromagnetic scattering for oblique incidence on impedance bodies of revolution,” IEEE Trans. Antennas Propagat., vol. 43, pp. 11–26, Jan. 1995. J. C. Nedelec, “Mixed ﬁnite elements in R3,” Numerische Mathematik, vol. 35, pp. 315–341, 1980. ´ s M. Djordjevic and B. M. Notaroˇ, “Three types of higher-order MoM basis functions automatically satisfying current continuity conditions,” in Proc. IEEE Antennas and Propagation Society Int. Symp. Digest, San Antonio, TX, June 16–21, 2002, pp. 610–613. , “Higher-order hierarchical basis functions with improved orthogonality properties for moment-method modeling of metallic and dielectric microwave structures,” Microwave Opt. Technol. Lett., vol. 37, no. 2, pp. 83–88, Apr. 2003. ´ B. M. Notaroˇ and B. D. Popovic, “Optimized entire-domain moments method analysis of 3D dielectric scatterers,” Int. J. Numerical Modelling: Electronic Networks, Devices and Fields, vol. 10, pp. 177–192, May–June 1997. , “Generalized excitations and loads for electromagnetic analysis with boundary elements,” Int. J. Engineering Analysis With Boundary Elements, ELSEVIER: Special Issue on Electromagnetics, vol. 27, no. 4, pp. 333–343, Apr. 2003. T. Griesser, C. A. Balanis, and L. Kefeng, “RCS analysis and reduction for lossy dihedral corner reﬂectors,” Proc. IEEE, vol. 77, no. 5, pp. 806–814, May 1989. S. M. Rao, D. R. Wilton, and A. W. Glisson, “Electromagnetic scattering by surfaces of arbitrary shape,” IEEE Trans. Antennas Propagat., vol. AP-30, pp. 409–418, May 1982. F. X. Canning, “Protecting EFIE-based scattering computations from effects of interior resonances,” IEEE Trans. Antennas Propagat., vol. 39, pp. 1545–1552, Nov. 1991. R. E. Hodges and Y. Rahmat-Samii, “An iterative current-based hybrid method for complex structures,” IEEE Trans. Antennas Propagat., vol. 45, pp. 265–276, Feb. 1997.

´ ˇ DJORDJEVIC AND NOTAROS: DOUBLE HIGHER ORDER MoM FOR SIE MODELING

2129

´ Miroslav Djordjevic (S’00–M’04) was born ´ in Cuprija, Serbia and Montenegro (former Yugoslavia), in 1973. He received the Dipl.Ing. (B.S.) degree from the University of Belgrade, Belgrade, Serbia and Montenegro, in 1998, the M.S. degree from the University of California, Los Angeles (UCLA), in 2000, and the Ph.D. degree from the University of Massachusetts (UMass) Dartmouth, in 2004. From 1998 to 2000, he was a Graduate Student Researcher at the Antenna Research and Measurement (ARAM) Laboratory, UCLA. Since 2000 to 2003, he was a Research Assistant at UMass where he is currently a Postdoctoral Associate. His research interests are in higher order modeling, hybrid methods, and analysis of vehicle mounted antennas.

Branislav M. Notaroˇ (M’00–SM’03) was born s in Zrenjanin, Yugoslavia, in 1965. He received the Dipl.Ing. (B.Sc.), M.Sc., and Ph.D. degrees in electrical engineering from the University of Belgrade, Belgrade, Yugoslavia, in 1988, 1992, and 1995, respectively. He is currently an Assistant Professor of electrical and computer engineering with the University of Massachusetts Dartmouth. From 1996 to 1998, he was an Assistant Professor with the Department of Electrical Engineering, University of Belgrade. He spent the 1998 to 1999 academic year as a Visiting Research Associate with the University of Colorado at Boulder. His teaching activities are in the area of theoretical and applied electromagnetics. He is the Co-Director of the Telecommunications Laboratory, Advanced Technology and Manufacturing Center, University of Massachusetts Dartmouth. He has authored or coauthored 15 journal papers, 40 conference papers, a book chapter, ﬁve university textbooks and workbooks, and a conceptual assessment tool for electromagnetics education. His research interests are predominantly in computational electromagnetics and antenna design. Dr. Notaroˇ was the recipient of the 1999 Institution of Electrical Engineers s (IEE) Marconi Premium.

2130

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Loop-Tree Implementation of the Adaptive Integral Method (AIM) for Numerically-Stable, Broadband, Fast Electromagnetic Modeling

Vladimir I. Okhmatovski, Member, IEEE, Jason D. Morsey, Member, IEEE, and Andreas C. Cangellaris, Fellow, IEEE

Abstract—The adaptive integral method (AIM) is implemented in conjunction with the loop-tree (LT) decomposition of the electric current density in the method of moments approximation of the electric ﬁeld integral equation. The representation of the unknown currents in terms of its solenoidal and irrotational components allows for accurate, broadband electromagnetic (EM) simulation without low-frequency numerical instability problems, while scaling of computational complexity and memory storage with the size of the problem of the are of the same order as in the conventional AIM algorithm. The proposed algorithm is built as an extension to the conventional AIM formulation that utilizes roof-top expansion functions, thus providing direct and easy way for the development of the new stable formulation when the roof-top based AIM is available. A new preconditioning strategy utilizing near interactions in the system which are typically available in the implementation of fast solvers is proposed and tested. The discussed preconditioner can be used with both roof-top and LT formulations of AIM and other fast algorithms. The resulting AIM implementation is validated through its application to the broadband, EM analysis of large microstrip antennas and planar interconnect structures. Index Terms—Fast algorithms, full-wave electromagnetic (EM) CAD, loop-tree (LT) decomposition, low frequency, method of moments (MoM).

I. INTRODUCTION T is a matter of common experience in the electromagnetic (EM) modeling community that the conventional integral equation based method of moments (MoM) exhibits two major shortcomings when used for broadband EM analysis of electrically large and/or geometrically complex structures. The ﬁrst shortcoming has to do with its numerical solution time and memory requirements, both growing at least as a square of the number of unknowns involved in the MoM approximation of the EM boundary value problem. State-of-the-art applications of relevance to integrated microwave/RF, and mixed-signal electronic devices and systems, call for EM models where the

Manuscript received June 20, 2003. This work was supported in part by the Defense Advanced Research Projects Agency (DARPA) NeoCAD program under Grant N66001-01-1-8921 and in part by the Semiconductor Research Corporation. V. I. Okhmatovski is with Cadence Design Systems, Incorporated, Tempe, AZ 85282 USA (e-mail: okhmatov@cadence.com). J. D. Morsey was with the Center for Computational Electromagnetics, Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801 USA. He is now with the IBM T. J. Watson Research Center, Yorktown Heights, NY 10598 USA. A. C. Cangellaris is with the Center for Computational Electromagnetics, Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801 USA. Digital Object Identiﬁer 10.1109/TAP.2004.832326

I

number of unknowns is in the order of tens and even hundreds of thousands. This makes the direct implementation of MoM-approximations of EM integral equations computationally prohibitive and thus impractical. The second shortcoming of the MoM approximation of EM integral equations is the so-called low frequency breakdown [1], [2] that occurs when the harmonic ﬁeld wavelength becomes substantially larger than the characteristic size of the MoM grid. In the application of MoM-based solvers for the narrow band analysis of traditional EM devices of resonant lengths (e.g., antennas, waveguides, and various types of RF/microwave passive components) this problem of geometry over-sampling tends to be the exception rather than the rule. However, over the last decade, the advent of miniaturization and high-density integration of electronic devices has led to new classes of RF/microwave passive components and associated integrated waveguides and packaging structures of increasing geometric complexity, primarily due to the presence of numerous minute features, the accurate modeling of which results in the aforementioned geometric over-sampling in the development of the MoM discrete model. Even in the absence of minute features, the aggressive push toward the integration of mixed-signal (e.g., high-speed digital and RF/microwave) functional blocks, calls for broadband EM modeling (from dc to multi-GHz frequencies) in support of computer-aided design of electromagnetically compatible, EMI-immune devices [3], [4]. While it is applications of this type that this paper is concerned with, it is worth mentioning that another important class of applications where low-frequency breakdown of MoM solvers is encountered is the kilohertz-range remote sensing of the buried objects [5], [6]. There are two major classes of methods that are capable of tackling effectively the computational complexity and large memory requirements of MoM approximations to EM integral equations. The ﬁrst class includes the fast multipole method (FMM) [7] and its multilevel modiﬁcations (MLFMM) [8]. The second class includes all the acceleration schemes that take advantage of the convolution nature of the EM ﬁeld integral equation to expedite its calculation through the use of fast Fourier transforms (FFT) algorithms [9]–[14]. For quasiplanar geometries both the FMM and FFT-based algorithms reduce the required CPU time in the iterative solution of the complexity MoM system from the aforementioned complexity. Also within the framework of to memory consumption of MoM is fast algorithms the substantially reduced and scales proportionally to the problem

0018-926X/04$20.00 © 2004 IEEE

OKHMATOVSKI et al.: LT IMPLEMENTATION OF AIM FOR FEM MODELING

2131

size . Even though the FMM and FFT acceleration schemes lead to similar improvements in computational efﬁciency, the FFT-based schemes are much more simple to implement and more ﬂexible in usage of various types of Green’s functions accounting for the surrounding medium. With regards to the theme of this paper, it is important to point out that, irrespective of the fast solution methodology employed, the problem of low-frequency breakdown remains and has to be addressed in the context of the speciﬁc fast solution scheme. Thus, if a robust and broadband EM simulator is to be developed, it is necessary to introduce appropriate modiﬁcations in both the development of the MoM approximation and the acceleration schemes to avoid the numerical instability at low frequencies. For the FMM class of fast solvers such a modiﬁcation was presented in [16], [17]. As discussed in [18], the fact that in the regular FMM the threshold distance separating the near and far interactions in the system and/or the order of the multipole expansion are related to the wavelength makes necessary the reformulation of the entire FMM scheme in order to ensure numerical stability at low frequencies. For the class of FFT-based fast solvers, the separation into the near and far interactions is either not required at all (in the case when the MoM mesh is bound to the FFT grid [13]) or the threshold distance separating far and near interactions of the algorithm is not related to the wavelength (this is the case for the adaptive versions of such solvers where the MoM and FFT grids are independent [10], [12]). In view of this property of the FFT-based fast solvers it was proposed in [15] that the problem of low-frequency breakdown may be circumvented through the use of a higher-order, locally corrected Nyström method [19], combined with a nonuniform grid FFT scheme [20]. To understand how the Nyström scheme stabilizes the solution of the electric ﬁeld integral equation (EFIE) at low frequencies one must recall that the nature of low-frequency breakdown is purely numerical [1]. It is caused by the loss due to round-off error of the small term compared to much larger term in the scattered ﬁeld representation (1) . Since higher-order methods evaluate both of when these terms with much higher accuracy, low-frequency numerical instability occurs at much lower frequencies compared to the conventional MoM with Rao–Wilton–Glisson (RWG) expansion functions [21]. In this paper, an alternative implementation of the adaptive integral method (AIM) [also referred to as the precorrected FFT algorithm] is proposed for overcoming the low-frequency breakdown. First, within the framework of the MoM approximation of the EFIE using roof-top expansion functions, the loop-tree (LT) decomposition of the unknown current is introduced as prescribed in [1], [2]. This change of expansion functions from roof-tops to loops and branches of the tree allows for the explicit separation of the irrotational and solenoidal components of the current. Subsequently, the AIM solution process is adjusted to accommodate the new representations of the discretized current densities. As a result, the overall

solution complexity and memory usage of the conventional AIM are preserved, while the numerical stability of the solution is maintained down to very low frequencies. The paper is organized as follows. In Section II, MoM approximation in the standard and mixed-potential forms is outlined, with emphasis on conversion from roof-top into the LT basis. In Section III, the proposed alternative AIM process is presented in standard and mixed potential forms, highlighting the differences and similarities between the new and the conventional formulations. In the following, the proposed implementation will be referred to as AIM-LT. Section IV discusses a new preconditioning strategy utilizing near interactions in the system. The numerical studies of Section V are used to demonstrate the validity of the AIM-LT methodology. Finally, Section VI concludes the paper with a brief summary of the method and a few remarks about its attributes. II. LOOP-TREE DECOMPOSITION OF THE MoM WITH ROOF-TOP BASIS The EFIE statement of the problem is usually obtained through the application of Green’s theorem to the distinct volumes of the structure. Assuming perfect electrically conducting (PEC) surfaces, the EFIE forces the tangential electric ﬁeld, ﬂowing on PEC surface of the produced by the current on circuit , to cancel the applied tangential electric ﬁeld the same surface . This results in standard [22]

(2) or mixed potential formulation [12], [23] of EFIE

In (2), is the unit normal vector to ; and and are the position vectors to the observation and source points, respectively. over a set of roof-top exExpanding the unknown current pansion functions with triangular [21] or rectangular support [22] (3) and testing the scattered ﬁeld with the same basis functions , the integral equation (2) is reduced to a set of linear algebraic equations (4) In (4), and contain, respectively, the coefﬁcients of expansion (3) and the discrete form of the excitation. The impedance

2132

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

matrix elements and excitation vector elements represented by the following inner products

are

, and are In practice submatrices and resulting from the evaluated based on the matrices roof-top formulation of MoM. For this purpose basis conversion and are introduced as follow:s matrices (12)

(5) where product is deﬁned as and the Hilbert space inner transformation matrix from the where is the sparse loop-based expansion functions to roof-tops, is the sparse transformation matrices from the branch-based expansion functions to the roof-tops, and the symbol denotes matrix transposition [1], [2]. The total number of loops and , is equal to the number of the branches in the tree, roof-top expansion functions. Since the tree represents an undirected graph deﬁned on the original MoM roof-top grid, the th contains only one nonzero element equal to row of the matrix unity in the th column corresponding to the th roof-top taken as the th branch in the tree. Each loop presents a directed graph contains on the roof-top mesh and the th row of matrix nonzero values in the columns corresponding to the roof-tops composing the th loop. The sign of these elements is positive if the corresponding roof-top function is collinear with the orientation of the loop, and negative if it has the opposite direction. is equal to the The absolute value of the nonzero elements in inverse of the corresponding roof-top function width in order to satisfy the condition of continuous current ﬂux through any cross section of the loop. Using (12) we can represent matrix in terms of the roof-top based impedance matrices and as (13) where the sparse matrices and (14) are the conversion matrices from the LT basis to the roof-top basis. The zero in the matrix signiﬁes the loop properties (8). As a result, the following formulas for the submatrices in (11) and can be obtained in terms of

(6) In the form (4), the primary cause of the low-frequency breakdown of the MoM approximation is easily recognized. As , the ﬁrst term in (4) drops below the numerical error level of the second term due to the ﬁnite accuracy of computations. In order to avoid this problem, the representation of unknown (rotational component) current in terms of loop currents (irrotational component) is inand tree-branches currents troduced

(7) The current ﬂux through any cross section of the loop remains constant forcing the following conditions to be satisﬁed (8) where is a potential in the range of operator on the surface . Application of the Galerkin process with loops and the tree branches as expansion and testing functions yields an alternative form of the MoM matrix that is free of the low-frequency breakdown problem (9) which in abbreviated form can be written as (10) The matrix elements in (9) are as follows:

(15) Thus, the matrix equation (9) can be interpreted as the original and capacitive parts MoM matrix (4) with its inductive preconditioned to the left and to the right with matrices and , respectively, in order to improve spectral properties of the roof-top based impedance matrix at low frequencies and/or for electrically oversampled structures. III. LOOP-TREE IMPLEMENTATION OF THE AIM ALGORITHM In the iterative solution of (9) a repetitive computation of the is required and can be accelermatrix-vector product

(11) In the process of evaluation of the matrix elements in (9) conditions (8) were taken into consideration. Also it was assumed that the media and the PEC object are such that the transfer of is allowed [23]. the divergence operator to the current

OKHMATOVSKI et al.: LT IMPLEMENTATION OF AIM FOR FEM MODELING

2133

placed by a set mulas

dipoles, as described by the following for-

(18) of the dipoles have where for each basis function only nonzero amplitudes, with being the total number of FFT grid points. The projection of the rectangular shaped roof-top functions on the FFT grid for the case of two-dimensional (2-D) structure is illustrated in Fig. 1. Various criteria can be devised for the evaluation of the dipole weights and , the most economical among them being the so-called multipole reproduction criteria discussed in [10]. Substitution of the dipole representations (18) for the basis and testing functions into the inner product formulae for the impedance matrix elements (5) yields the following expressions for the elements

Fig. 1. Illustration of a pair of closely interacting loops and roof-top function projection on the FFT grid.

ated by means of the AIM algorithm. Namely, the matrix-vector product is decomposed into near and far interactions in the following manner (16) A. Near Interactions The matrix of near interactions is sparse and the comis . In order to elaborate on plexity of its product with the deﬁnition of closely interacting loops let us consider loops and shown in Fig. 1. The loop being a concatenation of roof-top functions , is considered closely inter, acting with the loop composed of the roof-tops if there exists at least one pair of roof-top functions and , such that the distance between and is smaller than a certain preset threshold. The same rule is applied when it is to be decided whether or not a tree branch interacts closely with a given loop. Alternatively, if the matrices of and deﬁned in the roof-top basis near interactions are available from the conventional implementation of AIM, a is aldifferent deﬁnition of the matrix of near interactions lowed (17) Clearly, the more spatially localized the loops are the sparser the and become. Whether the deﬁnition matrices or is adopted the matrix of near interactions is computed once and stored, the storage requirement being of order due to its sparsity. The part of the matrix responsible for the is neither computed directly nor stored. far interactions is computed on the Instead the matrix-vector product ﬂy using FFT’s as described next. B. Far Interactions First, all the basis roof-top functions involved in the original MoM discretization of the object are projected onto the FFT grid enclosing the object in the same manner as done in the conventional AIM [10]. If mixed-potential formulation is used, then are also projected on the same FFT charge duplets grid. Projection means that each basis function and is re(21) where subscript FFT is used to indicate that the matrices are computed using projections of the basis and testing functions on the FFT grid. (19) In (19) multidimensional matrices are used. In order to distinguish two groups of dimensions in them, both boldface typing and square brackets are utilized. The boldface characters imply the spatial vector and dyadic structure of the quantities while the square brackets emphasize their discrete nature due to the projection on the FFT grid. In expanded form, for the case of a -plane, (19) becomes 2-D object located in

where

(20) and are the dimensions of the FFT grid along In (20), and , respectively, while and are the corresponding grid steps. Substitution of formula (19) into (15) leads to the following expressions for the blocks of the impedance matrix in LT basis

2134

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

From (21) it is apparent that matrix products and stand for the projections of the testing and basis loops and on the FFT grid. Similarly, the products characterize the projection of the testing and basis trees on the FFT grid. Also it is important to observe that these products are nothing else but the linear combinations of the projections from the individual basis/testing functions composing them. Thus, in a way analogous to the conventional AIM, the matrix-vector can be computed with ﬂops product using FFT due to the Toeplitz structure of the discretized dyadic Green’s functions. The part of the matrix-vector multiply associated with inductive interactions in the system is evaluated as follows:

Since the branch-based expansion functions are nothing else but selected roof-top expansion functions, the precorrection procedure for the interactions between branch-based testing and expansion functions is identical to that in the conventional AIM. At the same time, extraction of the near interactions corresponding to closely located loops and branches allows for two alternative approaches as has been discussed in the Subsection A. In the ﬁrst approach, one may extract contributions of the closely interacting loops and branches by removing entire elements from pertinent to them. Alternatively, one has the matrix due the option of extracting only the contributions to to the closely interacting roof-tops that contribute to the formation of the pertinent loops or branches. Let us refer to these contributions as (26) Clearly, whichever of the two approaches, extraction of or from in (25), is used the same precorrection method must be applied when the elements of are evaluated in (16). From the point of the matrix computational performance the two precorrection strategies are is equivalent. However, the second approach, where evaluated, may be easier to implement when the conventional AIM algorithm is already in place. This is due to the fact that, if the near interactions for roof-top functions are available, there is no need for the computation of any additional elements and in order to construct of matrices and . Irrespective of which approach is adopted, the matrix of near interactions is computed only once and then stored; thus, both the computational complexity associated with its construction and the memory requirements for its . storage scale as IV. PRECONDITIONER BASED ON NEAR INTERACTIONS contains information about the strongest The matrix interactions in the impedance matrix ; hence, its inverse can serve as an effective preconditioner in such commonly used iterative solvers as the conjugate gradient (CG) method [24] or the generalized minimal residual (GMRES) is sparse [Fig. 2(a)], its algorithm [25]. Even though computational complexity and direct inversion exhibits memory requirement since is in general a full matrix. The practiced solution to this problem is the class of so-called incomplete factorization methods [26]. The basic idea behind these methods is to discard the elements having values below a certain threshold in the process of the LU factorization, thus obtaining a sparse approximation to the originally full L and U matrices. Although these methods can lead to reduction of both memory storage and CPU time , our numerical experiments with their complexity to application to problems involving planar structures found them unable to provide acceptable iteration counts in certain cases. To remedy the situation an alternative preconditioning strategy is introduced. The key idea of this alternative approach through a reindexation of the is to achieve sparsity of unknowns (i.e., the coefﬁcients in the loop- and branch-based

(22) The part corresponding to the capacitive coupling in the system can be evaluated in two different ways depending on whether the standard or mixed potential form of EFIE is considered

(23) In (22) and (23), the operators and denote forward and backward discrete FFT, respectively. Clearly, if the AIM in the roof-top basis has been implemented and the fast computational kernel for evaluation of the and is available, the matrix-vector products can be easily computed as product (24) which is equivalent to formulas (22) and (23). C. Precorrection The dipole representation (18) for basis and testing functions provides accurate approximation of the impedance matrix elements only when the th and th elements are at a sufﬁciently large distance. Consequently, all elements of containing contributions from that do not satisfy this requirement must be removed from the matrix-vector product in order to obtain accurate approximation of the dein (16) sired product

(25)

OKHMATOVSKI et al.: LT IMPLEMENTATION OF AIM FOR FEM MODELING

2135

Fig. 2. Matrix of close interactions Z and U matrices of z , and (c).

(a) Reindexed matrix z

, (b) L

expansion functions) that would lead to a new matrix with its elements clustered near the diagonal. Hence, after has a banded structure, as reindexing, the new matrix depicted in Fig. 2(b), and contains all the nonzero elements of . Subsequently, taking advantage of the banded the matrix , its exact LU-factorization can be performed. structure of This LU factorization is done only once and is of complexity , where is the bandwidth of . The factors and of are also banded with the same bandwidth as that of [Fig. 2(c)]. Thus, storage of the banded and matrices of requires memory. The backward substitution process, yielding the desired product of matrix with the vector at each iteration, is of . complexity Clearly, performance of the preconditioner is strongly depenchanges with the problem size dent on how the bandwidth . In order to explore further this dependence of on size it is necessary elaborate on the details of the reindexing procedure used. The algorithm discussed next is very similar to the reverse Cuthill–McKee ordering [27], [28] which is enhanced by

considering the distance between two interacting elements as an additional parameter for reordering. To band the sparse matrix the process starts with an arbitrary element labeled R1. The testing function associated with it has expansion functions closely interacting with it, including the self-term. These expansion functions are then reindexed and assigned indexes in the order of increasing center-to-center disfrom 1 to tance from the testing function. Clearly, expansion function 2, the closest to R1 is indexed as R2. Given its close proximity to R1, most of its near interactions have been assigned a new index already. Those few interaction that have not been reindexed are and in order of increasing reindexed next, starting with distance from R2. This procedure continues until all elements have been reindexed and the banded matrix has been generated. The new preconditioner is sparse within the bandwidth, as depicted in Fig. 2(b). From the description of the reindexing algorithm it can be deduced that for geometries with continuous surfaces, such as a sphere or a square plate, the bandwidth is expected to grow with the size of the problem as . However, for the types of geometries encountered in planar integrated circuits (see, for example, Fig. 5), the bandwidth does not change with the number of unknowns provided that the radius of the sphere that deﬁnes the range of near interactions is kept smaller than the pitch between the lines. This point will become clear in Section V through the numerical examples used to investigate the performance of the proposed fast solver. Another observation worth making is that although the and preconditioner is sparse within the bandwidth, the matrices are in general full within the same bandwidth. Since memory usage is controlled by the bandwidth, one can ﬁll in the bandwidth of the preconditioner prior to the LU factorization. The storage requirements of the LU factorization remain unchanged; however, the quality of the preconditioner is improved signiﬁcantly. Finally, it is stressed that the proposed preconditioner can be applied with both the LT and roof-top basis functions. Once the near interactions and center-to-center distances between expansion functions have been deﬁned, the sorting algorithm remains essentially the same irrespective of the choice of the expansion functions. An exception is the case when large loops, referred to as super loops for the purposes of this manuscript, are present in the LT implementation [1]. Such super loops contain more near interactions since they cover a larger area of the analyzed structure. Even though every effort should be made to contain the size of the loops during the loop generation process, super loops are expected to occur, especially in conjunction with structures that include conducting portions forming closed loops (e.g., shorted sections of coplanar waveguides). When such super-loops are naturally present, the near interactions between these super loops and any elements except for other superloops must be ignored, in order to prevent the formation of an excessively oversized preconditioner.

V. NUMERICAL RESULTS AND DISCUSSION The fast solution methodology developed in the previous chapters is suitable for the integral equation-based EM

2136

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 3. Input reﬂection coefﬁcient of the 2

2 2 microstrip patch array computed with regular AIM, and AIM in LT basis.

modeling of both 3-D and 2.5D structures. However, the primary objective of our research has been the development of a fast and robust solver for the broadband EM analysis of dense, complex predominantly planar structures used in RF/microwave printed circuits, planar antenna arrays, and the high-density interconnect circuits used at all levels of packaging of high-speed/highfrequency mixed-signal integrated systems. The examples presented next involve representative members of the aforementioned categories of structures. A. Corporate-Fed Microstrip Antenna Array In the ﬁrst example, four different corporate-fed microstrip antenna arrays are considered. The matrix of patch antennas constituting these arrays had dimensions 2 2, 2 4, 4 4, and 8 4. The corresponding sizes of the MoM approximations were , and , respectively. The geometry and dimensions of the 2 2 antenna array are shown in Fig. 3. The 8 4 array and the current distribution, , at 20 GHz are depicted in Fig. 4. The air/dielectric interface was taken to be mm above the ground plane. The current distribution depicted in Fig. 4 is calculated for the , whereas for case of a substrate with permittivity the remaining numerical results generated for the purposes of demonstrating the performance of the proposed fast solver the surrounding medium of antennas was assumed to be air. The 2 2 array was modeled using the original AIM with roof-top expansion functions, and the proposed LT formulation of AIM both implemented in the standard EFIE form (2). The was computed magnitude of the input reﬂection coefﬁcient using standing wave characterization and is plotted versus frequency in Fig. 3. It can be seen that at higher frequencies the

Fig. 4. Geometry of 8 GHz.

2 4 array and current distribution log(jJ

(r ) ) at 20

j

methods are in very good agreement. However, at lower frequencies, the standard AIM exhibits numerical instability, while its LT version remains robust and produces accurate results. The parameters of the AIM implementation where chosen as follows. For the projection of each roof-top function on the FFT were used, as prescribed by forgrid sixteen dipoles mula (18). The area of near interactions for each basis/testing roof-top function, which establishes the threshold between near and far interactions, was deﬁned in terms of the FFT grid steps and remained the same for all frequencies. For the speciﬁc example considered here, two functions were assumed closely interacting if any two dipoles from their projections were separated by less than 11 points in either or directions.

OKHMATOVSKI et al.: LT IMPLEMENTATION OF AIM FOR FEM MODELING

2137

**TABLE I CPU TIME
**

AND MEMORY REQUIREMENTS ON A 1.4 GHz FOR CORPORATE-FED ARRAYS AT 20 GHz

PENTIUM4

the LU factorization of the preconditioner for the beneﬁt of the small number of iterations in the solution. For example, such a choice will be most appropriate when a multiport structure with a large number of ports (hence, multiple right-hand side vectors) is being modeled. B. Multiconductor Interconnect The second example considered is a multiconductor interconnect of enough nonuniformity along the direction of signal propagation to require full-wave EM modeling for the quantiﬁcation of its transmission and signal interference (crosstalk) properties. For high-speed digital signal transmission as well as for interference analysis pertinent to mixed-signal systems, the EM properties of such interconnects must be computed from dc to multi-GHz frequencies. Thus, in addition to the large computational complexity of such structures, the need for broadband EM analysis provides a signiﬁcant challenge for conventional full-wave integral equation solvers aimed for primarily narrow band frequency modeling of microwave and RF waveguides and passive components. The top view of the geometry of a generic 16-line microstrip interconnect is depicted in Fig. 5. The sixteen wires are organized into eight differential lines. It is stressed that the ﬁne discretization along the cross-section of the wires is needed for the accurate prediction of both the speed of signal transmission and the EM coupling between adjacent lines. For the purposes of this analysis the wires were assumed to be of zero thickness and perfectly conducting. In order to provide a reference solution, a portion of this interconnect structure was analyzed using the EM ﬁeld solver from Sonnet [29]. The structure analyzed in Sonnet consisted only of the differential line that is depicted in Fig. 5 as having terminals 1 through 4, and the two adjacent differential lines. Analysis of a truncated portion of the structure in Sonnet was necessitated by the signiﬁcant complexity of the 16-wire structure, the detailed modeling of which makes the use of this direct EM solver very time consuming. Magnitudes of some of the calculated -parameters are presented in Fig. 6. Good agreement of the results is observed over the entire bandwidth of analysis. The discrepancies between -parameters at higher frequencies, as well as for those frequencies and ports for which coupling becomes week, are attributed to the difference of the geometries analyzed by the two methods, namely, the entire 16-wire structure using the proposed method and the 8-wire portion of it using Sonnet. It is important to mention that no de-embedding of the port discontinuities was done in the generation of the presented -parameters [30]. In order to demonstrate the robustness of proposed algorithm the operating frequency of analysis was taken down to 0.001 Hz. versus frequency is plotted The admittance matrix element in Fig. 7 in the range from 0.001 Hz to 8 GHz. It can be seen that while the Sonnet EM solver experiences the low-frequency breakdown the LT implementation of the AIM algorithm provides accurate results in the range of frequencies stretching from dc to microwaves. The numerical statistics of computations for the interconnect structure of Fig. 5 is summarized in Table II. In order to demonstrate how the proposed algorithm scales with the size of the

Fig. 5. Geometry of the 16-wire microstrip interconnect structure. The circuit is printed on lossless grounded dielectric substrate of permittivity = 4 and thickness d = 0:15 mm.

Table I summarizes all information necessary for quantifying computational complexity of the proposed fast solver. The discussed AIM-LT implementation utilizing LT basis functions computational complexity and exhibits the same memory requirements as the conventional AIM implemented with roof-top functions. The low iteration count was made possible through the use of the preconditioner discussed in Section IV. The penalty paid for the effectiveness of the memory storage requirement preconditioner is the complexity of its LU factorization. The latter and the can deﬁnitely become a show-stopper in the use of such a preis in the order of hundreds of thousands. conditioner when Therefore, it is important that the numerous possibilities that exist for the construction of more efﬁcient preconditioners, in the sense that they provide for lower computational complexity without jeopardizing the efﬁciency of the iterative solver, are thoroughly investigated. The choice made here reﬂects the situation where it was preferable to trade-off up front CPU time for

2138

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 6.

Magnitude of some of Y -parameters for 16-wire interconnect.

TABLE II CPU TIME AND MEMORY ON 1.4 GHz PENTIUM4 FOR FOUR-, EIGHT-, AND 16-WIRE INTERCONNECTS AT 10 GHz

structure, portions of the structure including four, eight and 16 wires were considered separately. Also, the numerical solution was carried out using two different preconditioners. For the ﬁrst preconditioner only the matrix elements corresponding to interactions between testing/expansion functions separated by a distance of 0.15 mm or less were included. This means that the matrix elements of the basis and testing functions (loops and branches) located on different wires were not utilized in the preconditioner. We refer to this preconditioner as P1 in Table II. For the second preconditioner, referred to as P2, the radius of near interactions was extended so that the coupling between expansion and testing functions located on adjacent wires was included. In this case the penalty for the improved quality of pre-

conditioner and the faster convergence of the iterative solution was the substantial increase in memory usage. The number of iterations taken by the GMRES algorithm [25] to converge to outer residual error is plotted in Fig. 8 for the cases when the matrix equation was preconditioned with P1 and P2. Fig. 8 demonstrates that while at low frequencies the choice of preconditioner is irrelevant, the iteration count may become excessively large at higher frequencies if the coupling between the neighboring lines is not included in the preconditioner. It is important to point out that the total CPU time of the algorithm, presented in the last row of Table II, was calculated as the sum of the time required by the iterative solver and the time spent for the LU-factorization. Other computations associated with tasks such as the evaluation of the Green’s function and the computa, were tion of the matrix elements for the near interactions not included in the complexity evaluation. This is due to the subjective nature of the time estimates for these computations in view of the wide variety of methods that can be used for performing these calculations. VI. CONCLUSION This paper discusses a new fast algorithm for the iterative solution of MoM approximations of EM ﬁeld integral equations pertinent to the analysis of primarily planar and/or layered passive structures. The methodology is based on an alternative implementation of the AIM (pre-corrected-FFT algorithm), where the LT decomposition of unknown current is introduced to enhance the numerical stability of the iterative solution down to very low frequencies. Both mixed-potential and standard

OKHMATOVSKI et al.: LT IMPLEMENTATION OF AIM FOR FEM MODELING

2139

Fig. 7. Magnitude of some of the Y

-parameters for 16-wire interconnect at very low frequencies.

Fig. 8. Iteration count of GMRES solver versus frequency for 16-line interconnect depicted in Fig. 5.

formulations of the EFIE can be accommodated within the framework of the proposed method. For microstrip structures (2.5D geometries) the implementation of AIM utilizing LT complexity basis and testing functions exhibits memory, while for 3-D boundary element and requires structures the CPU time and memory scale as and , respectively. In order to expedite the convergence of the iterative solver, a new preconditioning strategy that utilizes near-ﬁeld interactions of the MoM impedance matrix organized in the LT basis was proposed and its efﬁciency

evaluated. Validation of the proposed fast solver was provided through its application to the numerical analysis of microstrip antenna arrays and multiconductor interconnect structures.

REFERENCES

[1] W. Wu, A. W. Glisson, and D. Kajfez, “A comparison of two low-frequency formulations for the electric ﬁeld integral equation,” in Proc. 10th Annu. Review of Progress in Applied Computational Electromagnetics, vol. 2, Monterey, CA, Mar. 1994, pp. 484–491.

2140

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[2] M. Burton and S. Kashyap, “A study of a recent, moment-method algorithm that is accurate to very low frequencies,” Appl. Computational Electromagn. Soc. J., vol. 10, no. 3, pp. 58–68, Nov. 1995. [3] B. Young, Digital Signal Integrity. New York: McGraw-Hill, 2001. [4] R. R. Tummala, Fundamentals of Microsystems Packaging. Englewood Cliffs, NJ: Prentice-Hall, 2001. [5] T. J. Cui and W. C. Chew et al., “Fast forward and inverse methods for buried objects,” in Fast and Efﬁcient Algorithms in Computational Electromagnetics, W. C. Chew et al., Eds. Boston, MA: Artech House, 2001, pp. 347–424. [6] S. Chen and W. C. Chew, “Low-frequency scattering from penetrable bodies,” in Fast and Efﬁcient Algorithms in Computational Electromagnetics, W. C. Chew, Ed. Boston, MA: Artech House, 2001, pp. 425–460. [7] R. Coifman, V. Rokhlin, and S. Wandzura, “The fast multipole method for the wave equation: A pedestrian prescription,” IEEE Antennas Propagat. Mag., vol. 35, pp. 7–12, June 1993. [8] J. M. Song, C. C. Lu, and W. C. Chew, “MLFMA for electromagnetic scattering from large complex objects,” IEEE Trans. Antennas Propagat., vol. 45, pp. 1488–1493, Oct. 1997. [9] J. R. Phillips and J. White, “A precorrected-FFT method for capacitance extraction of complicated 3-D structures,” in Proc. Int. Conf. ComputerAided Design, Santa Clara, CA, Nov. 1994. [10] E. Bleszynski, M. Bleszynski, and T. Jaroszewicz, “AIM: Adaptive integral method for solving large-scale electromagnetic scattering and radiation problems,” Radio Science, vol. 31, pp. 1255–1251, Sept.–Oct. 1996. [11] J. R. Phillips and J. K. White, “A precorrected-FFT method for electrostatic analysis of complicated 3-D structures,” IEEE Trans. ComputerAided Design, vol. 16, pp. 1059–1072, Oct. 1997. [12] F. Ling and J.-M. Jin et al., “Full-wave analysis of multilayer microstrip problems,” in Fast and Efﬁcient Algorithms in Computational Electromagnetics, W. C. Chew et al., Eds. Boston, MA: Artech House, 2001, pp. 729–772. [13] M. F. Catedra, R. P. Torres, J. Basterrechea, and E. Gago, The CG-FFT Method—Application of Signal Processing Techniques to Electromagnetics. Norwood, MA: Artech House, 1995. [14] N. N. Bojarski, “k -Space Formulation of the Electromagnetic Scattering Problem,” Air Force Avionics Lab., Tech. Report, AFAL-TR-71-75, Mar. 1971. [15] S. Gedney, A. Zhu, W.-H. Tang, and P. Petre, “High-order pre-corrected FFT solution for electromagnetic scattering,” in Proc. IEEE Antennas and Propagation Int. Symp. Digest, vol. 3, San Antonio, TX, June 2002, pp. 566–569. [16] J. Zhao and W. C. Chew et al., “Multilevel fast multipole algorithm at very low frequencies,” in Fast and Efﬁcient Algorithms in Computational Electromagnetics, W. C. Chew et al., Eds. Boston, MA: Artech House, 2001, pp. 151–202. [17] , “Integral equation solution of Maxwell’s equations from zero frequency to microwave frequencies,” IEEE Trans. Antennas Propagat., vol. 48, pp. 1635–1645, Oct. 2000. [18] S. Ohnuki and W. C. Chew, “Numerical accuracy of multipole expansion for 2-D MLFMA,” IEEE Trans. Antennas Propagat., vol. 51, pp. 1883–1890, Aug. 2003. [19] H. Contopanagos, B. Dembart, M. Epton, J. J. Ottusch, V. Rokhlin, J. L. Visher, and S. M. Wandzura, “Well-conditioned boundary integral equations for three-dimensional electromagnetic scattering,” IEEE Trans. Antennas Propagat., vol. 50, pp. 1824–1830, Dec. 2002. [20] G. X. Fan and Q. H. Liu, “The CGFFT method with a discontinuous FFT algorithm,” Microwave Opt. Technol. Lett., vol. 29, no. 1, pp. 47–49, 2001. [21] S. S. M. Rao, D. R. Wilton, and A. W. Glisson, “Electromagnetic scattering by surfaces of arbitrary shapes,” IEEE Trans. Antennas Propagat., vol. 30, pp. 409–418, May 1982. [22] G. V. Eleftheriades and R. Mosig, “On the network characterization of planar passive circuits using the method of moments,” IEEE Trans. Microwave Theory Tech., vol. 44, pp. 438–445, Mar. 1996. [23] K. A. Michalski and D. Zheng, “Electromagnetic scattering and radiation by surfaces of arbitrary shape in layered media, part I: Theory,” IEEE Trans. Antennas Propagat., vol. 38, pp. 335–344, Mar. 1990. [24] C. H. Smith, A. F. Peterson, and R. Mitra, “The biconjugate gradient method for electromagnetic scattering,” IEEE Trans. Antennas Propagat., vol. 38, pp. 938–940, June 1990. [25] V. Fraysse, L. Giraud, and S. Gratton. A Set of GMRES Routines for Real and Complex Arithmetics. [Online]. Available: www.cerfacs.fr [26] O. Axelsson, Iterative Solution Methods. Cambridge, MA: Cambridge Univ. Press, 1994.

[27] S. Pissanetsky, Sparse Matrix Technology, London, U.K.: Academic Press, 1984. [28] A. Georgem and J. W.-H. Liu, Computer Solution of Large Positive Definite Systems, London, U.K.: Prentice-Hall, 1981. [29] The Sonnet User’s Manual, Apr. 1999. [30] V. I. Okhmatovski, J. Morsey, and A. C. Cangellaris, “On de-embedding of port discontinuities in full-wave CAD models of multiport circuits,” IEEE Trans. Microwave Theory Tech., vol. 51, Dec. 2003. (in print).

Vladimir I. Okhmatovski (M’99) was born in Moscow, Russia, in 1974. He received the M.S. (with distinction) and Candidate of Science (Ph.D.) degrees from the Moscow Power Engineering Institute, Moscow, Russia, in 1996 and 1997, respectively. In 1997, he joined the Radio Engineering Department, Moscow Power Engineering Institute, as an Assistant Professor. From 1998 to 1999, he was a Postdoctoral Fellow in the Microwave Laboratory, National Technical University of Athens, Greece. From 1999 to 2003, he was a Postdoctoral Research Associate with the Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign. He is currently with the Department of CIC Advanced R&D, Cadence Design Systems, Tempe, AZ. He has authored and coauthored over 30 papers in professional journals and conference proceedings. His research interests include modeling of high-speed interconnects, fast algorithms in computational electromagnetics, geometrical and physical theories of diffraction and conformal antennas and arrays. Dr. Okhmatovski was the recipient of a 1995 scholarship of the Government of the Russian Federation, a 1996 Presidential Scholarship of the Russian Federation, and a 1997–2000 scholarship of the Russian Academy of Science. In 1996, he received Second Prize for the Best Young Scientist Report presented at the VI International Conference on Mathematical Methods in Electromagnetic Theory (MMET’96). He was also the recipient of the Outstanding Technical Paper Award at the 3rd Electronics Packaging Technology Conference (EPTC 2000).

Jason D. Morsey (S’01–M’03) received the B.S. (cum laude) and M.S. degrees in electrical engineering from Clemson University, Clemson, SC, in 1998 and 2000, respectively, and the Ph.D. degree from the University of Illinois at Urbana-Champaign, in 2003. He is currently with the IBM T. J. Watson Research Center, NY. His research interests include electromagnetic modeling of high-speed interconnects at all levels of integration and their signal integrity analysis.

Andreas C. Cangellaris (M’86–SM’97–F’00) received the M.S. and Ph.D. degrees in electrical and computer engineering from the University of California, Berkeley, in 1983 and 1985, respectively. From 1985 to 1987, he was a Senior Research Engineer in the Electronics Department, General Motors Research Laboratories, Warren, MI. From 1987 to 1992, he was an Assistant Professor on the faculty of Electrical and Computer Engineering at the University of Arizona, Tuscon, and then an Associate Professor from 1992 to 1997. He is currently a Professor of Electrical and Computer Engineering at the University of Illinois at Urbana-Champaign (UIUC). Over the past 15 years, he has supervised the development of numerous electromagnetic modeling methodologies and computer-aided design tools for high-speed/high-frequency signal integrity-driven applications, which have been transferred successfully to industry. He has coauthored more than 150 refereed papers and three book chapters on topics related to computational electromagnetics and interconnects and package modeling and simulation. His research work has been in the area of applied and computational electromagnetics with emphasis on their application to electrical modeling and simulation of RF/microwave components and systems, high-speed digital interconnects at the board, package, and chip level, as well as the modeling and simulation of electromagnetic compatibility and electromagnetic interference. Prof. Cangellaris is an active Member of the IEEE Microwave Theory and Techniques Society, the IEEE Components Packaging and Manufacturing Technology Society, the IEEE Antennas and Propagation Society, and the IEEE Magnetics Society, serving as a Member of technical program committees for major conferences and symposia sponsored by these societies. He has served as Associate Editor for the IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, and is currently serving as Associate Editor of the IEEE TRANSACTIONS ON ADVANCED PACKAGING and the IEEE Press Series on Electromagnetic Fields and Waves. He is the co-founder of the IEEE Topical Meeting on Electrical Performance of Electronic Packaging.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2141

**A Single-Level Low Rank IE-QR Algorithm for PEC Scattering Problems Using EFIE Formulation
**

Seung Mo Seo, Student Member, IEEE, and Jin-Fa Lee, Senior Member, IEEE

Abstract—This paper presents a single-level matrix compression algorithm, termed IE-QR, based on a low-rank approximation to speed up the electric ﬁeld integral equation (EFIE) formulation. It is shown, with the number of groups chosen to be proportional is the number of unknowns, the memory and to 1 2 , where CPU time for the resulting algorithm are both ( 1 5 ). The unique features of the algorithm are: a. The IE-QR algorithm is based on the near-rank-deﬁciency property for well-separated groups. This near-rank-deﬁciency assumption holds true for many integral equation methods such as Laplacian, radiation, and scattering problems in electromagnetics (EM). The same algorithm can be adapted to other applications outside EM with few or no modiﬁcations; and, b. The rank estimation is achieved by a dual-rank process, which ranks the transmitting and receiving groups, respectively. Thus, the IE-QR algorithm can achieve matrix compression without assembling the entire system matrix. Also, a “geometric-neighboring” preconditioner is presented in this paper. This “geometric-neighboring” preconditioner when used in conjunction with GMRES is proven to be both efﬁcient and effective for solving the compressed matrix equations. Index Terms—Electromagnetic (EM) scattering, low-rank QR factorization, method of moments (MoM).

I. INTRODUCTION

M

ETHOD of moments (MoM) using the electric ﬁeld integral equation (EFIE) formulation has been a very popular choice for solving electromagnetic scattering problems by perfect electric conducting (PEC) objects. However, it is well known that the traditional MoM suffers from the storage of a dense impedance matrix and computational complexity for large-scale problems. However, signiﬁcant progress has been made in using the fast multipole method (FMM) [1] to overcome these difﬁculties. The single-level FMM combined with the iterative techniques has reduced the numerical complexity to solve dense integral equation matrices that arise to from the Helmholtz equation. One major drawback of this approach is its dependence on the integral equation kernel. An) algorithm presented in other approach is the so called ( [2] and [3], which are based on low-rank representation of maapproach [2] is based on the FMM trix blocks. The basic idea without the closed-form formula. In [2], the matrix compression is employed with the explicit formulation of the matrix computablocks, which then results in an undesirable tional complexity. Thus, the basic algorithm has been ﬁne tuned in [3] through partial assembling of the matrix blocks. However,

Manuscript received March 17, 2003; revised September 16, 2003. The authors are with the ElectroScience Laboratory, Department of Electrical and Computer Engineering, The Ohio State University, Columbus, OH 432121191 USA (e-mail: seo.38@osu.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832367

the approach suggested in [3] was based on a statically-determined map which was used to assemble a reduced impedance matrix through interpolation. The exact procedure of the interpolation scheme adapted is not at all clear. In the present approach proposed here in this paper, we have made two distinct differences: 1) Our QR decomposition procedure is based on a dual-rank process which “ranks” both the transmitter and the receivers and 2) the convergence of the QR process is determined by both the orthogonal projection as well as the estimate error matrix norm smaller than the speciﬁed tolerance. This paper presents a low-rank IE-QR algorithm for efﬁciently compress the MoM matrix to reduce the memory requirement, matrix ﬁlling time, and the time of the iterative so. The single-level low-rank IE-QR algorithm lution to is based on the rank deﬁciency feature of the integral equation for well-separated groups of basis functions. The algorithm forms a low-rank QR factorization of a matrix block, nonself interaction and nontouching groups, with only a portion of it being formed. Each group has a bounding box from binary partition. The touching groups mean that the bounding boxes of two groups are overlapped. As a matter of fact, the total number of entries being computed for a matrix block, assuming the , is , where is the numermatrix dimension ically determined rank of the matrix block. The entire process can be viewed as the classical rank-revealing QR factorization using modiﬁed Gram-Schmidt (MGS) with partial pivoting. The rest of the paper is organized as follows: Section II gives a description of EFIE formulation for scattering problems; The single-level low-rank IE-QR algorithm is given in Section III; Section IV then details the “geometric-neighboring” preconditioner; and one open cone plate example is shown in Section V to validate and demonstrate the performance of the current approach. Moreover, we investigate the performance of the singlelevel low-rank IE-QR algorithm; and ﬁnally, in Section VI, we provide a brief summary. II. EFIE FORMULATION FOR SCATTERING PROBLEMS In this paper, we employ an integral equation for the surface current induced on a perfect conducting scatterer [4]. Application of the Galerkin method to the electric ﬁeld integral equation results in

(1)

0018-926X/04$20.00 © 2004 IEEE

2142

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

where is the problem domain or the scatterer, is the free space wavenumber, is the distance between observation and is the characteristic wave impedance in source points, and free space. In the current paper, we have employed surface divconformal vector basis functions for the surface current, namely the Rao–Wilton–Glisson (RWG) basis functions. III. SINGLE-LEVEL LOW-RANK IE-QR ALGORITHM The single-level low-rank IE-QR algorithm is based on the rank deﬁciency feature of the integral equation for well-separated groups of basis functions. The algorithm factorizes the due to the group and into local impedance matrix and matrices without a priori knowledge of , and are the number of receiving and transmitting where basis functions, respectively, and is the rank of interaction

For the detail of the MGS process, see [9]. Deﬁnition 2: (Orthogonal Component): Given an or, and a column thonormal matrix vector , we deﬁne by (5) It is clear that Deﬁnition 3: We say that a matrix if imate of . is an approx-

(6) . and we write rectDeﬁnition 4: (Column Index Selection): Given a , and assuming angular matrix ( ) exists, we deﬁne as the minimum index , , such that (7)

. . .

(2)

Here, is the column vector due to the th basis function in is the row vector the transmitting group (transmitter) and due to the th basis function in the receiving group (receiver). matrix is stored in the maThe information of the global trices and where ; , and the , , which are comself-impedance matrices puted directly from the integral equation formulation, where is the number of groups of basis functions. Also, the geometrical touching groups, in which the rank deﬁcient aspect cannot be preserved, are directly computed. The computing IE-QR of the touching groups has expensive computational complexity compared with the direct computation. The single-level low-rank IE-QR algorithm is used to genthat approximately span the erate orthogonal columns of and the columns of column space of the original matrix that are the expansion coefﬁcients of the corresponding columns of with respect to the column vectors in . columns of are obtained by solving the The rest equations extracted from (2). The important advantage of the single-level low-rank IE-QR algorithm is that it does not require . Storing and using and a priori knowledge of instead of helps both to save the required memory for each group interaction and to reduce the numerical complexity of both the matrix assembly and matrix-vector multiplication in the iterative solver. The detail of the algorithm can be shown below. A. Deﬁnitions and Notations Needed for the IE-QR Algorithm Deﬁnition 1: (MGS): Given linearly independent column , we denote the orthonormal matrix vectors, which is obtained by these vectors through the MGS process. Namely (3) and

Deﬁnition 5: (Row Index Selection): Given a gular matrix and its row partition according to . . .

rectan, we deﬁne

(8) Norm): Given a Deﬁnition 6: (Approximate , and a series of indexes angular matrix deﬁne an approximate norm of by rect, we

(9)

B. A Single-Level Low-Rank IE-QR Algorithm With Matrix Block Partially Formed The following is a straightforward theorem, so we just simply state it. and its column partition Theorem 1: Given a matrix , the matrix product with (10), shown at the bottom of the next page, is an approximate of , assuming is nonsingular. . Moreover, note that the That is approximate is error for the

. . . . . .

..

.

. . . . . .

. . .

..

.

. . .

..

.

(4)

(11)

SEO AND LEE: SINGLE-LEVEL LOW RANK IE-QR ALGORITHM FOR PEC SCATTERING

2143

Now we are ready to state our IE-QR algorithm with the matrix blocks partially formed. Suppose at the th step, we have approximate of the permuted explicitly formed a matrix, namely

. . . . . . .. . . . . . . ..

. . . . . . (12)

.

.

Fig. 1. Geometry of the open cone scatterer, with the height of 20 cm and the base diameter of 20 cm. (a) Surface triangulation and (b) mesh partitioning.

In (12), the entries marked by are available, i.e., have entries are not. Moreover, an been computed, whereas the orthonormal matrix , with dimensions , steps. Also, in has been constructed through previous and are row and column permutation matrices, (12), respectively. The detail of these permutation matrices will be described in the following algorithm. Algorithm: IE-QR Algorithm with Matrix Block Partially Formed Step 1) Let , we compute . Then the permutation matrix is an identity with and columns interchanges. to Step 2) Update

IV. GEOMETRIC-NEIGHBORING PRECONDITIONER In this section, we consider the efﬁcient solution of dense by preconditioned iterative linear system methods, particularly GMRES method. An insightful discussion of three types of preconditioners, the operator splitting preconditioner (OSP), the least squares approximate inverse preconditioner (LSAI), and the diagonal block approximate inverse preconditioner (DBAI) [5], for dense matrices arising from the application of BIE is provided. Our approach contains the idea of the mesh neighbor (MN) preconditioner in [6] and DBAI. The “mesh-neighboring” preconditioner proposed in the current paper, is based upon a two-step process [7]. In the ﬁrst step, we extract from the full impedance matrix, , a sparse version, , which includes the near range interactions as well as a heuristic bias toward geometrical , is obtained, the ﬁnal singularities. Once the sparse matrix, , will be formed through an incomplete preconditioner, factorization with a heuristic dropping strategy [8]. The detail of the preconditioner can be found in [7]. V. NUMERICAL RESULTS

. . . . . . .. . . . . . . ..

. . . . . . (13)

.

.

Let ,

, then compute , and

.

Update from to . , and the row Step 3) Compute is a identity with permutation matrix and rows interchanged. Update to . , where is a prescribed tolerance, Step 4) If according to (10). Find inthen compute , and compute dexes, . If , stop the procedure. Otherwise, continue the IE-QR process.

To demonstrate the efﬁcient and validate the current singlelevel IE-QR approach, we have conducted studies on one numerical example. In the example, we employed constant (as constant as we possibly can) mesh density while increase the operating frequency. A. An Open Cone Plate An open cone PEC scatterer is shown as inset in Fig. 1, whose height and diameter of the bottom are 20 cm. From Fig. 1(b), we clearly establish the open cone PEC scatterer is uniformly partitioned using a simple mesh-partitioning algorithm. The partitioned groups are well-separated for the single-level low-rank IE-QR algorithm. The rank map for the open cone mesh at

(10)

2144

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2.

Rank map of the open cone example at 5 GHz (N = 5;280).

5 GHz is shown in Fig. 2. The number of unknowns is 5820 and the maximum and minimum sizes of groups are 171 and 73, respectively. The black colored boxes represent the dense matrices (full MoM matrix assembling), which are for self-groups and touching (or overlapping) groups. The maximum and minimum ranks of all the coupling matrices of nonoverlapping groups are 19 and 7, respectively. Therefore, the use of a low-rank representation results in signiﬁcant CPU and memory reductions. The radiation patterns computed by the single-level low-rank IE-QR algorithm are plotted in Fig. 3 along with the results from full EFIE code. The 41 % of memory compared to full EFIE code is used. The approximation in the single-level low-rank IE-QR process, using a tolerance of 0.01, does not affect the solution quality, as evidenced in Fig. 3. The radiation patterns computed by the single-level IE-QR algorithm at 8 GHz are shown in Fig. 4. We see our results agree well with the results of full EFIE matrix.

Fig. 3. Monostatic RCS patterns of the open cone example at 5 GHz.

B. Performance of the Single-Level IE-QR Algorithm To study the memory and CPU time complexities of the single-level low-rank IE-QR algorithm, we increase the operating frequencies and subsequently enlarge the problem size. The memory consumption and CPU time of the IE-QR process are shown in Figs. 5 and 6, respectively, and they both exhibit an complexity for small to moderate electrical size

SEO AND LEE: SINGLE-LEVEL LOW RANK IE-QR ALGORITHM FOR PEC SCATTERING

2145

TABLE I PERFORMANCE OF SINGLE-LEVEL LOW-RANK IE-QR ALGORITHM WITH THE OPEN CONE PLATE (tolerance = 10 )

TABLE II PERFORMANCE OF SINGLE-LEVEL LOW-RANK IE-QR ALGORITHM WITH THE OPEN CONE PLATE (tolerance = 10 ) Fig. 4. Monostatic RCS patterns of the open cone example at 8 GHz.

problems. Finally, in Tables I and II, we summarize the computational details of the application of the proposed single-level IE-QR algorithm to the open cone example with the tolerance and , respectively. The computations were done of on Pentium II 400 MHz.

Fig. 5. Plot of the memory consumption for the open cone example. Note that ) line. the reference is the (

ON

VI. SUMMARY This paper presents the novel single-level low-rank IE-QR algorithm. The algorithm proves memory consumption and CPU time are reduced signiﬁcantly. REFERENCES

[1] R. Coifman, V. Rokhlin, and S. Wandzura, “The fast multipole method for the wave equation: A Pedestrian prescription,” IEEE Trans. Antennas Propagat. Mag., vol. 35, pp. 7–12, June 1993. [2] S. Kapur and J. Zhao, “A fast method of moments solver for efﬁcient parameter extraction of MCMs,” in Proc. 34th Design Automation Conf., vol. 39, June 1997, pp. 141–146. [3] S. Kapur and D. E. Long, “IES : A fast integral equation solver for efﬁcient 3-dimensional extraction,” in Proc. 37th Int. Conf. Computer Aided Design, Nov. 1997. [4] S. M. Rao, D. R. Wilton, and A. W. Glisson, “Electromagnetic scattering by surfaces of arbitrary shape,” IEEE Trans. Antennas Propagat., vol. AP-30, pp. 409–418, May 1982. [5] K. Chen, “An analysis of sparse approximate inverse preconditioners for boundary integral equations,” SIAM J. Matrix Anal. Appl., vol. 22, pp. 1058–1078, 2001. [6] S. Vavasis, “Preconditioning for boundary integral equations,” SIAM J. Matrix Anal. Appl., vol. 13, pp. 905–925, 1992. [7] J. F. Lee, R. Lee, and R. Burkholder, “Loop star basis functions and a robust preconditioner for EFIE scattering problems,” IEEE Trans. Antennas Propagat., vol. 51, pp. 1855–1863, Aug. 2003.

Fig. 6. Plot of the CPU time for the open cone example. Note that the reference is the ( ) line.

ON

2146

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

[8] I. Gustafson, “Modiﬁed incomplete Choleski (MIC) methods,” in Preconditioning Methods: Analysis and Applications, D. J. Evans, Ed. New York: Gordon and Breach, 1983, pp. 265–293. [9] G. H. Golub and C. F. Van Loan, Matrix Computations. Baltimore and London: The Johns Hopkins Univ. Press, 1996, pp. 223–236.

Seung Mo Seo (S’00) was born in Seoul, Korea. He received the B.S. degree in electrical engineering from Hong-Ik University, Seoul, in 1998 and the M.S. degree from The Ohio State University, Columbus, in 2001, where he is currently working toward the Ph.D. degree in electrical engineering. From 1999 to the present, he has been a Graduate Research Associate with the ElectroScience Laboratory, Department of Electrical and Computer Engineering, The Ohio State University, where he focused on development of fast integral equation methods.

Jin-Fa Lee (SM’99) received the B.S. degree from National Taiwan University, Taiwan, R.O.C., in 1982 and the M.S. and Ph.D. degrees from Carnegie-Mellon University, Pittsburg, PA, in 1986 and 1989, respectively, all in electrical engineering. From 1988 to 1990, he was with ANSOFT Corporation, where he developed several CAD/CAE ﬁnite element programs for modeling three-dimensional microwave and millimeter-wave circuits. His Ph.D studies resulted in the ﬁrst commercial three-dimensional FEM package for modeling RF/Microwave components, HFSS. From 1990 to 1991, he was a Postdoctoral Fellow at the University of Illinois at Urbana-Champaign. From 1991 to 2000, he was with the Department of Electrical and Computer Engineering, Worcester Polytechnic Institute, Worcester, MA. Currently, he is an Associate Professor in the ElectroScience Laboratory, Department of Electrical and Computer Engineering, The Ohio State University, Columbus. His current research interests are analyzes of numerical methods, fast ﬁnite element methods, integral equation methods, hybrid methods, three-dimensional mesh generation, domain decomposition methods, and mortar ﬁnite elements.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2147

**Accelerated Gradient Based Optimization Using Adjoint Sensitivities
**

Natalia K. Nikolova, Member, IEEE, Reza Saﬁan, Ezzeldin A. Soliman, Associate Member, IEEE, Mohamed H. Bakr, Member, IEEE, and John W. Bandler, Fellow, IEEE

Abstract—An electromagnetic feasible adjoint sensitivity technique (EM-FAST) has been proposed recently for use with frequency-domain solvers . It makes the implementation of the adjoint variable approach to design sensitivity analysis straightforward while preserving the accuracy at a level comparable to that of the exact sensitivities. The overhead computations associated with the estimation of the sensitivities in addition to the system analysis are due largely to the calculation of the derivatives of the system matrix. Here, we describe the integration of the EM-FAST with two methods for accelerated estimation of these derivatives: the boundary-layer concept and the Broyden update. We show that the Broyden update approach (Broyden-FAST) leads to an algorithm whose efﬁciency is problem independent and allows the computation of the response and its gradient through a single system analysis with practically no overhead. Both approaches are illustrated through the design of simple antennas using method of moments solvers. Index Terms—Adjoint sensitivities, antenna design, Broyden update, design methodology, method of moments (MoM), optimization, sensitivity.

I. INTRODUCTION

T

RADITIONAL full-wave electromagnetic (EM) solvers do not compute the gradient of the response (e.g., -parameters, input impedance or antenna gain) with respect to the design parameters, which relate to the geometry and the materials of the structure. Commercial high-frequency CAD software typically resorts to ﬁnite-difference approximations of the response sensitivities, which are numerically inefﬁcient but simple to implement with existing EM solvers. To compute the response and its sensitivities, such an approach requires full-wave analyses, being the number a minimum of of the design parameters. This approach is also known as the perturbation approximate sensitivity technique (PAST) [1]. Higher-order approximations may also be used at the expense of an increased number of simulations. They are feasible when

Manuscript received December 12, 2002; revised September 30, 2003. This work was supported in part by the Natural Sciences and Engineering Research Council of Canada under Grants OGP0227660-03, OGP0007239, OGP0249780-02, STR234854-00, through the Micronet Network of Centres of Excellence and Bandler Corporation. N. K. Nikolova, R. Saﬁan, E. A. Soliman, and M. H. Bakr are with the Department of Electrical and Computer Engineering, McMaster University, Hamilton, ON L8S 4K1, Canada. J. W. Bandler is with with the Department of Electrical and Computer Engineering, McMaster University, Hamilton, ON L8S 4K1, Canada and also with Bandler Corporation, Dundas, ON L9H 5E7, Canada. Digital Object Identiﬁer 10.1109/TAP.2004.832313

sufﬁcient database for the system response in the design parameter space becomes available. Such response data, for example, would gradually accumulate during optimization. It is possible to derive exact sensitivity expressions for the state variables of a system by directly differentiating its equations with respect to the desired design parameters. For example, in [2], a mixed potential integral equation is developed for the current density derivatives with the method of moments (MoM) applied to planar multilayer structures. This equation, when solved together with the original electric ﬁeld integral equation, yields both the currents and their derivatives with respect to the design parameters. Such an approach—generally referred to as sensitivity analysis via direct differentiation—can be applied to both steady-state [3], [4] and dynamic [5] systems. For each design parameter, an additional linear system analysis is required to obtain the respective response derivative. Each of these analyses is characterized by the same system matrix, which is also identical with the original system matrix. A more efﬁcient design sensitivity analysis is provided by the adjoint variable method [3]–[7]. It reduces the computational overhead of the sensitivity computation to just one additional linear system analysis where the system matrix is the transpose of that of the original problem. Thus, its computational overhead is times smaller than that of the direct differentiation approach and is practically independent of the number of design parameters . Adjoint-based design sensitivity analysis of microwave structures has been ﬁrst formulated in terms of circuit concepts rather than ﬁeld concepts, and it is referred to as the adjoint network method [7]–[12]. To obtain exact sensitivities, both the direct differentiation and the adjoint-variable techniques require the analytical derivatives of the system matrix with respect to the design parameters. This constitutes a major difﬁculty in applications with full-wave EM solvers for research or commercial design software. Recently, adjoint variable approaches were used with the ﬁnite-element method (FEM); see, for example, [13]–[15]. The FEM is well suited for exact sensitivity calculations because of the analytical relation between the coefﬁcients of the FEM matrix and the coordinates of the vertices of the ﬁnite element grid. This analytical relation, however, is not trivial. Its implementation in the computation of the derivatives of the FEM system matrix with respect to any geometrical or material design parameter is in practice difﬁcult and, to our knowledge, has not been exploited yet in commercial high-frequency CAD software. A similar difﬁculty exists with the exact sensitivities for the MoM. The different varieties of MoM techniques rely on speciﬁc Green’s functions, as well as different basis and weighting

0018-926X/04$20.00 © 2004 IEEE

2148

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

functions. The dependence of the system matrix coefﬁcients on possible geometry perturbations is involved and case speciﬁc [2], [16]. In summary, exact sensitivities appear to be often impractical in full-wave EM analysis for two reasons: (1) the analytical preprocessing is involved and solver speciﬁc; (2) the implementation requires thorough reworking of the analysis engine. The second requirement is especially unattractive in the development of commercial software. Recently, a feasible adjoint-sensitivity technique (FAST) for applications with full-wave EM solvers (EM-FAST) has been proposed [17]. It uses ﬁnite differences to approximate the derivatives of the system matrix. Its implementation in a versatile CAD environment is straightforward since it requires minor additions to existing frequency-domain computational algorithms. Its accuracy is comparable to that of the analytical exact sensitivities. Its overhead is mostly due to the ﬁnite-difference computation of the derivatives of the system matrix, and it is equivalent to that of the exact sensitivity calculations. Here, we propose the use of two techniques—the boundary-layer concept and the Broyden update—to enhance the speed of the EM-FAST, which is crucial in gradient-based optimization. There is a certain loss of accuracy; however, the approximated sensitivities are sufﬁciently accurate to efﬁciently guide the optimization toward the optimal design. In applications requiring higher accuracy of the response gradient such as tolerance or yield analysis, the original EM-FAST may be preferable. We start with a brief outline of the EM-FAST [17] and its computational requirements. We then discuss ways to accelerate its performance through the boundary layer concept (BLC) and the Broyden update. The resulting algorithms offer signiﬁcant CPU time reduction in comparison with the original EM-FAST on the order of the number of design parameters . The savings in comparison with the traditional ﬁnite-difference gradient approximation applied directly to the set of responses (e.g., PAST) are drastic, especially in the case of multiple design variables. II. FEASIBLE ADJOINT SENSITIVITY ANALYSIS A. Deﬁnitions and Notations in Adjoint Sensitivity Analysis Consider the system of complex-valued equations arising form the discretization of a linear EM problem (1) where, is the vector of design parameters. These parameters typically have real values related to the geis ometry and the materials of the structure. the state variable vector, e.g., complex-valued current distribuis the excitation vector; is the system tion in the MoM; matrix whose complex coefﬁcients depend on the geometry and materials. The objective of sensitivity analysis is to determine the graat the dient of a properly deﬁned response function current solution of (1) with respect to the design parameters (2)

We assume that the response is a scalar function, which is differentiable in and . We deﬁne the gradient operator as a row operator (3) When the gradient operator acts on a vector, e.g., is a matrix . . . . . . , the result

(4)

The optimization problem is formulated as (5) where is the objective function to be minimized, and is the vector of optimal design parameters. Gradient-based optimizers require both the response of the current design and its sensitivity (3) in order to predict the next design iterate. The sensitivities of the objective function are obtained as [18]

(6) where is the solution of the complex adjoint problem (7) in which the adjoint excitation is deﬁned by . . .

(8)

Here, and denote the real and imaginary parts, respectively, reﬂects the explicit of a complex variable. The gradient dependence of on . The matrix would typically be analytically available. In fact, the excitation is often insensitive to . For example, changes in the design parameters, i.e., in a microstrip circuit, the excitation is deﬁned at ports located at feed lines. If the design parameter variations affect neither the dielectric constant nor the height of the substrate, nor the width of the feed line, the excitation remains unchanged. is a constant vector representing the solution at In the current design, i.e., (6) can be written explicitly as

(9) The sensitivity expression (6) is a generalization of the wellknown, linear, real-system sensitivity formula [3], [17]. As evident from (6) and (7), the adjoint approach provides the gradient of the response with respect to all design parameters with just one additional system analysis (7) whose system matrix is simply related to that of the original problem (1). When factorization of is used to solve (1), the factors of

NIKOLOVA et al.: ACCELERATED GRADIENT BASED OPTIMIZATION USING ADJOINT SENSITIVITIES

2149

are easily obtained by rearranging the factors of . Thus, the additional system analysis (7) is practically avoided, the overhead being due only to the forward-backward substitutions. In the case of iterative solvers—often used when is large and/or sparse—a complete additional system analysis seems imperative. B. The Feasible Technique , which we refer to as The matrices derivative matrices, may be analytically available, as is the case with the FEM. Then the sensitivities obtained with (9) are exact. matrix at The calculation of an analytically available the current design is computationally equivalent to a -matrix ﬁll; therefore, at each design iteration, the equivalent of matrix ﬁlls is needed. Thus, the advantage of analytically available derivative matrices is in the accuracy of the derivative estimation rather than in its computational efﬁciency. When the factors of the system matrix are available from the analysis of (1), the computation of the derivative matrices determines the overhead associated with the sensitivity analysis since it is far more computationally demanding than the forward-backward substitutions when solving the adjoint problem (7). When the system equations of (1) and (7) are solved iteratively, the additional (adjoint) system analysis determines the computational overhead. Even in this case, the reduction of the time to estimate the derivative matrices is desirable. In full-wave EM analysis usually the derivative matrices , are not analytically available or they are too complicated to obtain for the purposes of general and versatile design software. Then, we can resort to the ﬁnite-difference [17], which requires approximation additional -matrix ﬁlls if forward (or backward) ﬁnite differences are used. The associated computational overhead is equivalent to that of the exact sensitivity estimation discussed above. The important advantage here is that the implementation with existing software is simple. The technique does not require any analytical preprocessing, which often restricts the versatility of the algorithm. We have investigated the accuracy of the sensitivity estimation with the feasible adjoint technique [17] and we have found between that it is excellent for relative perturbations 0.5% and 2%. The relative error in comparison with the exact sensitivities is well below 1% for a broad range of values of the design parameters, close to or far from the nulls of the sensitivity curves. A detailed comparison between the computational requirements of the EM-FAST and the commonly used ﬁnite differences applied directly to the response is made in [17]. Here, we only note that the EM-FAST reduces the number of required being the number of full-wave analyses by a factor of design parameters. However, there are overhead computations associated with the additional matrix ﬁlls in order to compute . In certain cases, e.g., electrically small problems, the MoM matrix ﬁll may account for a signiﬁcant portion of the CPU time required by the overall analysis (matrix ﬁll plus linear system solution). Such an overhead should not be overlooked in a sequence of repetitive analyses performed during optimization.

III. ACCELERATED OPTIMIZATION WITH APPROXIMATED ADJOINT SENSITIVITIES There are two techniques which can lead to faster calculation of the derivative matrices. The ﬁrst one is the boundary-layer concept (BLC) ﬁrst proposed by Amari [19] in the sensitivity analysis with the direct differentiation method. The acceleration offered by the BLC depends on the relation between the respective design parameter and the geometry of the structure as we explain below. Its computational requirements are dependent on the number of design parameters . It requires modiﬁcations of existing EM analysis software, which relate to meshing and matrix building subroutines. Its advantage is that it yields sensitivity estimates of very good accuracy. The second approach uses Broyden’s update to iteratively compute approximate derivative matrices. This approach reduces the overhead drastically since its computational requirements—negligible compared to a matrix ﬁll—practically do not depend on . The Broyden-update approach does not require any modiﬁcations of the EM analysis algorithms. A. BLC With the EM-FAST The BLC can be applied with solvers which allow nonuniform discretization and/or unstructured grids, e.g., the FEM and the MoM. The idea is to perturb a certain geometrical parameter (the design parameter ) of a structure by respective deformations of as few grid elements as possible. This makes most of the -matrix coefﬁcients insensitive to the perturbais mostly tion. Consequently, the matrix derivative sparse and only few nonzero coefﬁcients need to be calculated. This is in contrast with the conventional EM-FAST where full remeshing is applied to the perturbed structure, which results matrix. in a full We present two examples, which illustrate the BLC. Through them, we investigate the accuracy of the modiﬁed EM-FAST algorithm which exploits the BLC. 1) A Dipole of Finite Thickness: We analyze the sensitivity of a dipole with respect to the norof the input impedance malized length of the dipole . The dipole is discretized into segments whose normalized length is uniform and [see Fig. 1(a)]. Here, is the number equal to . This example is suitof segments. In this example, able for design sensitivity tests because the input impedance of a dipole is highly sensitive to its length, especially close to resonance. The thickness of the dipole is represented by the radius of its cross-section, which is constant and set to . The derivatives and are calcuand . We use the symlated, where metry of the structure and analyze half of it. The analysis algorithm is based on Pocklington’s equation, which is discretized using pulse basis functions and a point-matching technique [20]. Magnetic frill excitation is applied. Fig. 1(b) shows the perturbed geometry corresponding to a at the th design iteration where only change of length the boundary-layer (edge) segments are changed accordingly. has only one row and The resulting derivative matrix one column of nonzero elements. Fig. 1(c) shows the same parameter perturbation this time realized with the conventional

2150

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 2. Derivative of the input resistance of the dipole with respect to its normalized length.

Fig. 1. Perturbing the length of the dipole at the k th iteration with and without a boundary layer.

EM-FAST approach. Since the centers of all segments in the perturbed structure change their mutual positions, the matrix is dense. The input impedance sensitivities are calculated in four different ways. First, the forward ﬁnite differences are applied directly to the response

(10)

Fig. 3. Derivative of the input reactance of the dipole with respect to its normalized length.

For each , the MoM solver is invoked twice to perform the analysis at , and at , where . The sensitivity of is evaluated in the range from 0.3 to 1.2 (see Figs. 2 and 3). Second, the input impedance sensitivity is computed with the conventional EM-FAST [17]. The derivative matrix is dense and its coefﬁcients are calculated using forward ﬁnite differences applied to each matrix element . This requires numerical integrations. The increment is again set at . The derivative matrix is then used in (9) to compute and . The resulting sensitivity curves are used as a reference as they are the closest to the exact sensitivities [17]. The third and the fourth derivative estimations use the adis very joint technique with the BLC. The matrix sparse and its computation is fast as it involves only numerical integrations. The perturbations are set so that the length of the edge elements is increased by and for the third and the fourth analysis, respectively. Some accuracy is sacriﬁced as is clear from Figs. 2 and 3; however, it is sufﬁcient for the purposes of gradient-based optimization. The slight deterioration in accuracy is due to the nonuniformity of the segment

size introduced by the edge-only perturbation. We expect such deterioration to be less when higher-order basis and test functions are used such as triangular functions for wire antennas or rooftops for planar structures. We now proceed with the optimization of the dipole for an . The objective function is deinput impedance of ﬁned as (11) We allow two geometrical parameters to vary: the normalized dipole length and the normalized dipole diameter . The vector of design parameters is thus . The following constraints are imposed: (12) since this problem is known to be nonunique. The BLC is used to compute the matrix derivative . Notice, however, that it cannot be exploited in the case of the design parameter because a change in the antenna diameter affects all -matrix matrix is computed with the concoefﬁcients. The ventional EM-FAST technique, which requires a full matrix ﬁll.

NIKOLOVA et al.: ACCELERATED GRADIENT BASED OPTIMIZATION USING ADJOINT SENSITIVITIES

2151

TABLE I OPTIMIZATION OF THE INPUT IMPEDANCE OF THE DIPOLE

Fig. 5.

BLC and the perturbed mesh related to the design parameter x .

Fig. 6. Progress of R

= <Z

of the patch antenna during the optimization.

TABLE II DESIGN PARAMETERS OF THE PATCH AT EACH ITERATION Fig. 4. Geometry of the microstrip-fed patch antenna.

A similar situation would arise in the design of another wire antenna, a Yagi–Uda array. While the BLC is very useful when a design parameter represents the length of a wire, it can offer little or no computational savings if the design parameter is a separation distance between wires. Its efﬁciency is case speciﬁc. where the obThe initial design is . After seven iterations, an jective function is with optimal solution is found at , which corresponds to an input impedance of . The progress of the optimization is summarized in Table I. The gradient-based optimization routine of MATLAB1 fmincon is used. 2) A Microstrip-Fed Patch Antenna: The EM-FAST is also integrated with an in-house MoM solver, which performs analysis of layered structures. The analysis technique is based on the electric ﬁeld integral equation. Here, we show an application with the BLC to the optimization of a microstrip-fed rectangular of the patch are patch antenna. The length and the width optimized for a maximum real input impedance. The objective function to be minimized is deﬁned as (13) The geometry is shown in Fig. 4. The patch is printed on a and height substrate of relative dielectric constant mm. The initial design is given by mm mm. The operating frequency is set at 2 GHz. and

1MATLAB

The BLC is applied as illustrated in Fig. 5. The calculation and is sigof the derivative matrices niﬁcantly faster than one matrix ﬁll. A matrix ﬁll is equivaintegrations, where and show lent to the number of discrete steps along the length and the width of the patch, respectively. On the other hand, the estimations of and with the BLC are equivalent to and numerical integrations, respectively. The progress of the objective function is shown in Fig. 6 in . The changes of the design parameters with each terms of design iteration are listed in Table II. B. The Broyden-Update Approach to Matrix Derivative Estimation The Broyden update is a classical rank-one formula proposed by Broyden [21] for the approximation of the Jacobian of a vector function . If the approximated Jacobian is denoted as at the th iteration, Broyden’s formula is written as

is a registered trademark of The MathWorks, Natick, MA.

(14)

2152

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 7. Geometry of the Yagi–Uda array. TABLE III OPTIMIZATION OF THE INPUT IMPEDANCE OF THE YAGI–UDA ARRAY WITH BROYDEN-FAST

where is the increment vector in the design parameter space. It has elements corresponding to the increment of each design parameter. Broyden’s update has been used in a number of applications such as gradient-based optimization where analytical sensitivities are not available [22], the aggressive space mapping technique [23], etc. We apply Broyden’s update to estimate iteratively the deriva, which are subsequently tive matrices used in the sensitivity expression (9). We refer to this modiﬁed adjoint-based technique as Broyden-FAST. In the implementation of (14), every complex-valued matrix coefﬁcient is a nonlinear function of the design parameters. We deﬁne as a vector which consists of the real and imagas a matrix inary parts of all elements of the matrix, and which consists of their derivatives. To construct the vector , we in a vector followed by the vector stack all the columns of formed by all columns of . Thus, when is an matrix, is a vector with elements. A row of the matrix contains the derivatives of the respective element of the vector with respect to all design parameters. Therefore, is a matrix. The approximate derivative matrices generated by the Broyden formula are typically less accurate [22] than those obtained by perturbations in the EM-FAST. Our experience shows that as the optimization proceeds, the response sensitivity estimates produced by Broyden-FAST converge toward the exact sensitivities. As a precaution, in the case of a diverging objective function, the algorithm defaults to the conventional EM-FAST technique.

The advantage of the Broyden update is that it is problem-independent and does not require any modiﬁcations of the analysis algorithm. Moreover, its computational requirements are negligible in comparison with the EM-FAST. The response and its gradient are obtained by a single system analysis with practically no overhead regardless of the number of design parameters . The potential of the Broyden update is demonstrated by two examples: the optimization of a Yagi–Uda array and the optimization of a microstrip-fed patch antenna. 1) Optimization of a Yagi–Uda Array: An initial design of the six-element Yagi–Uda antenna is given in Fig. 7. All dimensions are normalized with respect to the free-space wavelength . We vary the normalized lengths of the reﬂector and the driven and , as well as the normalized element, and . Thus, the separation distances vector of design parameters is . The ob. The progress of jective function is set as in (11) with the optimization is summarized in Table III where the changes of the design parameters, the input impedance and the objective function are recorded at each iteration. An optimal solution is reached in nine iterations. At the th design iteration, we update the four derivative , with Broyden’s formula and matrices use them to compute the response sensitivities according to (9). The response sensitivities are then used by the optimization algorithm (fmincon) to produce the next design iterate. The Broyden-FAST sensitivities are then compared with the sensitivities calculated off-line where the derivative matrices

NIKOLOVA et al.: ACCELERATED GRADIENT BASED OPTIMIZATION USING ADJOINT SENSITIVITIES

2153

Fig. 8. Sensitivity of the objective function with respect to the length of the reﬂector during the optimization of Z of the Yagi–Uda antenna.

Fig. 10. Sensitivity of the objective function with respect to the separation s during the optimization of Z of the Yagi–Uda antenna.

Fig. 9. Sensitivity of the objective function with respect to the length of the driven element l during the optimization of Z of the Yagi–Uda antenna.

Fig. 11. Sensitivity of the objective function with respect to the separation s during the optimization of Z of the Yagi–Uda antenna.

are obtained by the ﬁnite-difference approach of our original technique, the EM-FAST [17]. The sensitivity curves are plotted in Figs. 8–11. At the ﬁrst iteration only, we compute the derivative matrices using our original approach with forward ﬁnite differences and 1% perturbation over the initial design . That is why, at the ﬁrst parameters and assign those to iteration, the Broyden-FAST sensitivities and the EM-FAST sensitivities are identical. For all subsequent design iterations, Broyden-FAST uses (14). It is evident that our approach based on the Broyden update produces sufﬁciently accurate sensitivity results that converge toward the exact sensitivities as the optimization progresses. To quantify the accuracy of the derivative matrices produced by the Broyden update in the FAST, we compute their global relative errors in norm

Fig. 12. Global error in the Z -matrix derivative estimates of Broyden-FAST and EM-FAST.

(15)

Here, the exact derivative is computed using an analytical formula valid for this speciﬁc MoM solver [24]. Note that the error estimate (15) operates on complex matrix elements. The errors and associated with the normalized separation distances

2154

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 13. Sensitivities of the objective function in the optimization of the Yagi–Uda antenna using the direct Broyden update: comparison with reference sensitivities computed with the EM-FAST.

are plotted in Fig. 12. For comparison, we also plot the global errors of the matrix derivatives of the forward ﬁnite differencing in the original FAST computed for the same design iterates. As expected, the original FAST is robust while the accuracy of the Broyden estimates may vary throughout the optimization. These variations lead to the small (but observable) differences between the EM-FAST response derivatives and the respective Broyden-FAST estimates (see Figs. 10 and 11). Notice that close to the optimum solution where the optimizer takes very small steps (see Table III), the Broyden update may not perform very , due to the nearly idenwell for all design parameters, e.g., tical -matrices of the consecutive design iterates; see (14). If necessary, this can be avoided by defaulting to EM-FAST when sufﬁciently small value of the objective function is achieved. The improved accuracy of the sensitivity estimates may thus improve the convergence of the optimization at its ﬁnal stages. The hybrid approaches, however, are not a subject of our current discussion. The Broyden update, of course, can be applied directly to the objective function . However, the objective function usually exhibits strongly nonlinear behavior and sharp sensitivities with respect to the designable parameters. At the same time, the Broyden formula is based on a local linear approximation of the function and thus it performs better with only mildly nonlinear functions. The -matrix elements, on the other hand, are smooth functions of the shape or material parameters. In fact, the majority of the MoM matrix elements are almost insensitive to

shape perturbations except for the diagonal (self-impedance) elements due to the intrinsic dependence on the distance between observation and integration points. Thus, when the Broyden update is applied at the level of the system matrix, better convergence of the sensitivity estimates and of the overall optimization process is expected. In support of this observation, we repeat the Yagi–Uda antenna design, this time using Broyden’s update directly at the level of the objective function in order to estimate its derivatives (direct Broyden approach). We keep the optimization set-up identical to that before: (1) the initial design is as shown in Fig. 7; (2) the objective function is deﬁned as in (11) with ; (3) the same optimization function fmincon of MATLAB is used; (4) the stop criteria and are the same; and (5) the value of the response sensitivity at the ﬁrst iteration is supplied by the EM-FAST estimate. The direct-Broyden derivatives are used by the optimizer to determine the subsequent design iterates. We also compute the objective function derivatives with the EM-FAST technique off-line in order to supply reference values for comparison. The direct-Broyden and the reference sensitivity curves are plotted in Fig. 13 for all four designable parameters. The progress of the optimization is summarized in Table IV. It is evident that the direct-Broyden derivatives do not converge well toward the reference values, and the objective function converges to a different (worse) solution than that of the

NIKOLOVA et al.: ACCELERATED GRADIENT BASED OPTIMIZATION USING ADJOINT SENSITIVITIES

2155

TABLE IV OPTIMIZATION OF THE INPUT IMPEDANCE OF THE YAGI–UDA ARRAY WITH DIRECT BROYDEN UPDATE OF THE OBJECTIVE FUNCTION

Fig. 14.

Progress of the objective function during the optimization of the microstrip-fed patch antenna.

Broyden-FAST optimization. This is due mostly to three interrelated factors: (1) the objective function is very sensitive to the and ); (2) the Broyden designable parameters (especially sensitivity estimation can not track well such a rapidly changing function; (3) the incorrect sensitivity information misleads the optimizer. Possible solutions to the problems encountered with the direct Broyden sensitivity analysis are provided by the trust region optimization approaches. This topic, however, is outside of the scope of our work. With this example, we only illustrate the improved convergence of both the sensitivity analysis and the optimization when the Broyden update is applied at the level of the system matrix. 2) Optimization of a Microstrip-Fed Patch Antenna: We now apply the Broyden-FAST to the optimization of the patch antenna in Fig. 4. The length and the width of the patch are optimized for a minimum magnitude of the reﬂection coefﬁcient. Thus, the objective function to be minimized is (16) is the input impedance of the antenna, and is the Here, characteristic impedance of the feeding microstrip line, which

approximately equals 50 in this example. The operating frequency is 2 GHz. The initial values of the designable paramemm and mm. The patch is meshed ters are with rectangular segments. The number of segments along the length and the width of the patch are 11 and 17, respectively. One segment is used along the width of the feeding microstrip line. The optimization is carried out using the fmincon function of MATLAB. We compute the sensitivities of the objective function with three methods: 1) the direct-Broyden update; 2) the traditional EM-FAST which employs ﬁnite differences to approximate the derivative matrices (without a boundary layer); and 3) the proposed Broyden-FAST which employs the Broyden update at the level of the system matrix. This time, we run three optimizations, each being driven by the respective sensitivity analysis technique. The three objective functions are plotted in Fig. 14 versus the optimization iteration number. The design parameters versus the iteration number are plotted in Fig. 15. It is clear from both ﬁgures that the direct-Broyden method fails in meeting the optimum design while the other two methods, EM-FAST and Broyden-FAST, are capable of achieving the optimum design. mm The obtained optimal patch dimensions are mm. and

2156

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 15.

Progress of the design parameters during the optimization of the microstrip-fed patch antenna. [3] E. J. Haug, K. K. Choi, and V. Komkov, Design Sensitivity Analysis of Structural Systems. Orlando, FL: Academic, 1986. [4] A. D. Belegundu and T. R. Chandrupatla, Optimization Concepts and Applications in Engineering. Upper Saddle River, NJ: Prentice Hall, 1999. [5] R. Tomoviæ, Sensitivity Analysis of Dynamic Systems. New York: McGraw-Hill, 1963. [6] J. W. Bandler, “Computer-aided circuit optimization,” in Modern Filter Theory and Design, G. C. Temes and S. K. Mitra, Eds. New York: Wiley, 1973, ch. 6. [7] K. C. Gupta, R. Garg, and R. Chadha, Computer-Aided Design of Microwave Circuits. Dedham, MA: Artech, 1981. [8] J. W. Bandler and R. E. Seviora, “Current trends in network optimization,” IEEE Trans. Microwave Theory Tech., vol. MTT-18, pp. 1159–1170, Dec. 1970. , “Wave sensitivities of networks,” IEEE Trans. Microwave Theory [9] Tech., vol. MTT-20, pp. 138–147, Feb. 1972. [10] G. Iuculano, V. A. Monaco, and P. Tiberio, “Network sensitivities in terms of scattering parameters,” Electron. Lett., vol. 7, pp. 53–55, Jan. 1971. [11] J. W. Bandler, Q. J. Zhang, and R. M. Biernacki, “A uniﬁed theory for frequency-domain simulation and sensitivity analysis of linear and nonlinear circuits,” IEEE Trans. Microwave Theory Tech., vol. AP-36, pp. 1661–1669, Dec. 1988. [12] F. Alessandri, M. Mongiardo, and R. Sorrentino, “New efﬁcient full wave optimization of microwave circuits by the adjoint network method,” IEEE Microwave and Guided Wave Letters, vol. 3, pp. 414–416, Nov. 1993. [13] H. Akel and J. P. Webb, “Design sensitivities for scattering-matrix calculation with tetrahedral edge elements,” IEEE Trans. Magn., vol. 36, pp. 1043–1046, July 2000. [14] H.-B. Lee and T. Itoh, “A systematic optimum design of waveguide-tomicrostrip transition,” IEEE Trans. Microwave Theory Tech., vol. 45, pp. 803–809, May 1997. [15] J. P. Webb, “Design sensitivity of frequency response in 3-D ﬁnite-element analysis of microwave devices,” IEEE Trans. Magnetics, vol. 38, pp. 1109–1112, Mar. 2002. [16] J. Ureel and D. De Zutter, “Shape sensitivities of capacitances of planar conducting surfaces using the method of moments,” IEEE Trans. Microwave Theory Tech., vol. 44, pp. 198–207, Feb. 1996. [17] N. K. Georgieva, S. Glavic, M. H. Bakr, and J. W. Bandler, “Feasible adjoint sensitivity technique for EM design optimization,” IEEE Trans. Microwave Theory Tech., vol. 50, pp. 51–2758, Dec. 2002. [18] N. K. Nikolova, J. W. Bandler, and M. H. Bakr, “Adjoint techniques for sensitivity analysis in high-frequency structure CAD,” IEEE Trans. Microwave Theory Tech., Special Issue on Electromagnetics-Based Optimization of Microwave Components and Circuits,, to be published. [19] S. Amari, “Numerical cost of gradient computation within the method of moments and its reduction by means of a novel boundary-layer concept,” in Proc. IEEE MTT-S Int. Symp. Dig., vol. 3, 2001, pp. 1945–1948.

The reason that Broyden-FAST succeeded where the directBroyden approach failed is similar to the reason outlined in the previous example with the Yagi–Uda antenna. The directBroyden approach is incapable to represent properly the fast variations in the sensitivities especially close to an optimal solution, which is associated with a resonance. On the other hand, Broyden-FAST deals with the sensitivities of the -matrix elements, which are slowly varying functions unaffected by resonance. IV. CONCLUSION We propose a novel technique for accelerated gradient based optimization. It integrates the recently developed feasible adjoint sensitivity technique for full-wave EM analysis [17] with methods for accelerated estimation of the derivatives of the system matrix. We consider two such methods: the boundary-layer concept and the Broyden update, and investigate their accuracy and versatility. We show that the Broyden update of the matrix derivatives is efﬁcient, problem-independent and sufﬁciently accurate for the purpose of gradient based optimization. When integrated with our feasible adjoint sensitivity technique, it allows the computation of the system response and its gradient in the design parameter space through a single system analysis. The overhead associated with the gradient estimation is negligible in comparison with the computational requirements of the full-wave analysis. Our Broyden-FAST technique combines the efﬁciency of the adjoint-variable sensitivity analysis with the simplicity of the Broyden update. Its implementation is straightforward and does not require any analytical preprocessing. REFERENCES

[1] J. W. Bandler, Q.-J. Zhang, J. Song, and R. M. Biernacki, “FAST gradient based yield optimization of nonlinear circuits,” IEEE Trans. Microwave Theory Tech., vol. 38, pp. 1701–1710, Nov. 1990. [2] J. Ureel and D. De Zutter, “A new method for obtaining the shape sensitivities of planar microstrip structures by a full-wave analysis,” IEEE Trans. Microwave Theory Tech., vol. 44, pp. 249–260, Feb. 1996.

NIKOLOVA et al.: ACCELERATED GRADIENT BASED OPTIMIZATION USING ADJOINT SENSITIVITIES

2157

[20] C. A. Balanis, Antenna Theory, 2nd ed. New York: Wiley, 1997. [21] C. G. Broyden, “A class of methods for solving nonlinear simultaneous equations,” Mathematics of Computation, vol. 19, pp. 577–593, 1965. [22] J. W. Bandler, S. H. Chen, S. Daijavad, and K. Madsen, “Efﬁcient optimization with integrated gradient approximations,” IEEE Trans. Microwave Theory Tech., vol. 36, pp. 444–455, Feb. 1988. [23] J. W. Bandler, R. M. Biernacki, S. H. Chen, R. H. Hemmers, and K. Madsen, “Electromagnetic optimization exploiting aggressive space mapping,” IEEE Trans. Microwave Theory Tech., vol. 43, pp. 2874–2882, Dec. 1995. [24] S. Glavic, “Electromagnetic Design Sensitivity Analysis of High-Frequency Structures,” M.A.Sc. thesis, McMaster University, May 2002.

Natalia K. Nikolova (formerly Georgieva) (S’93–M’97) received the Dipl.Eng. degree in radioelectronics from the Technical University of Varna, Bulgaria, in 1989 and the Ph.D. degree in electrical engineering from the University of Electro-Communications, Tokyo, Japan, in 1997. From 1998 to 1999, she held a Postdoctoral Fellowship of the Natural Sciences and Engineering Research Council of Canada (NSERC), during which time she was initially with the Microwave and Electromagnetics Laboratory, DalTech, Dalhousie University, Halifax, Canada. For approximately one year, she was with the Simulation Optimization Systems Research Laboratory, McMaster University, Hamilton, ON, Canada. In July 1999, she joined the Department of Electrical and Computer Engineering, McMaster University, where she is currently an Associate Professor. Her research interests include theoretical and computational electromagnetism, high-frequency analysis techniques, as well as CAD methods for high-frequency structures and antennas. Dr. Nikolova currently holds a University Faculty Award of NSERC, which she received in 2000.

Mohamed H. Bakr (S’98–M’00) received the B.Sc. degree in electronics and communications engineering (distinction with honors) and the Master’s degree in engineering mathematics from Cairo University, Giza, Egypt, in 1992 and 1996, respectively, and the Ph.D. degree from the Department of Electrical and Computer Engineering, McMaster University, Hamilton, ON, Canada, in September 2000. In 1997, he was a Student Intern with Optimization Systems Associates (OSA), Inc. From 1998 to 2000, he worked as a Research Assistant with the Simulation Optimization Systems (SOS) Research Laboratory, McMaster University. In November 2000, he joined the Computational Electromagnetics Research Laboratory (CERL), University of Victoria, Victoria, Canada as an NSERC Postdoctoral Fellow. He is currently an Assistant Professor with the Department of Electrical and Computer Engineering, McMaster University. His research areas of interest include optimization methods, computer-aided design and modeling of microwave circuits, neural network applications, smart analysis of microwave circuits and efﬁcient optimization using time/frequency domain methods.

Reza Saﬁan received the B.Sc. degree from the Isfahan University of Technology (IUT), Isfahan, Iran, in 1999 and the M.A.Sc. degree from McMaster University, Hamilton, Canada, in 2003, both in electrical engineering. He is currently working toward the Ph.D. degree at the University of Toronto, Toronto, ON, Canada. During 1999 to 2002, he was a Research Engineer in the Electrical and Computer Engineering Research Center (ECERC), IUT. From 2002 to 2003, he was a Research Assistant at McMaster University working with the Computational Electromagnetics Laboratory and the Simulation Optimization Systems Research Laboratory. His research interests include theory of electromagnetism and computational electromagnetics.

Ezzeldin A. Soliman (S’97–A’99) was born in Cairo, Egypt, on May 18, 1970. He received the B.Sc. degree (distinction with honors) in electronics and communications engineering and the M.Sc. degree in engineering physics, both from Cairo University, Giza, Egypt, in June 1992 and Nov. 1995, respectively, and the Ph.D. degree (summa cum laude) in electrical engineering from the University of Leuven, Leuven, Belgium, in February 2000. From 1992 to 1996, he was a Research and a Teaching Assistant with the Department of Engineering Physics, Faculty of Engineering, Cairo University. From 1996 to 2000, he has been a Research Assistant at both Interuniversity MicroElectronics Center (IMEC), Leuven, Belgium, and the Department of Electrical Engineering, University of Leuven. From April 2002 to July 2002, he was a Visiting Professor at IMEC. From October 2002 to September 2003, he was on a Visiting Researcher in the Department of Electrical and Computer Engineering, McMaster University, Hamilton, ON, Canada. He is currently an Assistant Professor with the Department of Engineering Physics, Faculty of Engineering, Cairo University. His research interests include computational electromagnetics, development and characterization of planar antennas in the multilayer thin ﬁlm technology, neural network modeling of electromagnetic problems, and the EM-based optimization techniques.

John W. Bandler (S’66–M’66–SM’74–F’78) was born in Jerusalem, on November 9, 1941. He studied at the Imperial College of Science and Technology, London, U.K., from 1960 to 1966, and received the B.Sc.(Eng.), Ph.D., and D.Sc.(Eng.) degrees from the University of London, London, U.K., in 1963, 1967, and 1976, respectively. In 1966, he joined Mullard Research Laboratories, Redhill, Surrey, U.K. From 1967 to 1969, he was a Postdoctorate Fellow and Sessional Lecturer at the University of Manitoba, Winnipeg, Canada. He joined McMaster University, Hamilton, ON, Canada, in 1969, where he has served as Chairman of the Department of Electrical Engineering and Dean of the Faculty of Engineering, and is currently Professor Emeritus in the Electrical and Computer Engineering Department, directing research in the Simulation Optimization Systems Research Laboratory. He was President of Optimization Systems Associates Inc. (OSA), which he founded in 1983, until November 20, 1997, the date of acquisition of OSA by Hewlett-Packard Company (HP). OSA implemented a ﬁrst-generation yield-driven microwave CAD capability for Raytheon in 1985, followed by further innovations in linear and nonlinear microwave CAD technology for the Raytheon/Texas Instruments Joint Venture MIMIC Program. OSA introduced the CAE systems RoMPE in 1988, HarPE in 1989, OSA90 and OSA90/hope in 1991, Empipe in 1992, Empipe3D and EmpipeExpress in 1996. OSA created the product empath in 1996 which was marketed by Sonnet Software, Inc., USA. Dr. Bandler is President of Bandler Corporation, which he founded in 1997. He has published more than 350 papers from 1965 to 2003. He contributed to Modern Filter Theory and Design (Surrey, U.K.: Wiley-Interscience, 1973) and to Analog Methods for Computer-aided Analysis and Diagnosis Germany: Marcel Dekker, Inc., 1988). Four of his papers have been reprinted in Computer-Aided Filter Design (New York: IEEE Press, 1973), one in each of Microwave Integrated Circuits (Norwood, MA: Artech House, 1975), Low-Noise Microwave Transistors and Ampliﬁers (New York: IEEE Press, 1981), Microwave Integrated Circuits, (Norwood, MA: Artech House, 1985, 2nd ed.), Statistical Design of Integrated Circuits (New York: IEEE Press, 1987)and Analog Fault Diagnosis (New York: IEEE Press, 1987). He joined the Editorial Boards of the International Journal of Numerical Modeling in 1987, the International Journal of Microwave and Millimeterwave Computer-Aided Engineering in 1989, and Optimization and Engineering in 1998. He was Guest Editor of International Journal of Microwave and Millimeter-Wave Computer-Aided Engineering, Special Issue on Optimization-Oriented Microwave CAD (1997). He was Guest Coeditor Optimization and Engineering Special Issue on Surrogate Modeling and Space Mapping for Engineering Optimization (2001). Dr. Bandler is a Fellow of the Canadian Academy of Engineering, the Royal Society of Canada, the Institution of Electrical Engineers (IEE), London, U.K., anf the Engineering Institute of Canada. He is a Member of the Association of Professional Engineers of the Province of Ontario (Canada), the MIT Electromagnetics Academy, and the Micronet Network of Centres of Excellence. He received the Automatic Radio Frequency Techniques Group (ARFTG) Automated Measurements Career Award in 1994. He was an Associate Editor of the IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES (1969–1974), and has continued serving as a member of the Editorial Board. He was Guest Editor of the IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES Special Issue on computer-oriented microwave practices (1974) and Guest Coeditor of the of the IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES Special Issue on Process-Oriented Microwave CAD and Modeling (1992), and Guest Editor, IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, Special Issue on Automated Circuit Design Using Electromagnetic Simulators (1997). He is Guest Coeditor, IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, Special Issue on Electromagnetics-Based Optimization of Microwave Components and Circuits (2004). He has served as Chair of the MTT-1 Technical Committee on Computer-Aided Design.

2158

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

**A Theoretical Study of the Stability Criteria for Hybridized FDTD Algorithms for Multiscale Analysis
**

M. Marrone, Student Member, IEEE, and R. Mittra, Life Fellow, IEEE

Abstract—In this paper, we propose two new hybrid, two-dimensional, generalized ﬁnite-difference time-domain algorithms—formulated by using the cell method—and designed to analyze objects with ﬁne details, without using very small time steps dictated by the Courant condition. A detailed analysis of the stability of the proposed algorithms is presented, with special attention devoted to the phenomenon of late time instabilities. Finally, some rules are provided that would ensure the stability of the proposed algorithms. Index Terms—Cell method (CM), generalized ﬁnite-difference time-domain (FDTD) method, hybrid algorithms, stability, subgridding.

I. INTRODUCTION

W

HEN analyzing complex and multiscale structures using the time domain methods, it often becomes necessary to combine two algorithms, e.g., the ﬁnite-difference time-domain (FDTD) and ﬁnite-element time-domain (FETD), to improve the accuracy of the solution without placing an inordinately heavy burden on the CPU. For instance, it is often desirable to use the FDTD on a coarse, structured Cartesian mesh in the bulk of the region, and to employ either a FDTD subgridding scheme in a smaller region containing objects with ﬁne features [1], [2], or to use the FETD on an unstructured mesh to further improve the modeling accuracy in this region [3]. Other alternative approaches that are still under investigation are the hybrid alternating direction implicit (ADI)-FDTD schemes on subgridding [4] and the ADI-Multiresolution time domain (MRTD) methods [5]. These methods circumvent the problem of having to use a very small time step throughout the computational domain to satisfy the Courant condition [6], which is associated with the smallest length of the mesh edges in the entire computational domain. But they do introduce the burden of temporal and spatial interpolations, and the distinct possibility of instabilities introduced by the spurious reﬂections at the interface of the two domains with dissimilar meshes. What is even more disconcerting is that, although the two algorithms applied in the different domains may be independently stable (in the absence of the other domain), the combination may still be unstable and produce the so-called late time instabilities. It has been demonstrated that this kind of instabilities related to spatial interpolation schemes, such as those employed in the sub-

gridding schemes, are due to a lack of symmetry in some discrete operators, and are not dependent on the Courant condition [7], [8]. In this paper we propose two new hybrid generalized FDTD algorithms where each of them is a combination of two algorithms that work on different grids—a coarse mesh FDTD in one domain, and either a ﬁne mesh or a triangular one in the other—and with different time stepping schemes. To investigate their stability, we will need to account for the instabilities that may arise from the spatial as well as temporal interpolations. The two hybrid algorithms will be developed by using the recently-introduced cell method (CM) [[10]–[13]] that enables us to address both the problems of instability and connectivity between the different algorithms. This paper is organized as follows. Section II presents the CM and Section III lays the foundations of the main algorithm wave propagation in two dimenfor the analysis of the sional (2-D) cases on a coarse mesh FDTD. Section IV presents two new algorithms, suitable for either a ﬁne or an unstructured mesh, which are later combined with the algorithm in Section III to set up two hybrid algorithms, whose stability analyzes are presented in Section V, followed by a brief conclusion given in Section VI. II. THE CM A study of the mathematical structure common to many physical theories [9] provides a discrete mathematical framework for the electromagnetic ﬁeld theory, which can be utilized to develop a numerical method for solving electromagnetic static and dynamic problems on unstructured grids. This approach is referred to as the CM, which employs the global (integral) variables instead of the local ones. To deﬁne the global variables, and to avoid any sign ambiguities in their values, we need to associate them in a physically coherent way, with certain oriented space and time elements [9], [10]. Speciﬁcally, we need two types of oriented space elements: 1) inner-oriented space elements (points , lines , surfaces , volumes ) and 2) complementary outer-oriented space elements (points , lines , surfaces , volumes ). In addition, we need two types of oriented time elements: i) inner oriented time elements (instants , intervals ) and ii) complementary outer oriented time elements (instants , intervals ). The principal global variables of electromagnetics are as follows: • Electric voltage impulse ; ; • Magnetic ﬂux ; • Magnetic voltage impulse

Manuscript received May 15, 2003; revised September 30, 2003. The authors are with the Electromagnetic Communication Laboratory, Pennsylvania State University, University Park, PA 16802 USA (e-mail: mum17@psu.edu; mittra@engr.psu.edu). Digital Object Identiﬁer 10.1109/TAP.2004.832332

0018-926X/04$20.00 © 2004 IEEE

MARRONE AND MITTRA: THEORETICAL STUDY OF STABILITY CRITERIA FOR HYBRIDIZED FDTD ALGORITHMS

2159

Fig. 1. Primal and dual space cell complexes and the global variables associated with their primal and dual cells.

Fig. 2. Primal and dual time cell complexes and the global variables associated with their primal and dual cells.

• Electric ﬂux • Electric charge content • Electric current ﬂow

; ; .

The above are associated, respectively, with the oriented space and time elements appearing inside the square brackets. A suitable geometrical structure that supports the global variables in CM is comprised of two staggered and oriented space cell complexes. A space complex, which is synonymous with the 3-D grid, is a structured collection of space elements (cells), such as nodes, edges, faces and volumes. In particular, we need two oriented space cell complexes: 1) a primal complex, made up of inner-oriented space elements, supporting the global variables associated with the above elements and 2) a complementary dual complex, comprising of outer-oriented space elements, supporting the global variables associated with the outer-oriented space elements. Moreover, to avoid instability problems, we need to establish a relationship of duality between the two complexes. Speciﬁcally, we need as many primal nodes as dual volumes ; and the same is true for the primal edges and the dual faces , the primal faces and the dual edges , the primal volumes and the dual nodes . In Fig. 1 we show an example of primal and dual cubical space cell complexes, and the association of the global variables with the space cells. The time structure utilized in the CM also involves two staggered and oriented time cell complexes, viz., a primal complex, comprised of inner-oriented time elements, and a dual complex made up of outer-oriented time elements. Moreover, in common with the space structure, the two time complexes must be related by duality, i.e., we need an equal number of primal instants and dual intervals ; the same is true for the primal intervals and the dual instants . For the sake of brevity, and taking into account the duality between the primal and the dual time cell complexes, we will use the following notation: primal instants , primal intervals , dual instants and dual intervals . In Fig. 2 we show an example of primal and dual time cell complexes with the common choice of the primal and dual intervals equal to the , and the association of the global time step variables with the time cells. In the CM, the electromagnetic laws can be formulated in an algebraic form on the cell complexes, in terms of the global variables. In particular, these laws can be divided into the following

two categories, depending upon the type of global variables that are involved. • Field equations (topological equations). These equations link the global variables associated with space (time) cells belonging to the same type of complex (either the primal cell complex or the dual one). The ﬁeld equations can be enforced on the cell complexes in an exact discrete form by denote using appropriate incidence matrices. Let the incidence matrices between the oriented primal elements, such as edges and nodes, faces and edges, volumes and faces, respectively, and let represent the corresponding incidence matrices between the oriented dual elements, such as edges and nodes, faces and edges, volumes and faces, respectively [7], [9], [12], [13]. The above may be viewed as discrete counterparts of the differential operators gradient, curl and divergence. Thus, the ﬁeld equations can be expressed in a Yee-like form [15] as follows: — Faraday-Neumann law (1) which relates the global variables on the primal complex; Magnetic Gauss law (2) which relates the global variables on the primal complex; Ampère-Maxwell law (3) — which relates the global variables on the dual complex; Electric Gauss law (4) which relates the global variables on the dual complex where are arrays of scalars, the superscript indicates the association with either a primal time instant or a dual time interval, while indicates the association with either the superscript a primal time interval or a dual time instant. • Space-time constitutive relations. These relations link the global variables associated with space and time cells belonging to different types of cell complexes. The electric

—

—

2160

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

space-time constitutive relation links the variables and , whereas the magnetic space-time constitutive relation provides a link between the variables and . These relations employ the material parameters and can be enforced on the cell complexes only in an approximate and discrete form by using suitable algebraic relationships. Next, by introducing the following intermediate variables: ; — electric voltage ; — magnetic voltage ; — electric current that are global variables in space, but local variables in time, we can formulate the space-time constitutive relations by combining two simpler relations. For example, to calculate the electric voltage impulse from the electric ﬂux , we can ﬁrst calculate the electric voltage from the electric ﬂux , by using the constitutive relation , and we can then calculate the electric voltage impulse from the electric voltage , by a time integral . The time integration relations and the relation constitutive relations we need are given below: — Time integration relationships: These relationships can be extracted in an algebraic form from the following time integral relationships:

are referred to where the matrices as constitutive matrices. These matrices can be generated explicitly from the knowledge of the geometrical features of the cell complexes, and the features of the materials inside the computational domain. As we will see in Section IV, the constitutive matrices must be symmetric and positive deﬁnite in order for the time updating scheme to be stable. These features can be obtained via the symmetrized microcell interpolation scheme (SMIS) [14], which also ensures that the constitutive matrices are physically consistent. III. GENERALIZED FDTD 2-DEXPLICIT ALGORITHM FORMULATED BY THE CM The formulation of the generalized FDTD algorithm on two staggered cell complexes follows from the combination of the algebraic electromagnetic laws of the CM. In particular, we wave propagation, because the focus here only on the 2-D wave propagation leads to similar theoretical results. 2-D Since the ﬁelds do not vary along the z-axis we can use a 2-D cell complex obtained from a projection of a 3-D cell complex onto the xy-plane (Fig. 3). The resulting 2-D cell complex is characterized by primal nodes ; edges ; faces ; and dual nodes ; edges and faces . On this 2-D cell complex, the associations of the physical variables to the space elements are related to those in the 3-D case as follows: ; • electric voltage ; • electric ﬂux ; • magnetic voltage • magnetic ﬂux ; ; • electric current and we need to modify the ﬁeld equations slightly from the 3-D case. In particular, for the FDTD algorithms that we need to set up, we are interested only in the Faraday–Neumann and Am2-D case, we can use the time père–Maxwell laws. For the integration relation in (5), to reduce these modiﬁed equations to

In practice, we utilize the following two sets of time integration relations. The ﬁrst set is (5) and it leads to the leap-frog time-stepping scheme employed in conventional FDTD [15], [6]. The second set is (6) and leads to the Newmark time-stepping scheme employed in FETD [16], which is an implicit scheme. Constitutive relations: These relations approximate the following local constitutive relationships in an algebraic form:

(9) We point out that we have translated the quantities in the above equations by half a time-step, without loss of generality, in order to simplify the analytical formalism of the stability criterion to be discussed in Section V. Finally, by combining the ﬁeld equations given in (9) with the constitutive relations, we can derive an explicit time stepping algorithm (see Fig. 4), governing the 2-D wave propagation, which reads (10) In general, the incidence matrices that multiply the array vectors and in (10) must be adjoint one to the other, and the constitutive matrices must be symmetric and positive deﬁnite, otherwise late time instabilities may arise. Moreover, a generalized Courant condition must still be satisﬁed to avoid instabilities [7]. If the cell complexes we use are Cartesian orthogonal

—

Depending on the type of algorithms, we can utilize either one of the following two algebraic constitutive relations: (7) or the: (8)

MARRONE AND MITTRA: THEORETICAL STUDY OF STABILITY CRITERIA FOR HYBRIDIZED FDTD ALGORITHMS

2161

Fig. 3.

Associations of the physical variables with the elements of a 2-D cell complexes for the study of

TM

waves propagation.

Fig. 4.

Graphical visualization of the CM-FDTD algorithm with the leap-frog time stepping.

grids, as shown in Fig. 3, the algorithm in (10) becomes identical to the classical FDTD algorithm, once we replace the electric and magnetic voltages with the electric and magnetic ﬁeld components. In this case, the conditions sufﬁcient for avoiding late time instabilities are satisﬁed automatically; hence, just as in the case of the conventional FDTD, the algorithm in (10) is stable if the time step satisﬁes the following Courant condition [6] (11) where and are the grid step sizes and is the speed of light. Since (10) can be utilized on arbitrary 2-D primal and dual cell complexes, we can derive a generalized FDTD algorithm for such grids, and we will refer to it as the CM-FDTD algorithm. IV. HYBRID ALGORITHMS When analyzing ﬁne details in a small region via the FDTD method, it is often convenient to use subgridding in that region. Although this technique improves the accuracy of the results, it

places a heavy burden on the CPU, since it requires the use of a very small time step throughout the computational domain to satisfy the Courant condition (11), which is associated with the smallest length of the mesh edges in the entire computational domain. It is possible to overcome this problem by combining two algorithms into a hybrid one, where we employ the FDTD of the computational domain , algorithm in the main part and an alternate one in a smaller region , which either uses the subgridding (Fig. 5) or an unstructured mesh (Fig. 6). The alternate algorithms should be chosen such that their stability characteristics are not so tied to the smallest length of the grid as those ones of the FDTD algorithm. in the region The use of a complementary grid that is dual to the primal grid throughout the entire computational domain, as for instance in the cases shown in Figs. 5 and 6, allows us to maintain the self-adjointeness between the matrices that multiply the array vectors V and F in (10). This approach is general and it is different from that in [8] where one must replace the matrices and (or and in the 3-D cases) with some extended discrete operators constructed via the use of some particular interpolation schemes.

2162

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

where is the incidence matrix related to the primal cell is the incidence matrix related to the primal complex in ; is the incidence matrix related to cell complex in ; and, connected to the primal the part of the primal cell complex in cell complex in . Moreover, we assume that the common zone and in , is connected between the two cell complexes in via rectangular surfaces (see Figs. 5 and 6) in order for two of the four blocks to be zero in the partitioned form of the constitutive matrices in (8). For instance (12) (13)

Fig. 5. Example of two subdomains. The domain

: square grid, the domain

: square grid with subgridding.

Finally, in practice, the only variables we are interested in calculating are the electric and the magnetic voltages associated with and . Speciﬁcally, the elements of the two subdomains they are • electric voltages ; . • magnetic voltages A. Algorithm for the Subdomain In the subdomain we have a coarse Cartesian orthogonal mesh, and we can use the CM-FDTD algorithm given in (10). In this region, the blocks of the incidence matrix and of the constitutive matrices (12), (13) can be reduced to yield: • • . For the stability analysis, discussed in Section V, it is convenient to rewrite the CM-FDTD algorithm as a two-step one as follows:

**Fig. 6. Example of two subdomains. The domain
**

: square grid, the domain

: square grid with a triangular grid.

To set up the hybrid algorithms for analyzing the wave propagation in a 2-D domain , we consider splitting it into two and , where we have a coarse mesh in the resubdomains gion and a ﬁne one in . When we discretize using the primal and dual cell complexes, partitioning it into two subdomains creates a decomposition of the primal nodes set into two and . The primal edge set is also divided into sets, viz., two sets, namely and , where all the edges in connect only nodes belonging to the set . Thus, in the subdomain we have the primal nodes and the primal edges whereas there are the primal nodes and the primal edges in with their associated dual elements. In addition, a decomposition of the set of the primal nodes as well as the edges into two introduces a partitioning of the original incidence matrices and the constitutive matrices into four blocks. These blocks are suitable not only for the derivation of the algorithms, but also for connecting them in a compact form that enables us to study the stability of the resulting hybrid algorithms more conveniently. In particular, for the incidence matrix , we have the following partition:

(14) B. Algorithms for the Subdomain In contrast to the subdomain contains objects with ﬁne features that we want to model either by using a subgridding scheme (see Fig. 5), or with an unstructured mesh (Fig. 6). In wave propagation in this region, we order to analyze the propose two different algorithms for which the maximum allowable time step is not dependent on the smallest length of the mesh edges in the same manner as it is in the FDTD algorithm. The ﬁrst algorithm, which we will refer to as CM-NEW, is an implicit algorithm, whereas the second one, which we will name the CM-TS-FDTD algorithm, is explicit in nature. 1) CM-NEW Algorithm: The CM-NEW algorithm is obtained in a way that is similar to the CM-FDTD algorithm, but it employs a Newmark time integration scheme. Speciﬁcally, it utilizes the time-stepping (6) • •

MARRONE AND MITTRA: THEORETICAL STUDY OF STABILITY CRITERIA FOR HYBRIDIZED FDTD ALGORITHMS

2163

Fig. 7.

**Graphical visualization of the hybrid algorithm CM-FDTD (domain
**

) and CM-NEW (domain

).

Fig. 8.

**Graphical visualization of the hybrid algorithm CM-FDTD (domain
**

) and CM-TS-FDTD (domain

).

The above can be rewritten as the following two-step implicit algorithm:

•

C. The Hybrid Algorithms The hybrid algorithms are derived by combining the CM-FDTD algorithm with either the CM-NEW or the CM-TS-FDTD. The are graphically represented in Figs. 7 and 8, respectively. (15) 2) CM-TS-FDTD Algorithm: The CM-TS-FDTD algorithm is a type of CM-FDTD algorithm with a leap-frog time-stepping scheme, with a time subgridding (TS) of • • • • • V. STABILITY ANALYSIS In order to study the stability of the hybrid algorithms, it is convenient to restate them, without loss of generality, in terms of and , which are related to the variables the new variables and as follows: (16) and are real symmetric and where the matrices is symmetric and positive deﬁnite, provided that the matrix positive deﬁnite. Moreover, let us deﬁne the following matrices,

2164

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

that are useful for rewriting the hybrid algorithms in terms of the and new variables

(17)

Fig. 9. Examples of domains

. (a) Case SUB, (b) Case TRI-05.

where

. If the constitutive matrices are symmetric and positive deﬁnite, are symmetric, and is symmetric positive defthen the inite. Then there exists the matrix

A. Stability Analysis of the Hybrid Algorithm Set by CM-FDTD (Domain ) and CM-NEW (Domain ) To simplify the derivation of the stability criterion, we begin by assuming that there are no input sources. Then, by using (14), (15), (16), and (17), we can rewrite the hybrid time-stepping algorithm as follows:

such that the matrix

, given by

(18) The system of equations in (18) has the following form: (19) where

and

Using the basic notions of discrete system theory (see Appendix), it can be shown that the system in (19) is stable if two of the following conditions are satisﬁed. The ﬁrst condition be symmetric, or that there exists a requires that the matrix nonsingular matrix such that

is symmetric. Therefore, the ﬁrst condition for the stability of the hybrid algorithm is that the constitutive matrices be symmetric and positive deﬁnite. Because the complexity of the combined algorithm, it is not possible to derive an explicit Courant criterion, such as (11), which is employed in the FDTD algorithm to satisfy the second condition (20) for the stability. For instance, for the domains in Fig. 9, where the grid step sizes of are m, it folthe FDTD grid in the domain – s. For these domains, we lows from (10) that with varying from 0 to 2.5e–9 s and have calculated steps of 0.05e–9 s, and have veriﬁed that (20) is satisﬁed when – s. Moreover, the absence of instabilities has been conﬁrmed by means of time domain simulations, on the same – s, for 100 000 time steps. domains, run with a This leads us to conclude that this algorithm does not suffer for late time instabilities, provided that the constitutive matrices are symmetric and positive deﬁnite, and the Courant condition reduce in practice to (10) regardless of the type of grid utilized (either the subgridding or the triangular grid), in the domain due to the implicit nature of the CM-NEW algorithm. B. Stability Analysis of the Hybrid Algorithm Set by CM-FDTD (Domain ) and CM-TS-FDTD (Domain ) Using (16), and (17), we can rewrite the CM-TS-FDTD algorithm, in the absence of input sources, as shown in (21) at the bottom of the next page. From (14), (16), and (17), and deﬁning

where is a symmetric matrix. One can show that late time instabilities would occur if the latter condition is not satisﬁed, of are complex [7]. The and some of the eigenvalues second condition, a generalized Courant criterion, demands that max of must satthe maximum eigenvalue in modulus isfy the bound given by (20)

we can write the hybrid algorithm as

(22)

MARRONE AND MITTRA: THEORETICAL STUDY OF STABILITY CRITERIA FOR HYBRIDIZED FDTD ALGORITHMS

2165

The system in (22) has the same form as that in (19), provided to be we deﬁne

TABLE I

and

TABLE II

If the constitutive matrices are symmetric and positive deﬁnite, then is symmetric, and if (23) then is symmetric and positive deﬁnite. In this case are symmetric; hence, the matrix

is such that the matrix

is also symmetric. Unfortunately, it is not possible to ﬁnd a more practical criterion than the one given in (23), one which is valid irrespective of the type of grid used in the domain . Let us consider the following three cases: 1) the subdomain has a subgridding [case SUB, Fig. 9(a)]; 2) a triangular grid with a reﬁnement factor of 0.5 [case TRI-05, Fig. 9(b)]; and 3) a triangular grid with reﬁnement factor of 0.33 (case TRI-033). , which is the upper bound For each of these cases the that avoids late time instabilities, is presented in Table I. From depends on the reﬁnement of this Table we see that the the grid (cases TRI-05 and TRI-033). In summary, the stability conditions of the hybrid algorithm (22) are that the constitutive matrices be symmetric and positive deﬁnite, and that the time-step has an upper bound given by (23). Turning now to the additional condition for the stability, viz., (20), it is not possible to ﬁnd a criterion that is both practical and general. However, for the three cases analyzed above, we have calculated the max, with varying from 0 to 2.5e–9 s, and steps of 0.05e–9 s, to ﬁnd the maximum sufﬁcient to meet (20). The results are

presented in Table I. As we can see from this Table, with the exception of the triangular grid with a reﬁnement factor of 0.33, is equal or close to the , which ensures the the stability of the FDTD on the subdomain . Unfortunately, as we see from the Tables I and II, for each that avoids the late time instabilities is approxcase the imatively one half of the corresponding to the Courant limit. However, since (23) represents a sufﬁciency condition, it does not necessarly follow that the late time instabilities would . occur if we use a time step such that As a matter of fact, in the time domain simulations run for each there were no late time instabilities even case with after 100 000 time steps. This leads us to conclude that this algorithm is stable, provided that the constitutive matrices are symmetric and positive deﬁnite, and the time step either satisﬁes or, as in the cases analyzed here, satisﬁes the criterion in (23). For the examples provided allowed was always less than ; however, herein, the algorithm is still less expensive then the one without the time . subgridding, so long as VI. CONCLUSION In this paper, we have proposed two new hybrid 2-D time domain algorithms, implemented by the CM, in order to analyze ﬁne details without placing a heavy constraint on the maximum time step imposed by the stability criterion. Additionally, the matrix notation of the CM has enabled us to carry out the stability analysis of the hybrid algorithms, and to derive two conditions that guarantee stable numerical solutions. Finally, the theoretically derived stability criteria have been tested via a numerical study. Although the analysis presented in this paper has only addressed the 2-D case, it can be readily extended to 3-D geometries, and the authors are currently in the process of pursuing this effort.

(21)

2166

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

APPENDIX We wish to demonstrate that the discrete system (24) is stable if the matrix matrix such that is symmetric, or there is a nonsingular

REFERENCES

[1] M. W. Chevalier, R. J. Luebbers, and V. P. Cable, “FDTD local grid with material traverse,” IEEE Trans. Antennas Propagat., vol. 45, pp. 411–421, Mar. 1997. [2] M. Okoniewski, E. Okoniewska, and M. A. Stuchly, “Three dimensional subgridding algorithm for FDTD,” IEEE Trans. Antennas Propagat., vol. 45, pp. 422–429, Mar. 1997. [3] M. Feliziani and F. Maradei, “Mixed ﬁnite-difference/whitney-elements time domain (FD/WE-TD) method,” IEEE Trans. Magn., vol. 34, pp. 3222–3227, Mar. 1998. [4] B. Z. Wang, Y. Wang, W. Yu, and R. Mittra, “A hybrid 2-D ADI-FDTD subgridding scheme for modeling on-chip interconnects,” IEEE Trans. Adv. Packaging, vol. 24, pp. 528–533, Nov. 2001. [5] Z. Chen and J. Zhang, “An unconditionally stable 3-D ADI-MRTD method free of the CFL stability condition,” IEEE Microwave Wireless Comp. Lett., vol. 11, pp. 349–351, Aug. 2001. [6] A. Taﬂove and S. C. Hagness, Computational Electrodynamics: The Finite-Difference Time-Domain Method. Norwood, MA: Artech House, 2000. [7] T. Weiland and R. Schuhmann, “Space and time stability of discrete time domain algorithms,” in Proc. 4th Int. Workshop on Computational Electromagnetics in the Time Domain, Nottingham, U.K., July 17–19, 2001, pp. 155–161. [8] O. Pod˘ brand, M. Clemens, and T. Weiland, “New ﬂexible subgridding e scheme for the ﬁnite integration technique,” IEEE Trans. Magn., vol. 39, pp. 1662–1665, May 2003. [9] E. Tonti, On the Formal Structure of Physical Theories, Italian National Research Council, 1975. preprint. , “Finite formulation of the electromagnetic ﬁeld,” IEEE Trans. [10] Magn., vol. 38, pp. 333–336, Mar. 2002. [11] M. Marrone, A. M. F. Frasson, and H. E. Hernández-Figueroa, “A novel numerical approach for electromagnetic scattering: The cell method,” in Proc. IEEE AP-S URSI, vol. 1, San Antonio, TX, June 16–21, 2002, pp. 160–163. [12] M. Marrone, V. F. Rodríguez-Esquerre, and H. E. Hernández-Figueroa, “Novel numerical method for the analysis of 2D photonic crystals: The cell method,” Opt. Express, vol. 10, no. 22, pp. 1299–1304, Nov. 2002. [13] M. Marrone, R. Mittra, and W. Yu, “A novel approach to deriving a stable hybridized FDTD algorithm using the cell method,” in Proc. IEEE AP-S URSI, Columbus, OH, June 22–27, 2003. [14] M. Marrone, “A novel technique to build constitutive matrices for generalized FDTD algorithms,” in Proc. IEEE AP-S URSI, Columbus, OH, June 22–27, 2003. [15] K. S. Yee, “Numerical solution of initial boundary value problems involving Maxwell’s equations in isotropic media,” IEEE Trans. Antennas Propagat., vol. 14, pp. 302–307, Mar. 1966. [16] S. D. Gedney and U. Navsariwala, “An unconditionally stable ﬁnite element time-domain solution of the vector wave equation,” IEEE Trans. Microwave Guided Wave Lett., vol. 5, pp. 332–334, Oct. 1995. [17] G. H. Golub and C. F. Van Loan, Matrix Computations. Baltimore, MD: The Johns Hopkins Univ. Press, 1996. [18] P. Henrici, Error Propagation for Difference Methods. New York: Wiley, 1963.

where is symmetric matrix, and if the maximum eigenvalue satisﬁes the bound in modulus of (25) The matrix ical form can always be transferred into the Jordan canonthrough a nonsingular matrix as follows:

and, since is symmetric or it has the same eigenvalues as is a diagonal matrix. By those of a symmetric matrix, then we can transform the deﬁning the new variable system (24) into a new one which reads (26) The solution of the above equation is stable if and only if (24) is stable and if each of the following one component systems

are stable. The root conditions given in [18] states that in general the discrete system, (27) with , is stable, regardless of the nature of the input source, if the roots of the polynomial equation (28) satisfy . Using basic algebraic tools, we can demonstrate , that the roots of the polynomial equation in (27) are if (29) Since the eigenvalues of the matrix are all real, the system (24) is stable if either , or, in general, the in modulus of is such that maximum eigenvalue

Massimiliano Marrone (S’02) received the Laurea degree (summa cum laude) in electrical engineering and the Ph.D. degree in information technology from the University of Trieste, Italy, in 1999 and 2003, respectively. From July 2001 to January 2002, he was a Visiting Scholar at the Microwaves and Optics Department, Universidade Estadual de Campinas (UNICAMP), Sao Paulo, Brazil. From June 2002 to December 2002, he was a Visiting scholar at the Electromagnetic Communication Laboratory, Pennsylvania State University, University Park, where he is currently a Postdoctoral Researcher. His research interests include numerical techniques for modeling electromagnetic ﬁelds in complex enviroments (FDTD, FEM, ﬁnite volumes), network theory, and foundations of electromagnetics. Dr. Marrone is a Member of the International Compumag Society (ICS).

MARRONE AND MITTRA: THEORETICAL STUDY OF STABILITY CRITERIA FOR HYBRIDIZED FDTD ALGORITHMS

2167

Raj Mittra (S’54–M’57–SM’69–F’71–LF’96) is a Professor in the Electrical Engineering Department, Pennsylvania State University, University Park. He is also Director of the Electromagnetic Communication Laboratory, which is afﬁliated with the Communication and Space Sciences Laboratory of the Electrical Engineering Department. Prior to joining Penn State, he was a Professor in the Electrical and Computer Engineering Department, University of Illinois, Urbana-Champaign. He was a Visiting Professor at Oxford University, Oxford, U.K., and the Technical University of Denmark, Lyngby, Denmark. He is President of RM Associates, a consulting organization that provides services to industrial and governmental organizations. He has published more than 600 technical papers and more than 30 books or book chapters on various topics related to electromagnetics, antennas, microwaves, and electronic packaging. He has received three patents on communication antennas. Currently, he is the North American Editor of AEÜ. He has advised more than 80 Ph.D. students and about an equal number of M.S. students, and has mentored approximately 50 postdoctoral research associates and visiting scholars in the EMC Laboratory. His professional interests include the areas of communication antenna design, RF circuits, computational electromagnetics, electromagnetic modeling and simulation of electronic packages, EMC analysis, radar scattering, frequency-selective surfaces, microwave and millimeter-wave integrated circuits, and satellite antennas. Prof. Mittra received the Guggenheim Fellowship Award in 1965, the IEEE Centennial Medal in 1984, and the IEEE Millennium Medal in 2000. He is a past President of AP-S and was Editor of the IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION.

2168

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Communications______________________________________________________________________

Full-Wave Analysis of a Waveguide Printed Slot

Giorgio Montisci and Giuseppe Mazzarella

Abstract—In this paper, we propose a new waveguide slot array conﬁguration, in which the slots are printed on a copper-clad substrate. This conﬁguration can be obtained using a printed slot and therefore allows both a simpler realization of a slot array and permits its reuse when the pattern requirement changes. An innovative and highly effective method of moments procedure has been devised. It employs a spectral domain approach for the computation of the internal Green function. Our code has been validated against a commercial ﬁnite element method code, and the agreement is highly satisfactory. The code has then been used to obtain the scattering and radiating properties of the proposed printed slot. Such a slot behaves as a shunt admittance, so Elliott’s method can be used to design an array of printed slots. Index Terms—Method of moments (MoM), slot antennas, waveguide arrays.

Fig. 1.

Alternate phase feeding for the proposed slot array.

I. INTRODUCTION Waveguide slot arrays have a number of advantages over other microwave antenna systems. Among the others they have a very large efﬁciency, due to the use of closed guiding structures for feeding, and a considerable mechanical strength. On the other hand, they have no ﬂexibility since, once an array is realized, its electromagnetic behavior cannot be changed. If the antenna requirements change, the array itself must be changed. Moreover the array realization requires, mainly in the K-band, expensive procedures, such as electroerosion. In order to get a structure mechanically ﬂexible and easier to realize, but with all the other advantages of a waveguide slot array, it is possible to replace the upper waveguide wall with a copper-clad laminate, with a single metal layer. The slots are etched on this metal layer using one of the standard technologies for printed antennas. The electromagnetic behavior of the array then depends only on this slotted copper-clad laminate, which can be easily replaced at only a small fraction of the total array cost. Moreover such an array can be realized in an easier way and with the same, or even higher, accuracy than usual waveguide slot arrays. The copper layer can be placed below or above the dielectric laminate, and a metallic frame is required to sustain the laminate (see Fig. 1). If the copper layer is below the laminate, this frame will cause almost unpredictable diffraction effects (due to both space and surface waves). Therefore we propose here to insert the dielectric inside the waveguide. In this way the slots radiate in an open half-space, exactly as in a standard array. Of course, the power handling capability is smaller than in a standard array, but this can be a problem only in some applications. The proposed structure (see Fig. 1) consists of a metallic “comb-like” structure, formed by the bottom and sidewalls of the waveguides, covered with the copper-clad laminate. A standard slot array can be accurately designed using Elliott’s procedure [1]. This procedure can be directly applied to the structure proposed here, as soon as the self-admittance of a single slot is known, since the mutual coupling is exactly equal to the standard case one. At a ﬁrst glance, another difference appears, namely the internal coupling

Manuscript received May 13, 2003; revised September 29, 2003. The authors are with the Dipartimento di Ingegneria Elettrica ed Elettronica, Università di Cagliari, 09123 Cagliari, Italy (e-mail: giorgiom@diee.unica.it). Digital Object Identiﬁer 10.1109/TAP.2004.832333

between waveguides. This coupling has been evaluated with HFSS using two slots, each in one of two adjacent waveguides with a common internal dielectric layer: we consider a principal waveguide, which is fed with an incident TE10 mode and a parasite one. External boundary conditions have been chosen so that only internal coupling is involved in the analysis. We found that the voltage excitation of the slot in the parasite guide is about 030 dB with respect to that of the slot in the principal waveguide. Therefore this small coupling does not affect the array behavior and can be neglected expecially if the waveguides are fed with alternating phases [2]. This case has been analyzed, too, and the waveguide ﬁeld distribution and the slots excitation are the same as in the case of two isolated waveguide. To evaluate the slot self-admittance, we present here a specialized and very effective full-wave analysis, based on the method of moments (MoM) [3] which allows also to take exactly into account the thickness of the metallic plate in which the slots are cut. An accurate and effective procedure has been realized using entire domain basis functions [4]. The key point in every MoM procedure for waveguide slots is the evaluation of the waveguide magnetic ﬁeld inside the equivalent magnetic currents. As it is well known [5] the relevant Green function is singular and this can lead to numerical instabilities and poor convergence, both in the space domain (expressing the Green function as a modal series) and in the spectral domain. Since the F potential [6] is less singular in a magnetic current the elements of the MoM matrix have been expressed in terms of F . Previous papers on slots in inhomogeneous waveguide [7] employ a modal series for the Green function. In this paper a spectral domain approach has been used since it is of far simple use in presence of an air-dielectric boundary. Our MoM code has been validated using a commercial ﬁnite element method (FEM) code (Ansoft HFSS). The agreement is highly satisfactory, though our tailored code is signiﬁcantly more effective than HFSS, which is a general-purpose commercial software and is not optimized for the proposed structure. II. PROBLEM FORMULATION Using the alternate phase feeding [2] for the array, we can assume that short circuits exist at the upper corners of two adjacent waveguides (see Fig. 1). Therefore, the slot self-admittance is not affected by the adjacent waveguides, and, in order to evaluate it, we can consider a longitudinal slot in an isolated waveguide (Fig. 2). Let t be the dielectric substrate thickness and "r its permittivity. The fundamental mode of

0018-926X/04$20.00 © 2004 IEEE

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2169

the z -directed magnetic currents are used as unknowns, and each current (M i ; M e ) is expressed as truncated sinusoidal series with respect to z N N p a sin (2) M = (z + L) iz = a mp (z): p p 2L p=1 n=1 In (2) the apex can assume the values i; e. Expansions (2) are then substituted into (1) and the resulting equations are scalar multiplied by mq and integrated over the apertures of the slots. Only a few terms in (2) are needed, at least for resonant or almost-resonant slots, so that the resulting linear system in the unknowns ai ; ae is small and quite well conditioned. p p Both the internal and external magnetic ﬁelds in (1) are expressed using the potential F

H Apq = j!"

1 = j!"F + j! rr 1 F 1

(3)

so the MoM matrix element is

Slot

F p 1mq ds 0 2 (r 1 F p ) (r 1 mq ) dS k0 slot

(4)

Fig. 2.

Geometry of the Slot. (a) Front view and (b) top view.

this waveguide impinges upon the slot and is scattered and radiated by the slot itself. In order to compute this interaction, and obtain the slot equivalent circuit, both slot apertures (6i ; 6e ) are replaced by equivalent surface magnetic currents, backed by perfect electric conductors. According to the equivalence theorem, these currents M i ; M e , are proportional to the tangential electric ﬁeld on the apertures, and are therefore unknowns. In order to compute them, we enforce, at all interfaces 6i ; 6e , the continuity of the tangential magnetic ﬁeld

wherein F p is the potential produced by mp . The external and slot region integrals are computed respectively using a standard plane-wave spectrum representation of the external ﬁeld and the admittance matrix formalism, much as done in [6]. It is worth noting that mp are the modes of the slot region (considered as a waveguide) and therefore the admittance matrix allows to include exactly the wall thickness effect. The internal terms in (4) have been computed in the spectral domain expressing both the potentials and the currents as a Fourier transform respect to the waveguide axis coordinate and as a (truncated) Fourier series respect to the transverse direction. In this latter case, the expansion is actually an eigenmodes one. Due to the waveguide inhomogeneity the z -directed magnetic cur~ ~ ~ rent M produces both Fy and Fz components

~ Mz (w) = ~ Fy (x; y; w) =

where

1 1 1

n=0 n=0 n=0

~ Mzn (w) cos ~ Fyn (y; w) sin ~ Fzn (y; w) cos

nx a nx a nx a

(5)

in H w [M i ] + in H inc = in H s [M i ; M e ] on 6i i n H s [M i ; M e ] = i n H e [M e ]

2 2

2

2

~ Fz (x; y; w) = ~ F (x; y; w) =

(6)

2

on 6e

(1)

1 +1 F (x; y; z) exp(jwz) dz 2 01

(7)

wherein H w ; H s ; H e are the waveguide region, slot region and external region magnetic ﬁeld, H inc is the incident magnetic ﬁeld of the fundamental mode and in the normal to each surface. As usual for radiating slots [8], we assume that the slots radiate into a half-space bounded by an inﬁnite, perfectly conducting, ground plane. All the involved regions are regular and need not to be discretized in a MoM procedure, since their Green function is known in (a simple) closed form. In (1) we have also included explicitly the currents that produce each ﬁeld. Since each magnetic ﬁeld is an integral operator acting on its source, (1) is already cast as a set of coupled integral equations. The actual form of these operators is of course required, and their computation will be discussed later in this section. The resulting integral equations are then solved by the MoM, to obtain the unknown currents. To do this, all unknowns are expressed as a linear combination of the same basis functions. Usually the slots are narrow enough to neglect the longitudinal component of the electric ﬁeld on them [9]. Therefore only

~ ~ and likely for M . The use of expansions with different parity for Fy ~ and Fz is dictated by the boundary conditions on the lateral walls of the waveguide. Since the effect of the unknowns magnetic currents has been included in the continuity condition of the electric ﬁeld at the source location, then we can compute the potential F by solving two equal ~ ~ homogeneous Helmholtz equations, one for Fy and the other for Fz , which become: ~ @ Fn @y2 ~ @ Fn @y2 + 2 n Fn = 0 0 y d 0 ~ + 2 F n = 0 d y b n~

wherein

n 2 2 2n = k0 0 w2 0 0 a 2 = "r k2 0 w2 0 n 2 : n 0 a

(8)

2170

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 3. Comparison between our MoM analysis procedure and Ansoft HFSS (t = 0:3 mm, " = 2:2; slot length = 13:2 mm, slot width = 1 mm, slot oset = 4 mm).

Fig. 4. Resonant length of the slot as a function of the slot offset (frequency = 9:7 GHz, t = 0:3 mm, slot width = 1 mm).

**The general solution for (8) is
**

~ Fn = A0n cos(0n y) + B0n cos(0n y) 0 y d ~ Fn = An cos(n (b 0 y)) (9) + Bn cos(n (b 0 y)) d y b wherein the coefﬁcients A0n ; B0n ; An , and Bn are computed as a

function of the unknowns magnetic currents, by enforcing the continuity and boundary conditions. The MoM matrix elements are then computed in the spectral domain using the Parseval identity in all the internal terms of (4). The integration has been performed on the real axis, following the suggestions of [10]. III. RESULTS The presented procedure has been applied to a number of different cases. We discuss here only a typical set of results, using a WR90 waveguide (A = 22:86 mm, B = 10:16 mm) partially ﬁlled with a dielectric slab. The ground plane thickness is equal to 50 m and the losses have been neglected. Actually, this is correct at the chosen X-band frequency but could be questionable, at least for the dielectric losses, at higher frequencies. In any case such losses can be taken into account, if required, by using a complex permittivity in the Green function and in the pole computation. Our pole-search procedure can be directly generalized to look for complex poles. The results of our code have been validated through a comparison with the results of a commercial FEM code (HFSS 6.0). As it is well known a longitudinal slot in an empty waveguide behaves as a shunt admittance and this behavior is retained also for a dielectric-backed printed slot since our simulations show that the magnitude of j1+S11 0 S12 j is very small [9]. Therefore only the normalized slot admittance has been considered for the comparison. The agreement between our code and the commercial FEM one is very good (see Fig. 3). It is worth noting that our code is about one order of magnitude faster than Ansoft HFSS and requires at least two orders of magnitude less memory than HFSS. In Fig. 4 the normalized resonant length of the slot has been shown, as a function of the slot offset, for different values of the substrate dielectric constant at the design frequency of 9.7 GHz. Finally in Fig. 5–7 the normalized slot admittance is reported respectively as a function of the slot offset, the substrate dielectric constant and the dielectric substrate thickness.

Fig. 5. Normalized slot admittance at resonance as a function of the slot offset (t = 0:3 mm, " = 2:2; slot length = 13:2 mm, slot width = 1 mm).

Fig. 6. Normalized slot admittance at resonance as a function of " (t = 0:3 mm, Oset = 4:0 mm, slot length = 13:2 mm, slot width = 1 mm).

In all the simulations we used seven expansion functions for each slot [N = 7 in (2)], since we found that this choice is enough to achieve a good convergence.

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2171

**Dual Polarized Wide-Band Aperture Stacked Patch Antennas
**

K. Ghorbani and R. B. Waterhouse

Abstract—A wide-band, dual polarized printed antenna is designed and developed in this paper. The antenna is based upon an aperture stacked patch layout and incorporates a simple dual-layered feeding technique to achieve dual-polarized radiation. The printed antenna has a measured 10 dB return loss bandwidth of 52% and an isolation between the excitation ports of greater than 39 dB over this frequency range. The gain of the antenna is 7 4 dBi 0 4 dB and the typical issues associated with incorporating an aperture excited solution are resolved by using a cross-shaped reﬂector patch to ensure the front-to-back ratio is greater than 20 dB. Index Terms—Aperature antennas, microstrip antennas, polarization. Fig. 7. Normalized slot admittance at resonance as a function of the dielectric substrate thickness t (Oset = 4:0 mm, " = 2:2; slot length = 13:2 mm, slot width = 1 mm).

I. INTRODUCTION As mobile communication services become more sophisticated, the need for larger spectral bandwidth for delivery of these provisions is apparent. It was because of this trend the International Telecommunications Union recommended that for IMT-2000 several frequency bands be utilized over almost a 50% range centered near 2 GHz [1]. Thus, base station antennas must be able to operate efﬁciently over ever increasing frequency ranges than were originally required. Also, to further enhance the link performance between the base station and wireless user, diversity techniques have been proposed and subsequently used at the base station. Polarization diversity is a useful technique to reduce the detrimental effects of multipath fading and is a common procedure utilized at base stations of most mobile communication systems. Printed antennas have many salient features that have made them primary candidates for cellular base stations including their inherent ease of mass construction and their conformal nature. However, in their original form and subsequent bandwidth enhancement procedures, such as stacking patches [2], or for the case of an aperture excited patch, using a large slot [3] it is difﬁcult to achieve the previously mentioned bandwidth and therefore difﬁcult to provide a single element solution. A single element solution allows for simple, low cost arrays to be developed to provide the necessary sectoral radiation patterns typically required at a base station terminal. Over past few years there have been several reported printed antennas that can achieve 50% bandwidth: Aperture stacked patches (ASPs) [4], Quasi Yagi–Uda printed antennas [5] and Suspended patches with three-dimensional feeds [6], to name a few. Each of these printed antennas has their relative ﬁgures of merit and can readily satisfy the necessary bandwidth requirements for IMT-2000. However, of these solutions, the easiest to achieve good quality dual polarization is the aperture-stacked patch, due to the inherent polarization purity of a thin slot excited patch antenna. There have been several dual polarized slot coupled patch antennas investigated recently [7]–[11], however the antennas considered are relatively narrow band, with bandwidths less than or equal to 25%. In this paper, we present the design and develop of a dual polarized broadband printed antenna capable of operation over a 50% impedance

IV. CONCLUSION A full-wave MoM procedure for printed waveguide slots has been presented and validated against a commercial FEM code. From it both the tangential electric ﬁeld and the scattering data of a printed slot can be obtained. From the latter we derive that a waveguide printed slot behaves as a shunt element, as a standard waveguide longitudinal slot does. The MoM procedure has been devised in order to exploit the simple geometric form of a slot to realize an accurate and very effective code. As a result, an accurate electromagnetic characterization of this structure can be obtained, with the scattering data recast in a form suitable to be used in slot array design procedures based on Elliot’s method. This allows to design slot arrays with the slots printed on the aperture plane, with the same accuracy of a standard waveguide slot array. As a consequence, the new slot conﬁguration proposed here can be effectively used to allow an easier, less expensive array realization, which permits a simple reuse of the array itself when the pattern requirements changes. REFERENCES

[1] R. S. Elliott, “An improved design procedure for small arrays of shunt slots,” IEEE Trans. Antennas Propagat., vol. AP-31, pp. 48–53, 1983. [2] Y. Kimura, T. Hirano, J. Hirokawa, and M. Ando, “Chokes for alternating-phase fed single-layer slotted waveguide arrays,” Proc. IEEE Int. Antennas Propagat. Symp. Dig., vol. 38, pp. 82–85, July 2000. [3] R. F. Harrington, Field Computation by Moment Methods. New York: IEEE Press, 1993. [4] R. W. Lyon and A. J. Sangster, “Efﬁcient moment method analysis of radiating slots in a thick-walled rectangular waveguide,” Proc. Inst. Elect. Eng., pt. H, vol. 128, no. 4, Aug. 1981. [5] R. E. Collin, Field Theory of Guided Waves. New York, NY: IEEE Press, 1990. [6] G. Mazzarella and G. Montisci, “A rigorous analysis of dielectric-covered narrow longitudinal shunt slots with ﬁnite wall thickness,” Electromagn., vol. 19, pp. 407–418, 1999. [7] J. Joubert and D. A. McNamara, “Analysis of radiating slots in a rectangular waveguide loaded with a dielectric slab,” IEEE Trans. Antennas Propagat., vol. 41, Sept. 1993. [8] S. R. Rengarajan, “Compound radiating slot in a broad wall of a rectangular waveguide,” IEEE Trans. Antennas Propagat., vol. AP-37, pp. 1116–1124, 1989. [9] T. Vu Khac and C. T. Carson, “Impedance properties of longitudinal slot antenna in the broad face of rectangular waveguide,” IEEE Trans. Antennas Propagat., vol. AP-21, pp. 708–710, 1973. [10] J. R. Mosig, R. C. Hall, and F. E. Gardiol, Handbook of Microstrip Antennas, J. R. James and P. S. Hall, Eds, London, U.K.: Peregrinus, 1993.

Manuscript received February 13, 2003; revised September 17, 2003. K. Ghorbani is with the School of Electrical and Computer Engineering, RMIT University, Melbourne, VIC 3000, Australia. R. B. Waterhouse was with Pharad Technologies, North Fitzroy, VIC 3068 Australia. He is now with Pharad Technologies, Baltimore, MD 21227 USA (e-mail: r.waterhouse@ieee.org). Digital Object Identiﬁer 10.1109/TAP.2004.832484

0018-926X/04$20.00 © 2004 IEEE

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2171

**Dual Polarized Wide-Band Aperture Stacked Patch Antennas
**

K. Ghorbani and R. B. Waterhouse

Abstract—A wide-band, dual polarized printed antenna is designed and developed in this paper. The antenna is based upon an aperture stacked patch layout and incorporates a simple dual-layered feeding technique to achieve dual-polarized radiation. The printed antenna has a measured 10 dB return loss bandwidth of 52% and an isolation between the excitation ports of greater than 39 dB over this frequency range. The gain of the antenna is 7 4 dBi 0 4 dB and the typical issues associated with incorporating an aperture excited solution are resolved by using a cross-shaped reﬂector patch to ensure the front-to-back ratio is greater than 20 dB. Index Terms—Aperature antennas, microstrip antennas, polarization. Fig. 7. Normalized slot admittance at resonance as a function of the dielectric substrate thickness t (Oset = 4:0 mm, " = 2:2; slot length = 13:2 mm, slot width = 1 mm).

I. INTRODUCTION As mobile communication services become more sophisticated, the need for larger spectral bandwidth for delivery of these provisions is apparent. It was because of this trend the International Telecommunications Union recommended that for IMT-2000 several frequency bands be utilized over almost a 50% range centered near 2 GHz [1]. Thus, base station antennas must be able to operate efﬁciently over ever increasing frequency ranges than were originally required. Also, to further enhance the link performance between the base station and wireless user, diversity techniques have been proposed and subsequently used at the base station. Polarization diversity is a useful technique to reduce the detrimental effects of multipath fading and is a common procedure utilized at base stations of most mobile communication systems. Printed antennas have many salient features that have made them primary candidates for cellular base stations including their inherent ease of mass construction and their conformal nature. However, in their original form and subsequent bandwidth enhancement procedures, such as stacking patches [2], or for the case of an aperture excited patch, using a large slot [3] it is difﬁcult to achieve the previously mentioned bandwidth and therefore difﬁcult to provide a single element solution. A single element solution allows for simple, low cost arrays to be developed to provide the necessary sectoral radiation patterns typically required at a base station terminal. Over past few years there have been several reported printed antennas that can achieve 50% bandwidth: Aperture stacked patches (ASPs) [4], Quasi Yagi–Uda printed antennas [5] and Suspended patches with three-dimensional feeds [6], to name a few. Each of these printed antennas has their relative ﬁgures of merit and can readily satisfy the necessary bandwidth requirements for IMT-2000. However, of these solutions, the easiest to achieve good quality dual polarization is the aperture-stacked patch, due to the inherent polarization purity of a thin slot excited patch antenna. There have been several dual polarized slot coupled patch antennas investigated recently [7]–[11], however the antennas considered are relatively narrow band, with bandwidths less than or equal to 25%. In this paper, we present the design and develop of a dual polarized broadband printed antenna capable of operation over a 50% impedance

IV. CONCLUSION A full-wave MoM procedure for printed waveguide slots has been presented and validated against a commercial FEM code. From it both the tangential electric ﬁeld and the scattering data of a printed slot can be obtained. From the latter we derive that a waveguide printed slot behaves as a shunt element, as a standard waveguide longitudinal slot does. The MoM procedure has been devised in order to exploit the simple geometric form of a slot to realize an accurate and very effective code. As a result, an accurate electromagnetic characterization of this structure can be obtained, with the scattering data recast in a form suitable to be used in slot array design procedures based on Elliot’s method. This allows to design slot arrays with the slots printed on the aperture plane, with the same accuracy of a standard waveguide slot array. As a consequence, the new slot conﬁguration proposed here can be effectively used to allow an easier, less expensive array realization, which permits a simple reuse of the array itself when the pattern requirements changes. REFERENCES

[1] R. S. Elliott, “An improved design procedure for small arrays of shunt slots,” IEEE Trans. Antennas Propagat., vol. AP-31, pp. 48–53, 1983. [2] Y. Kimura, T. Hirano, J. Hirokawa, and M. Ando, “Chokes for alternating-phase fed single-layer slotted waveguide arrays,” Proc. IEEE Int. Antennas Propagat. Symp. Dig., vol. 38, pp. 82–85, July 2000. [3] R. F. Harrington, Field Computation by Moment Methods. New York: IEEE Press, 1993. [4] R. W. Lyon and A. J. Sangster, “Efﬁcient moment method analysis of radiating slots in a thick-walled rectangular waveguide,” Proc. Inst. Elect. Eng., pt. H, vol. 128, no. 4, Aug. 1981. [5] R. E. Collin, Field Theory of Guided Waves. New York, NY: IEEE Press, 1990. [6] G. Mazzarella and G. Montisci, “A rigorous analysis of dielectric-covered narrow longitudinal shunt slots with ﬁnite wall thickness,” Electromagn., vol. 19, pp. 407–418, 1999. [7] J. Joubert and D. A. McNamara, “Analysis of radiating slots in a rectangular waveguide loaded with a dielectric slab,” IEEE Trans. Antennas Propagat., vol. 41, Sept. 1993. [8] S. R. Rengarajan, “Compound radiating slot in a broad wall of a rectangular waveguide,” IEEE Trans. Antennas Propagat., vol. AP-37, pp. 1116–1124, 1989. [9] T. Vu Khac and C. T. Carson, “Impedance properties of longitudinal slot antenna in the broad face of rectangular waveguide,” IEEE Trans. Antennas Propagat., vol. AP-21, pp. 708–710, 1973. [10] J. R. Mosig, R. C. Hall, and F. E. Gardiol, Handbook of Microstrip Antennas, J. R. James and P. S. Hall, Eds, London, U.K.: Peregrinus, 1993.

Manuscript received February 13, 2003; revised September 17, 2003. K. Ghorbani is with the School of Electrical and Computer Engineering, RMIT University, Melbourne, VIC 3000, Australia. R. B. Waterhouse was with Pharad Technologies, North Fitzroy, VIC 3068 Australia. He is now with Pharad Technologies, Baltimore, MD 21227 USA (e-mail: r.waterhouse@ieee.org). Digital Object Identiﬁer 10.1109/TAP.2004.832484

0018-926X/04$20.00 © 2004 IEEE

2172

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

Fig. 1. Geometry of dual polarized ASP antenna.

bandwidth, with polarization isolation greater than 35 dB and a gain of 7 dBi 6 0:4 dB over this band of frequencies. Our solution utilizes a dual polar, broadband reﬂector patch below the feed/antenna groundplane which improves the front-to-back ratio (FBR) of the element to more than 20 dB across the band of interest, a key aspect for sectorized cellular base stations and an issue with most aperture solutions. II. ANTENNA CONFIGURATION AND DESIGN The geometry of dual polarized ASP antenna is shown in Fig. 1. As can be seen from this diagram, 2 rectangular patches are mounted on multiple dielectric layers (in this case 2 each) located above a feed layer situated above a ground-plane. Another feed layer used to provide the orthogonal polarization exists below the ground-plane. Power is coupled to the patch radiators from the feed networks via a cross-shaped slot in the common ground-plane. Beyond the lower feed substrate are a foam spacer and then a cross-shaped patch use to enhance the front-to-back ratio. The strategy outlined in [4] was used to design a linearly polarized aperture-stacked patch with the appropriate bandwidth (50%). Ensemble 6.1, a planar method of moments (MoM) ﬁeld simulator, was utilized as the simulation tool. Symmetry is a very important aspect in the design of the antenna to achieve a high degree of isolation between the two polarization ports over all operation frequencies. By using a centered crossed slot to feed the patch element, a high degree of symmetry can be maintained, however this necessitates a balanced feed network for each polarization excitation. Incorporating a single feed-line on each polarization of the crossed slot would introduce asymmetry, which can degrade the isolation between the two ports. Using small slots (as in [9], [10]) is not an option as the bandwidth requirement is too large for such a solution. In the proposed conﬁguration each feed network is placed on a separate layer on opposite sides of the ground plane of the antenna as can be seen in Fig. 1. This feed layout is based on that presented in [8], [11] and avoids using an air-bridge (as in the case where both balanced feeds for each polarization are etched on the same board). Using separate feed layers improves the isolation between the ports due to limiting spurious radiation from the air-bridge as well as reduces the construction cost of the antenna. The input impedance for the two ports will differ slightly as the upper feed network can couple to the patch elements. However this is typically negligible as electrically thick layers are used between the slot and the patches of the ASP and so direct coupling from the feed-line to the patches should be minimal. Having said that, the height of two feeding layers must be chosen so that the mutual coupling between the slot and patch elements remains the same for both polarization ports. The proposed ASP consists of four

Fig. 2. Measured and Ensemble 6.1 return loss of dual polarized ASP antenna: (a) Port 1, (b) Port 2 (parameters: "r1 = 4:5; h1 = 1:58 mm, "r2 = 1; h2 = 35 mm, "r3 = 2:2; h3 = 0:508 mm, "r4 = 2:2; h4 = 0:508 mm, "r5 = 1:07; h5 = 9 mm, "r6 = 2:2; h6 = 3:175 mm, "r7 = 1; h7 = 13 mm, "r4 = 2:2; h8 = 1:58 mm, W1REF = 10 mm, L1REF = 85 mm, W2REF = 10 mm, L2REF = 85 mm, W1F = 0:4 mm, L1STUB = 12:3 mm, W2F = 0:4 mm, L2STUB = 13:1 mm, W2SL = 10 mm, L2SL = 46 mm, W2SL = 10 mm, L2SL = 46 mm, W1 = 47 mm, L1 = 47 mm, W2 = 45 mm, L2 = 45 mm.).

Fig. 3. Measured and Ensemble 6.1 isolation between Port 1 and Port 2.

layers of Rogers Duroid 5880 laminate ("r3 ; "r4 ; "r6 ; "r8 ), one layer of FR4 dielectric material ("r1 ) and three air/foam layers ("r2 ; "r5 ; "r7 ).

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2173

Fig. 4. Measured radiation patterns of dual polarized ASP antenna at 2.0 GHz: (a) H-plane Port 1, (b) H-plane Port 2, (c) E-plane Port 1, (d) E-plane Port 2.

ASP antennas normally produce a relatively low front to back ratio (typically 10–14 dB [4]) due to the resonant aperture used to enhance the bandwidth of the aperture-coupled patches. In order to increase the FBR a microstrip antenna element as a reﬂector (based on the method described in [12]) was used. The length of the reﬂector is chosen so the phase of the radiated ﬁeld due to the reﬂector is approximately opposite to the radiated ﬁeld from the aperture. Changing the spacing (between the reﬂector and the aperture) and width of the reﬂector can control the magnitude of the radiated ﬁeld. The total cancellation will occur when the magnitudes of the radiated ﬁeld from the aperture and the reﬂector are approximately equal while the phases are opposite. The spacing between the aperture and the reﬂector should also be sufﬁcient so the reﬂector has negligible affect on the input impedance of the antenna. For the application considered here a cross-shaped element was chosen to ensure minimal back radiation in both polarizations was achieved. To reduce the cost of the antenna and to make it more rigid in construction, an FR4 substrate was used for the microstrip patch reﬂector. The spacing between the reﬂector patch and the lower feedline was optimized in order to achieve a FBR of greater than 20 dB without affecting the input impedance behavior of both polarization networks. This optimization was done by simulation. The procedure outlined in [12] was followed here. This procedure is pretty robust and so for the sake of brevity, we simply referenced it in this paper. The optimization procedure is conducted over the entire 10 dB return loss bandwidth. Plots of the affect of spacing on the FBR are given in [12] and these were consistent with the study conducted here. Typically for

the optimum design the FBR peaks at a frequency between the lower band-edge and the center frequency. Slightly increasing the spacing from this value can enhance the peak FBR at the expense of the achievable 20 dB FBR bandwidth. The reﬂector patch enhanced the FBR by more than 5 dB across the entire band to greater than 20 dB. III. RESULTS AND DISCUSSION Fig. 2(a) and (b) show the measured and calculated return loss from Port 1 and Port 2 of the dual polarized ASP antenna, respectively (port 1 corresponds to feed 1 in Fig. 1). As can be seen from these responses, there is good agreement between the theoretical and experimental results. 10 dB return loss bandwidths of greater than 52% for both polarizations were obtained. The lower edge of the matched band is slighter lower for Port 2 than Port 1. This can be attributed to the height of the upper feed substrate and so the coupling between the slot and the patch element for this polarization is stronger than between the slot and the patch element for the Port 1 excitation. The shift is less than 1%. Fig. 3 shows the measured and calculated isolation between two ports. An isolation of more than 39 dB over the impedance-matched band has been achieved. As mentioned before, the two feed networks are separated by the ground plane, thereby eliminating any cross coupling between the microstrip-lines in the two networks. The magnitude of isolation presented here is comparable with the previous dual polarized aperture coupled patches presented in the literature and is achieved over a considerably larger impedance bandwidth.

2174

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

patch ensures the FBR of the antenna is greater than 20 dB over the entire 50% matched-impedance bandwidth. The measured results are in very good agreement with the theoretical predictions. The antenna is simple to manufacture and is suited to being incorporated into a mobile communications base station array. REFERENCES

[1] Handbook on Antennas in Wireless Communications, L. Godara, Ed., CRC Press, New York, 2001. [2] R. B. Waterhouse, “Design of probe-fed stacked patches,” IEEE Trans. Antennas Propagat., vol. 47, pp. 1780–1784, Dec. 1999. [3] J. F. Zurcher, “The SSFIP: A global concept for high performance broadband planar antennas,” Electron. Lett., vol. 24, pp. 1433–1435, Nov. 1988. [4] S. D. Targonski, R. B. Waterhouse, and D. M. Pozar, “A wideband aperture coupled stacked patch antenna using thick substrates,” Electron. Lett., vol. 32, pp. 1941–1942, Oct. 1996. [5] N. Kaneda, W. R. Deal, Y. Qian, R. Waterhouse, and T. Itoh, “A broadband planar quasiyagi antenna,” IEEE Trans. Antennas Propagat., vol. 50, pp. 1158–1160, Aug. 2002. [6] N. Herscovici, “A wide-band single-layer patch antenna,” IEEE Trans. Antennas Propagat., vol. 46, pp. 471–474, Apr. 1998. [7] E. Aloni and R. Kastner, “Analysis of a dual circularly polarized microstrip antenna fed by crossed slots,” IEEE Trans. Antennas Propagat., vol. 42, pp. 1053–1058, Aug. 1994. [8] M. Yamazaki, E. T. Rahardjo, and M. Haneishi, “Construction of a slotcoupled planar antenna for dual polarization,” Electron. Lett., vol. 30, pp. 1814–1815, 1994. [9] S. Hienonen, A. Lehto, and A. V. Raisanen, “Simple broadband dual-polarized aperture-coupled microstrip antenna,” in Proc. IEEE APS Symp. Digest, Orlando, FL, June 1999, pp. 1228–1231. [10] S. C. Gao, L. W. Li, P. Gardner, and P. S. Hall, “Wideband dual-polarized microstrip patch antenna,” Electron. Lett., vol. 37, pp. 1213–1214, Sept. 2001. [11] J. R. Sanford and A. Tengs, “A two substrate dual polarized aperture coupled patch,” Proc. IEEE Trans. Antennas and Propagation Symp. Digest, pp. 1544–1547, 1996. [12] R. B. Waterhouse, D. Novak, A. Nirmalathas, and C. Lim, “Broadband printed sectoral coverage antennas for millimeter-wave wireless applications,” IEEE Trans. Antennas Propagat., vol. 50, pp. 12–16, Jan. 2002.

Fig. 5.

Measured and Ensemble 6.1 gain of dual polarized ASP antenna.

A full radiation characterization of the antenna was conducted and the results at the band edges (1.5 and 2.5 GHz) as well as the center frequency (2 GHz). The results at 2 GHz are shown in Fig. 4. The cross-polarization levels for this antenna are considerably low, with a higher contribution in theE-planeofeachexcitedport.Thiscanbeattributedtotheinteractionof excited surface waves, the slot and the edges of the truncated substrates. Such an interaction is always more prevalent in the E-plane. The crosspolar level for each excitation and in each plane is not less than 12 dB below the copolar value for scan angles up to 660 , angles consistent with a three-sectored base station. Of course the worse cases are at the extreme scan angles,wherethecopolarradiateddecaysmorerapidly. The cross-polarization levels are the highest at 1.5 GHz, which is due to the size of the ground-plane (here 105 mm 2 105 mm). The ﬁnite size of the ground-planewillhaveagreaterimpactatlowerfrequencies,asitsrelative size issmaller. A measured FBR(the ratio ofpower at broadside ( = 0 ) and = 180 ) of approximately 20 dB was obtained for the antenna over the entire impedance bandwidth. The gain of the antenna was greater than 7 dBi over the entire VSWR < 2 : 1 band. A comparison between theory and experiment are shown in Fig. 5. As can be seen from this ﬁgure there is very good agreement between the predicted and measured results. The gain of the dual-polarized ASP drops off at higher frequencies, beyond the operation bandwidth of the antenna due to the slot becoming the dominant radiation mechanism [12]. There are many design parameters in the dual polarized ASP with reduced backward directed radiation. Because of this, it is important to ensure that the overall performance of the antenna is not sensitive to parameter variations (tolerances in the dielectric constants of the materials, or etching procedures). We conducted a parameter sensitivity investigation via simulation (a three-sigma test) on the proposed antenna, allowing all the parameters associated with antenna (conductor/slot dimensions (60:02 mm, materials ("r 6 0:02; h 6 0:025 mm), to be varied within the material tolerances and etching tolerances. The ﬁndings showed minimal impact on the return loss bandwidth and FBR performance. There were slight increases in the peaks of the return loss within the 10 dB return loss band by 1.5 dB when all the conductors were over-etched by 0.02 mm, however, these peaks were still below the 10 dB requirement. These ﬁndings show that the antenna is very tolerant to moderate variations in all the parameters. IV. CONCLUSION In this paper, a dual-polarized wide-band patch antenna has been presented. The antenna has a VSWR < 2 : 1 bandwidth of greater than 50%, an isolation between ports of more than 39 dB and a gain of 7:4 dBi 6 0:4 dB over this frequency range. A cross-shaped reﬂector

**Resonant Frequency of Equilateral Triangular Microstrip Antenna With and Without Air Gap
**

Debatosh Guha and Jawad Y. Siddiqui

Abstract—A tunable equilateral triangular microstrip patch (ETMP) antenna with a variable air gap between the substrate and the ground plane has been studied theoretically and experimentally. Calculated resonant frequencies for different air gap heights have been veriﬁed with measurements of a coax-fed antenna. The tunability of the antenna as a function of the air gap height is studied theoretically showing over 200% tunable range of an ETMP with 50 mm side length printed on a substrate with = 10 5. The computed results for the antennas with zero air gap height are compared with a standard spectral domain moment method analysis supported by other previously reported experiments. Very close agreement is revealed in all comparisons. Index Terms—Microstrip antenna, triangular microstrip patch, tunable microstrip antenna. Manuscript received October 22, 2002; revised June 30, 2003. This work was supported by the Center of Advanced Study in Radio Physics and Electronics, University of Calcutta. The authors are with the Institute of Radio Physics and Electronics, University of Calcutta, Calcutta 700 009, India (e-mail: dgirpe@yahoo.co.in; siddiqui@cal.vsnl.net.in). Digital Object Identiﬁer 10.1109/TAP.2004.832504

0018-926X/04$20.00 © 2004 IEEE

2174

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

patch ensures the FBR of the antenna is greater than 20 dB over the entire 50% matched-impedance bandwidth. The measured results are in very good agreement with the theoretical predictions. The antenna is simple to manufacture and is suited to being incorporated into a mobile communications base station array. REFERENCES

[1] Handbook on Antennas in Wireless Communications, L. Godara, Ed., CRC Press, New York, 2001. [2] R. B. Waterhouse, “Design of probe-fed stacked patches,” IEEE Trans. Antennas Propagat., vol. 47, pp. 1780–1784, Dec. 1999. [3] J. F. Zurcher, “The SSFIP: A global concept for high performance broadband planar antennas,” Electron. Lett., vol. 24, pp. 1433–1435, Nov. 1988. [4] S. D. Targonski, R. B. Waterhouse, and D. M. Pozar, “A wideband aperture coupled stacked patch antenna using thick substrates,” Electron. Lett., vol. 32, pp. 1941–1942, Oct. 1996. [5] N. Kaneda, W. R. Deal, Y. Qian, R. Waterhouse, and T. Itoh, “A broadband planar quasiyagi antenna,” IEEE Trans. Antennas Propagat., vol. 50, pp. 1158–1160, Aug. 2002. [6] N. Herscovici, “A wide-band single-layer patch antenna,” IEEE Trans. Antennas Propagat., vol. 46, pp. 471–474, Apr. 1998. [7] E. Aloni and R. Kastner, “Analysis of a dual circularly polarized microstrip antenna fed by crossed slots,” IEEE Trans. Antennas Propagat., vol. 42, pp. 1053–1058, Aug. 1994. [8] M. Yamazaki, E. T. Rahardjo, and M. Haneishi, “Construction of a slotcoupled planar antenna for dual polarization,” Electron. Lett., vol. 30, pp. 1814–1815, 1994. [9] S. Hienonen, A. Lehto, and A. V. Raisanen, “Simple broadband dual-polarized aperture-coupled microstrip antenna,” in Proc. IEEE APS Symp. Digest, Orlando, FL, June 1999, pp. 1228–1231. [10] S. C. Gao, L. W. Li, P. Gardner, and P. S. Hall, “Wideband dual-polarized microstrip patch antenna,” Electron. Lett., vol. 37, pp. 1213–1214, Sept. 2001. [11] J. R. Sanford and A. Tengs, “A two substrate dual polarized aperture coupled patch,” Proc. IEEE Trans. Antennas and Propagation Symp. Digest, pp. 1544–1547, 1996. [12] R. B. Waterhouse, D. Novak, A. Nirmalathas, and C. Lim, “Broadband printed sectoral coverage antennas for millimeter-wave wireless applications,” IEEE Trans. Antennas Propagat., vol. 50, pp. 12–16, Jan. 2002.

Fig. 5.

Measured and Ensemble 6.1 gain of dual polarized ASP antenna.

A full radiation characterization of the antenna was conducted and the results at the band edges (1.5 and 2.5 GHz) as well as the center frequency (2 GHz). The results at 2 GHz are shown in Fig. 4. The cross-polarization levels for this antenna are considerably low, with a higher contribution in theE-planeofeachexcitedport.Thiscanbeattributedtotheinteractionof excited surface waves, the slot and the edges of the truncated substrates. Such an interaction is always more prevalent in the E-plane. The crosspolar level for each excitation and in each plane is not less than 12 dB below the copolar value for scan angles up to 660 , angles consistent with a three-sectored base station. Of course the worse cases are at the extreme scan angles,wherethecopolarradiateddecaysmorerapidly. The cross-polarization levels are the highest at 1.5 GHz, which is due to the size of the ground-plane (here 105 mm 2 105 mm). The ﬁnite size of the ground-planewillhaveagreaterimpactatlowerfrequencies,asitsrelative size issmaller. A measured FBR(the ratio ofpower at broadside ( = 0 ) and = 180 ) of approximately 20 dB was obtained for the antenna over the entire impedance bandwidth. The gain of the antenna was greater than 7 dBi over the entire VSWR < 2 : 1 band. A comparison between theory and experiment are shown in Fig. 5. As can be seen from this ﬁgure there is very good agreement between the predicted and measured results. The gain of the dual-polarized ASP drops off at higher frequencies, beyond the operation bandwidth of the antenna due to the slot becoming the dominant radiation mechanism [12]. There are many design parameters in the dual polarized ASP with reduced backward directed radiation. Because of this, it is important to ensure that the overall performance of the antenna is not sensitive to parameter variations (tolerances in the dielectric constants of the materials, or etching procedures). We conducted a parameter sensitivity investigation via simulation (a three-sigma test) on the proposed antenna, allowing all the parameters associated with antenna (conductor/slot dimensions (60:02 mm, materials ("r 6 0:02; h 6 0:025 mm), to be varied within the material tolerances and etching tolerances. The ﬁndings showed minimal impact on the return loss bandwidth and FBR performance. There were slight increases in the peaks of the return loss within the 10 dB return loss band by 1.5 dB when all the conductors were over-etched by 0.02 mm, however, these peaks were still below the 10 dB requirement. These ﬁndings show that the antenna is very tolerant to moderate variations in all the parameters. IV. CONCLUSION In this paper, a dual-polarized wide-band patch antenna has been presented. The antenna has a VSWR < 2 : 1 bandwidth of greater than 50%, an isolation between ports of more than 39 dB and a gain of 7:4 dBi 6 0:4 dB over this frequency range. A cross-shaped reﬂector

**Resonant Frequency of Equilateral Triangular Microstrip Antenna With and Without Air Gap
**

Debatosh Guha and Jawad Y. Siddiqui

Abstract—A tunable equilateral triangular microstrip patch (ETMP) antenna with a variable air gap between the substrate and the ground plane has been studied theoretically and experimentally. Calculated resonant frequencies for different air gap heights have been veriﬁed with measurements of a coax-fed antenna. The tunability of the antenna as a function of the air gap height is studied theoretically showing over 200% tunable range of an ETMP with 50 mm side length printed on a substrate with = 10 5. The computed results for the antennas with zero air gap height are compared with a standard spectral domain moment method analysis supported by other previously reported experiments. Very close agreement is revealed in all comparisons. Index Terms—Microstrip antenna, triangular microstrip patch, tunable microstrip antenna. Manuscript received October 22, 2002; revised June 30, 2003. This work was supported by the Center of Advanced Study in Radio Physics and Electronics, University of Calcutta. The authors are with the Institute of Radio Physics and Electronics, University of Calcutta, Calcutta 700 009, India (e-mail: dgirpe@yahoo.co.in; siddiqui@cal.vsnl.net.in). Digital Object Identiﬁer 10.1109/TAP.2004.832504

0018-926X/04$20.00 © 2004 IEEE

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 52, NO. 8, AUGUST 2004

2175

I. INTRODUCTION Since the early phase of microstrip research the triangular geometry of microstrip patch is being investigated as planar circuit components [1]–[3] as well as microstrip antennas [4]–[17]. Many new conﬁgurations have also been studied in recent years to explore new characteristics [18]–[24]. Very recently, Gurel and Yazgan [25] have analyzed a tunable equilateral triangular microstrip patch (ETMP) antenna. They have introduced a variable air gap in between the ground plane and the substrate like some previous investigations with other geometries [26]–[28]. The resonant frequency fr of the ETMP becomes a critical design parameter since it is inherently a narrow bandwidth structure. Starting from [1], most of the formulations for fr are reported on the basis of the cavity model analyzes (CMA) [1], [4], [7]–[17], [25], though a few deal with other techniques like spectral domain technique [3], [19], geometrical theory [6] and spectral domain analysis with method of moment solution (SDA MoM) [12]. One recent advance in this ﬁeld [25] presents a new computation of the resonant frequency of an ETMP with and without air gap employing Wolff’s cavity model [29]. In [25], a better accuracy of their theory with respect to SDA MoM [12] and other available cavity model results is also outlined. But the formulation [25] itself suffers from some limitations. The effective side length of the ETMP derived in [25] is valid only for the substrates with "r < 10. These shortcomings are addressed in this paper and a simple formula of the resonant frequency of an ETMP with and without air gap is proposed on the basis of an improved cavity model recently proposed by one of the authors [30]. Like [30], the present model should be widely applicable, even to the MIC design on semiconductor materials with "r 12. Furthermore, the v