This action might not be possible to undo. Are you sure you want to continue?

29, 269–288, 2011

NEW TECHNIQUES FOR INCREASING ANTENNA BANDWIDTH WITH IMPEDANCE LOADING R. A. Formato Registered Patent Attorney & Consulting Engineer Cataldo & Fisher, LLC 400 Trade Center, Suite 5900 Woburn, MA 01801, USA Abstract—New methods are presented for increasing the bandwidth of wire antennas using impedance loading. This paper extends the seminal Wu-King theory of the internal impedance proﬁle that produces travelling-wave only current modes on a center-fed dipole antenna. It also presents a numerical optimization methodology based on Central Force Optimization, a new deterministic multidimensional search and optimization metaheuristic useful for problems in applied electromagnetics. A CFO-optimized loaded monopole antenna is described in detail and compared to the same structure loaded with a fractional Wu-King proﬁle. The CFO monopole generally performs better than other designs using either the full or fractional Wu-King proﬁles or the extended Wu-King proﬁles. The methods described in this paper should be useful in any wire antenna design that utilizes impedance loading to increase bandwidth. 1. INTRODUCTION This paper describes two methods for increasing wire antenna bandwidth using impedance loading: (1) an extension of the seminal analytic Wu-King loading proﬁle, and (2) a numerical approach based on Central Force Optimization. The notion of increasing antenna bandwidth by adding impedance loading has been around for long time, at least since the early 1950s. In 1953, for example, Willoughby published his Patent Speciﬁcation for “An Improved Wide Band Aerial” [1]. It discloses wire antenna elements that include a multiplicity of individual resistors, or wire segments with diﬀerent conductivity, whose resistance values increase exponentially with distance from the radio-frequency (RF) source.

Received 19 February 2011, Accepted 6 April 2011, Scheduled 8 April 2011 Corresponding author: Richard A. Formato (rf2@ieee.org).

270

Formato

In 1961, Altshuler described a “Traveling-Wave Linear Antenna” [2] whose input impedance was essentially constant over an octave in frequency. The ﬂat response resulted from inserting a 240 Ω resistor one-quarter wavelength (λ/4) from the end of a wire radiating element. Altshuler determined the value and location of the loading resistance by analogizing the antenna to an open-ended transmission line. His analysis is based on the relationship between a line’s characteristic impedance, Z0 , and a wire antenna’s expansion parameter Ψ. Physically, Ψ represents the ratio of the antenna’s surface vector potential to its axial current, and is approximately constant along the radiator as long as the current is not too small. For a dipole of half-length h and radius a, Ψ ≈ 2[ln( 2h ) − 1], and the corresponding loading resistance a that produces an approximately travelling wave current distribution is on the order of 60Ψ ohms. Altshuler’s initial work on resistive loading was extended by Wu and King (WK) in a seminal paper published in 1965 [3]. Instead of using a single lumped resistance, WK assumed a continuously variable resistance along a center-fed dipole (CFD) antenna. The objective of their analysis was to determine an impedance loading proﬁle that resulted in a purely travelling wave current mode. The WK proﬁle has been the basis of most if not all subsequent work on impedance-loaded wire elements. The performance achieved by continuously loaded antennas over the years has been impressive. For example, Kanda [4] built a small receive-only loaded-CFD ﬁeld probe using the full WK proﬁle that exhibited essentially ﬂat response from HF to beyond 1 GHz. Both resistive and capacitive loading were employed by depositing a segmented, conductive thin-ﬁlm of varying thickness on a glass rod substrate. Reactive loading was achieved by burning away with a laser narrow rings between conductive segments. Unfortunately, this ﬁeld sensor was so heavily loaded that its radiation eﬃciency was far too low for use as a transmit antenna (transfer function typically below −22 dB). The full WK proﬁle tends to load the antenna very heavily with an attendant substantial reduction in radiation eﬃciency. This issue was addressed by Rama Rao and Debroux (RD), who designed, fabricated and measured a continuously loaded high frequency (HF) monopole [5, 6]. Because the full WK proﬁle generally results in poor eﬃciency, the RD monopole used a fractional loading proﬁle equal to 0.3 times the Wu-King proﬁle (“30% proﬁle”), along with a ﬁxed, lumped-element matching network. Besides fractional proﬁles, other loading proﬁles have been proposed that combine resistance and inductance to improve bandwidth and eﬃciency [7]. Recently, Lestari et al. revisited the WK analysis in order to develop an

This paper is organized as follows: Section 2 presents an overview of the Wu-King theory. of which the WK proﬁle is a special case. 29. improves eﬃciency. Vol. There is no reﬂected wave propagating toward the source to form a standing wave pattern. while Section 3 describes the extension of that theory. 2011 271 optimal linear loading proﬁle for a ground penetrating radar bow-tie antenna [8]. in turn. Section 5 applies Central Force Optimization to the loaded monopole problem and compares its results to the performance of the analytic loading proﬁles. Because the ﬁelds radiated by an antenna are proportional to its Idl product (current moment).Progress In Electromagnetics Research B. This paper describes a modiﬁcation to the WK proﬁle that results in higher radiation eﬃciency by increasing the average antenna current while maintaining the traveling-wave current mode necessary for increased bandwidth. where Ψ = ΨR + jΨI is the complex expansion |z| 1− h . increasing the average current increases the radiated ﬁelds. The propagation factor represents a current wave progressing outward along each dipole arm. The required proﬁle is given by Z i (z) = 60(Ψ/h) . The extended theory is applied to the RD HF monopole as a design example in Section 4. The current distribution produced by the WK loading proﬁle decays linearly along the antenna. The improved proﬁle produces a traveling-wave current mode with a power law decay. where Z i (z) is the (complex) internal impedance per unit length (ohms/meter) consisting of lineal resistance Ri (z) and reactance X i (z) √ with j = −1. and consequently no resonance eﬀect. This current distribution exists only when the CFD element has a speciﬁc “1/z” internal impedance proﬁle. WU-KING THEORY The WK model assumes that the CFD has an internal impedance proﬁle along the wire element given by Z i (z) = Ri (z) + jX i (z). h which consists of the product of a linearly decreasing (“straight line”) amplitude factor (1 − |z| ) and a traveling wave propagation factor h in the complex exponential term exp (−j k0 |z|). The WK current distribution is Iz (z) ≈ (1 − |z| ) exp (−j k0 |z| ). WK develops the diﬀerential equation satisﬁed by the CFD’s current Iz (z) and then determines by inspection that a traveling-wave current mode exists for one particular impedance proﬁle Z i . which. z is the distance from the RF source. k0 = 2π/λ 0 is the wavenumber. Additional analyses of dipole structures generally appear in [9–12]. 2. Section 6 is the conclusion.

which then yields the condition that must be satisﬁed by the element’s internal impedance in order to generate traveling-wave-only modes. When ν = 1 the Wu and King case is recovered. that is. λ0 = 4h (see [3] for details). Of course. It is . x) = 0 sin(W ) du. Ψ is the ratio of the antenna element’s surface vector potential to axial current. the design frequency f0 (Hz) at which Ψ is evaluated and the corresponding wavelength λ0 (meters) are related by f0 λ0 = c ≈ 2. As discussed below. This approach is fundamentally diﬀerent than in [3] because the loading proﬁle for a particular traveling-wave current mode now is an unknown which is determined by solving the appropriate equations. 13] given by √ x x C(b. As in [3]. Figure 1 shows typical current amplitude distributions parametric in the proﬁle exponent ν. but when ν = 1 the more general case is obtained.998 × 108 m/s where c is the free-space velocity of light. respectively. the current amplitude now decays with a power law variation (h − |z| ) ν with exponent ν (proﬁle exponent). exp (−jk0 |z| ) is the travelling wave propagation factor. Note that the design frequency f0 for evaluating Ψ is not the RF source frequency. k0 h C and S are the generalized sine and cosine integrals [3. and is approximately constant along its length. W W Because Ψ is frequency dependent. The 1/z proﬁle is the basis for the resistive loading used in [4– 7].272 Formato parameter [2. The expansion parameter is deﬁned as [3] Ψ = 2 sinh−1 + j h a − C(2k0 a. S(b. 3. x) = 0 1−cos(W ) du. 3] with real and imaginary parts subscripted R and I. IMPROVED ANALYTIC LOADING PROFILE The WK 1/z proﬁle is a special case of a more general proﬁle that results in higher radiation eﬃciency. The generalized CFD current distribution is assumed to be of the form Iz (z) = C(h−|z| ) ν exp (−jk0 |z| ) where C is a complex constant determined by the current at the feed point. Instead of the WK linear decay. I0 being the CFD’s feed point current. The ﬁrst step in deriving the generalized proﬁle is to assume a power law traveling-wave current distribution. however. 2k0 h) [1 − exp (−j2k0 h) ] . it usually is evaluated at the antenna’s fundamental half-wave resonance. 2k0 h) − jS(2k0 a. this choice is not necessarily the best. W = u2 + b2 . The next step is to substitute this assumed current distribution into the current equation developed by WK.

Equation (11) in [3]. 6]. (h−|z|) which generalizes Equation (12) in [3] and recovers it when ν = 1. The extended impedance proﬁle that produces travelling-wave-only modes is determined by f (z). for X i ≥ 0 and X i < 0. The improved internal impedance proﬁle is determined as follows. But many practical designs employ only resistive loading because excellent results often are achieved even without the reactive component (see [5. 29. CFD current distributions parametric in proﬁle exponent ν. The resulting equation is satisﬁed by the auxiliary function f (z) introduced in Equation (9) in [3]. In the monopole antenna designs discussed below only resistive loading is considered. 2 The derivatives dIz and d I2z are computed and substituted into the dz dz equation satisﬁed by Iz (z).Progress In Electromagnetics Research B. respectively. 15]. Ri (z) = 60ν(h − |z|)ν−2 ΨR − . While f0 usually is the antenna’s half-wave resonant frequency. The solution for f (z) is f (z) = 2ν (h − |z|)ν−2 {1 − j 2k0ν−1 }. this is not necessarily the best choice as discussed in [14. for example). The improved loading proﬁle in Equation (1) above contains both resistance and reactance. apparent that values of ν less than unity create signiﬁcantly higher average antenna currents. Vol. 2011 273 Figure 1. Radiating elements with these current distributions are more eﬃcient than those using the 1/z loading proﬁle corresponding to ν = 1. The circular frequency is ω0 = 2π f0 where f0 is the design frequency (Hz) at which Ψ is computed. The loading proﬁle resistance and reactance per unit length (units of ohms/meter) are computed from f (z) to be (1 − ν)ΨI (1a) 2k0 (h − |z|) (1 − ν)ΨR (1b) X i (z) = 60ν(h − |z|)ν−2 ΨI + 2k0 (h − |z|) The corresponding lineal inductance (henrys/meter) or capacitance (farads/meter) are Li = X i /ω0 and C i = 1/(ω0 X i ).

often precipitously.” [5. instead of the RD loading proﬁles. The objectives of the RD design were (1) as ﬂat as possible an impedance bandwidth and (2) a stable radiation pattern at low take-oﬀ angles to support long range HF links from 5 to 30 MHz. This paper does not. to quote. As the impedance response is progressively ﬂattened in a loaded antenna. The results reported here also are computed using NEC. “The input impedance values predicted by both theories are in good qualitative agreement with the measured results and conﬁrm the non-resonant behavior of the antenna. Because VSWR ﬂuctuated substantially. continuously resistively loaded base-fed monopole antenna using full and fractional WK proﬁles. The antenna then was built using diﬀerent loading proﬁles. NEC-modeled monopole performance thus agrees well with measured data. 6]. a tradeoﬀ that appears to be typical. and vice versa. They designed. based on how well the RD measured and NEC-computed data agree. thereby validating NEC as an accurate predictive tool for loaded wire element antenna design. LOADED HF MONOPOLE USING ANALYTIC PROFILE Rama Rao and Debroux’s work provides a useful analytical and experimental benchmark for testing the improved loading proﬁle. and extensive measurements were made of its performance [5. the radiation eﬃciency drops. The measured RD data validated both the WK model and NEC’s results. The loading proﬁle was computed using WK theory and the antenna performance modeled with the NEC Method of Moments code version 3 (see below). achieving the VSWR objective required the use of a 6-element matching network and 4 : 1 transformer. a base- .274 Formato 4. Equation (1) above was used to compute a discrete resistanceonly loading proﬁle for the RD monopole geometry. 1226]. fabricated and measured a 35-foot tall. but. that is. The improved loading proﬁle described here mitigates this eﬀect. new proﬁles are applied using the extended WK theory and a numerical optimization algorithm. however. This paper therefore compares its analytical results directly to RD’s analytical results through the NEC modeling engine. The RD design resulted in a voltage standing wave ratio (VSWR) ≤ 2 : 1 relative to 50 Ω and radiation eﬃciency in the range 15%– 36%. p. A continuously varying resistance along the antenna was created by coating a dielectric substrate with variable thickness chromium ﬁlm. 2-inch diameter. report new measurement results using the modiﬁed loading proﬁles because. there is no reason even to suspect that NEC does not accurately calculate the monopole’s performance.

. although in this case the eﬀect is minor (a maximum bandwidth metallic monopole element has h ∼ 10a [16]). and 0.Progress In Electromagnetics Research B. The loading resistance k c value was calculated segment-by-segment as RL = Ri (zk ) · ∆. a ﬁlamentary antenna wire could be used instead of the 2-inch diameter substrate. and the corresponding loading proﬁles are plotted in Figure 3. A typical NEC input ﬁle appears in Figure 2. however.1D). which was used because it is freely available online [17]. Because singleelement lumped resistances are used. . ν = 0.668 meters tall.028 MHz was used as the design frequency. 3–8]. ν = 1 . Corresponding values for the least heavily loaded proﬁle. 14 where ∆ = h/14 is the constant segment length and zk = ∆ (2k − 1) 2 is the z coordinate at the center of the kth segment. 0.8. . Figures 4 and 5 plot the monopole’s input resistance and reactance. k = c 1. As will be seen below. 2011 275 fed element h = 10.27 Ω and 461.4. As for all plots in this paper. are 0. But for all values of ν the data reﬂect the typical characteristic of progressively increasing resistance with distance from the RF source as seen in [1. The antenna was modeled with NEC-2D (Numerical Electromagnetics Code Version 2 Double Precision). Runs were made using the proﬁle exponent values in Figure 1. and the continuous curves connecting data points have been interpolated using a natural cubic spline. The generalized sine/cosine integrals were numerically evaluated using 32-point Gauss quadrature (see generally [13]). 0. As expected.8.2 Ω. calculated data points are marked by symbols. 19] were followed.4 Ω Interestingly. The fundamental half-wave resonance of 7. The heaviest loading is the full WK proﬁle corresponding to ν = 1. Both versions of NEC yield essentially the same results.0254 meters radius on a PEC (perfectly electrically conducting) ground plane. a = 0. The most heavily loaded proﬁles. while the last segment at the top of the monopole is loaded with 1220 Ω. and as a check the antenna was modeled with the most recent version of NEC as well (NEC-4. Vol. The resistor in the ﬁrst segment has a value of 45. But the same geometry as the RD monopole was modeled because the antenna’s length-to-diameter ratio does inﬂuence its bandwidth. the loading resistance increases with increasing proﬁle exponent at all segments except the last where the three highest resistances in increasing order correspond to ν values of 1. 29. excursions in the antenna’s input impedance increase with decreasing proﬁle exponent. NEC’s modeling guidelines and the model validation procedure described in [18. . Fourteen equal-length segments were used with each segment loaded at its midpoint using NEC “LD” cards. .05. monotonically increasing proﬁles are not necessarily the best.

the heavily loaded WK proﬁle with ν = 1 results in an eﬃciency of less than 4%. Radiation eﬃciency appears in Figure 6. for example. it exhibits the greatest variability. As with . Figure 4.4. Figure 3. while the very light loading proﬁle ν = 0. The eﬀect of the proﬁle exponent is dramatic.05 increases it to nearly 70%. Extended loading proﬁle NEC input ﬁle for ν = 0.276 Formato Figure 2. For comparison. show the ﬂattest responses. ν = 0) is plotted as well. of course. especially for the input resistance. Loaded monopole input resistance. and. Discrete extended theory loading proﬁles. the input impedance of the antenna with no loading (“metallic” monopole. At 5 MHz.

result in higher and more variable VSWR’s. Compared to the full WK proﬁle. matching VSWR levels up to 10 : 1 . somewhat surprisingly. Between 5 and 8 MHz the VSWR decreases quickly from approximately 4. in that order. Vol. the ν = 0.4 provides what likely is the best tradeoﬀ between VSWR and radiation eﬃciency.15. Loaded monopole VSWR//300 Ω. the input impedance.8 curve is consistently better than the ν = 1 curve. Loaded monopole input reactance. Figure 7. But. lower values of proﬁle exponent result in greater variability with frequency.6 : 1 to 2 : 1. 2011 277 Figure 5. Loaded monopole radiation eﬃciency. the ν = 0.Progress In Electromagnetics Research B. As expected. As a general proposition. 29. the ﬂattest response is achieved with the heaviest loading. Loading at this level results in radiation eﬃciencies between about 18% and 41% with a VSWR below 2 : 1 at all frequencies above approximately 8 MHz without any matching network. Figure 7 plots the loaded monopole’s VSWR versus frequency assuming a 300 Ω feed system characteristic impedance (the value used in [5]). eﬃciency increases with decreasing exponent values. At any given frequency. Proﬁles with ν = 0. 0. The proﬁle with ν = 0. Figure 8. Figure 6. 0.05. Loaded monopole maximum power gain.8 proﬁle provides both better radiation eﬃciency and VSWR performance.4.

the ν = 1 proﬁle’s gain just above −10 dBi is more than 13 dB below the gain with ν = 0. greater than 7 dB or so. As a reference.4 over angles of interest are for the most part comparable. In all cases the pattern is smoothest at 5 MHz. Gain was computed at 10◦ increments in the polar angle 0◦ ≤ θ ≤ 90◦ in NEC’s standard right-handed polar coordinate system. The radiation patterns. and at a given frequency with lighter loading. But because some matching must be done for each proﬁle. low-loss network. and the penalty paid for heavy loading is substantial. the parameter-free implementation described in [20] was used to determine an optimized loading proﬁle for the RD HF monopole discussed in §4 .278 Formato in a fairly narrow band is readily accomplished with a simple. between about 39% and 62%. which were computed at 1◦ increments in θ. Maximum power gain (product of directive gain and radiation eﬃciency) and radiation patterns for the loaded monopole appear in Figures 8 and 9. The gap generally narrows with increasing frequency. At 5 MHz. with a maximum VSWR below 10 : 1 that falls quickly to approximately 2. as expected. respectively. The heavier loading proﬁles. as expected. CFO is a deterministic metaheuristic for global search and optimization that has been applied to problems in applied electromagnetics and tested against recognized benchmark suites [20–34]. so that moderate take-oﬀ angles for long-distance links generally are in the range 60◦ ≤ θ ≤ 80◦ . for example. exhibit substantially lower gains and consequently are less desirable. but it remains signiﬁcant across the HF band.15 proﬁle may in fact be the best because it provides much higher eﬃciencies. LOADED HF MONOPOLE USING NUMERICAL PROFILE This section describes a numerical technique for computing an optimized loading proﬁle using Central Force Optimization (CFO). The gains with 0. show generally smooth variations without a great deal of lobing. While more eﬃcient versions of the algorithm have been implemented recently on a GPU using multiple topologies (contact author for details). patterns for the unloaded metallic monopole are included. and. by contrast. The maximum gain value over these angles is plotted. the ν = 0. 5. Gain increases with lighter loading. Note that θ = 0◦ is in the direction of the zenith and θ = 90◦ the horizon. they show the greatest lobing.05 ≤ ν ≤ 0. and there is no clear basis for selecting one loading proﬁle over another. More scalloping develops with increasing frequency at all loading levels.05 loading.5 : 1 at 10 MHz and thereafter remaining below that level.

chosen because analytical results suggest that maximum required load resistances are on the order of several hundred ohms. R2 . previous experience. the decision space may be expanded or contracted based on many factors. k = 1. .Progress In Electromagnetics Research B. Of course. . . (see also [23]). Vol. . for example. 14. . R14 ) is the vector of load resistor values where RL = (RL L L segment-by-segment along the monopole. Loaded monopole radiation patterns. . 29.5 1 . . data suggesting regions of “good” solutions (for . . The objective function being maximized is f (RL )= min(ε)+min(Gmax ) ∆V SW R2. 2011 279 Figure 9. The feasible space (allowable k solutions) was 0 ≤ RL ≤ 1000 Ω.

For example. CFO allows the user to focus on the more important issues of the objective function’s form and its parameter values. The decision space size inﬂuences how well CFO or other optimizers work. an analytical proﬁle from §4). A still more general form might contain coeﬃcients and exponents that are frequency-dependent. in order to improve performance in certain frequency ranges. while min(Gmax ) is the corresponding minimum value of maximum total power gain.” or practical constraints related to fabricating the antenna. whereas CFO can be implemented parameterfree so that the only user input is the objective function to be maximized. The decision space’s topology is determined by the objective function’s form and parameter values. which provides greater ∆V SW Rc3 ﬂexibility by including three user-speciﬁed constants. The exponent of 2. many optimization algorithms require the user to provide several parameters speciﬁc to the algorithm. f (RL )’s functional form was chosen to return greater values for loading proﬁles that result in greater gain and eﬃciency and reduced variability in VSWR. a “best guess. min(ε) is the minimum radiation eﬃciency across the 5–30 MHz band (computed every 1 MHz). This question is more easily addressed with a deterministic optimizer such as CFO than it is with . which consequently also may be a factor in its speciﬁcation. One of CFO’s major advantages over other optimizers is its determinism. the eﬀects of changing the objective function’s form or parameter values can be investigated directly. it is entirely possible. which cannot be done with a stochastic optimizer whether or not it requires userinput parameters. that some other decision space may yield a better antenna.280 Formato example. In addition. Every CFO run with the same setup parameters yields the same results. For “real world” problems such as the loaded monopole design. c3 . ∆V SW R is the diﬀerence between maximum and minimum VSWR relative to a feed system characteristic impedance Z0 = 300Ω. Even when the optimizer accurately locates the global maxima for a particular topology. ci (fM Hz ). a more general form than the one above is f (RL ) = c1 ·min(ε)+c2 ·min(Gmax ) . in contrast to inherently stochastic optimizers that yield diﬀerent outcomes on successive runs. Because only one CFO run is required. even likely. other functional forms could be used that might provide even better results. and changing any of these changes the topology being searched. In general there is no way to know which topology actually produces the “best” antenna design. Of course.5 for this term was chosen empirically. Other functional forms could be used that reﬂect a completely diﬀerent balance between the monopole’s various performance measures. c2 . c1 .

Progress In Electromagnetics Research B. Evolution of the optimizer’s best ﬁtness (objective function value). . 29. Nt = 120. 2011 281 a stochastic optimizer whose results never are the same.1. The CFO optimization run reported here employs the parameterfree implementation described in [20] with directional information included in errant probe repositioning described in [31] and the init following hard-wired parameter values: Frep = 0. (Np /Nd )max = 4.5. Nd = 14. γstop = 1. Vol.1. Pseudocode appears in Figure 10. and CFO’s best probe number are [20] [20] [20] [31] (see [27. Parameter-free CFO Pseudocode (Ω is the decision space).29]): [31] [20] [20] Figure 10. γstart = 0. average probe distance (Davg ). ∆γ = 0. ∆Frep = 0. Readers wishing to implement their own versions of the program are advised to consult [20] for additional information that is beyond the scope of this paper.

The best probe number plot shows how the probe returning the best ﬁtness changes throughout a run. Figure 14 tabulates where each loading resistor is placed along the monopole’s height and its value. 25]).429 4.001 8.667 3. and may be characteristic of antenna decision spaces (see. respectively.239 8.2825 7.5575 24.7045 29.477 7.3102 27.4492 11.4537 9. for example.381 1.763 9.715 6.3978 7. Figure 12. CFO best probe Figure 14. Height (meters) 0.1540 7.953 5. Davg is the average distance step-by-step between CFO’s probe with the best ﬁtness and all other probes.525 10. [23. while Figure 15 plots the resistance as a function of segment number. loading proﬁle. CFO-optimized .905 2.5870 26.143 1. Davg decreases as probes cluster around the global maximum.8245 16. These ﬁgures are interesting because the numerically optimized proﬁle is quite diﬀerent than the proﬁles computed using the original WK theory or the extension developed in this paper.287 Resistance (Ω) 82.191 4.4720 Figure 13. number. The ﬁtness exhibits ﬁve main plateaus with three smaller (shorter duration) ones.282 Formato plotted in Figures 11–13. CFO Davg .7010 22. approaching zero as all probes coalesce onto the maximum. CFO best ﬁtness. The CFO-optimized loading proﬁle is shown in Figures 14 and 15. This type of behavior has been seen in other antenna related problems. These plots and their signiﬁcance are discussed at length in [20]. Figure 11.8015 20.3115 9.

Figure 16. 29. and the reactance exhibits a similar variation between approximately −220 Ω and +120 Ω. and shows clearly that the WK theory and its extension are not a complete solution to the problem of creating travelling wave-only current modes using impedance loading. These variations are moderate and should be expected to . CFO-optimized monopole loading proﬁle. In those proﬁles the loading resistance increases monotonically with increasing distance from the excitation point. 2011 283 Figure 15. then jumps to a higher value and thereafter continues to decrease monotonically. in marked contrast. and Figures 17–21 plot the NEC-2D results. Vol. starts with a high resistance that decreases approximately monotonically over about half the monopole height.Progress In Electromagnetics Research B. CFO-optimized NEC input ﬁle. This behavior is quite diﬀerent from the analytic proﬁles. Figure 16 shows the CFO-optimized monopole’s NEC input ﬁle. The CFO proﬁle. The input resistance varies gradually with frequency from about 175 Ω to 575 Ω.

CFO loaded monopole input impedance. The worst eﬃciency . Figure 18 plots the monopole’s radiation eﬃciency which ranges from a low of about 9% to a maximum of 45%. Figure 19. CFO loaded monopole radiation eﬃciency.5.5 MHz at alternate resonances and anti-resonances in the monopole structure. CFO Loaded Monopole VSWR//300 Ω. Figure 21. CFO loaded monopole radiation pattern. Figure 20. 15. result in a similarly moderate VSWR variation. CFO loaded monopole maximum gain. 21 and 27. Four resonant points (Xin = 0) appear below approximately 7. Figure 18.284 Formato Figure 17.

which is plotted in Figure 19. 2011 285 occurs at the lowest frequency. The radiation eﬃciency of the CFO-optimized monopole is comparable to that using the analytically computed proﬁles discussed in §4. CONCLUSION This paper has presented two new approaches to broadbanding wire antennas using impedance loading: (1) an analytic method based on extending the seminal Wu-King theory for the internal impedance proﬁle that produces purely travelling wave current modes.org). 1961. 2. 5 MHz. 29. Altshuler. Jul.” IEEE Transactions on Antennas . does not require a continuously varying proﬁle.. These new techniques. Maximum VSWR is about 2. O.” IRE Transactions on Antennas and Propagation. E. and (2) a numerical method based on the deterministic metaheuristic Central Force Optimization.” Patent Speciﬁcation. No. The ﬁnal measures of the CFO-optimized monopole’s eﬀectiveness are its maximum gain and radiation patterns. Wu. especially CFO. Willoughby. W. 3. and avoids the insertion losses inherent in any matching network. The extended analytic proﬁles provide results generally better than the full or fractional WK proﬁles. “The traveling-wave linear antenna. Perhaps the most important single measure of the loaded monopole’s eﬀectiveness is its VSWR. 324. VSWR//300 Ω is below 2 : 1 at all frequencies above approximately 5. E. should be useful in any broadband antenna design that incorporates impedance loading. the CFO optimized monopole meets the standing wave ratio design objective essentially across the entire 5–30 MHz HF band without any matching network. but it increases quickly to about 15% near 7 MHz. Vol. Commonwealth of Australia. Thus. T.Progress In Electromagnetics Research B. 13.. which are plotted in Figures 20 and 21. “An improved wide band aerial.2 : 1 at 5 MHz. E. Thereafter it exhibits an oscillatory behavior with a minimum slightly less than 20% near 19 MHz. 6. 1953. These patterns are similar to those produced by the analytic proﬁles. uses lower value resistance. “The cylindrical antenna with nonreﬂecting resistive loading. REFERENCES 1. P. 162009. T. CFO source code and any papers not freely accessible online are available upon request to the author (rf2@ieee. Aug. which is quite signiﬁcant because none of the analytic loading proﬁles performs as well. King. The CFO-optimized antenna is simpler. as expected. and the CFO-optimized loading proﬁle is better yet.3 MHz. and R.

438. Lestari. Ant. May 1965. 6. Ninth Printing.. 257– 259. G.” IEEE Military Communications Conference. M. B. 2004. 1970. 7. 11. 1965). Stegun.” IEEE Transactions on Antennas and Propagation. New York. and P. A. 114. Soc. 51–66. 3. 2089. Ramabi. L. A. R. London. US Department of Commerce. Graphs. W. and R.. (see also Corrections. 2184–2192. 231. Fikioris. E. 9. Little.” Progress In Electromagnetics Research. Valagiannopoulos. A. and Propagation. 1990. Valagiannopoulos. 14. 55. Kurniawan. 30–Oct. “A modiﬁed bow-tie antenna for improved pulse radiation. Mack. USA. Conference Proceedings. A. 2002. A.. G. J. B. Rama Rao. “Wideband HF monopole antennas with tapered resistivity loading. Sep. 1992. Rama Rao. 369–373. Handbook of Mathematical Functions with Formulas. Ligthart. Canada. 14–17. O. National Bureau of Standards Applied Mathematics Series 55 (AMS 55). 2005. Monterey. A. Jul. and R. Ukraine. Jul.” 10th International Conference on Mathematical Methods in Electromagnetic Theory.. Vol. C.” IEEE Transactions on Antennas and Propagation. May 1983. Formato. and Mathematical Tables. 13. and L. P. and G. Fikioris. 10. Cambridge University Press. and I. “Optimized tapered resistivity proﬁles for wideband HF monopole antenna.. Prop. “Electromagnetic scattering of the ﬁeld of a metamaterial slab antenna by an arbitrarily positioned cluster of metallic cylinders. “Loading proﬁles for wideband antennas. Soc. Yarovoy. 12.. S. Mittra. G. Symposium. Vol. “On choosing kernel and feed when analyzing dipole antennas via integral equations. & Prop.286 Formato 4. Ontario. King. Sep. B. IEEE Trans.” . C. 285–306. Debroux. Valagiannopoulos. CA. and C.. “Combined tapered resistive and inductive loading to increase the bandwidth of small antennas. B. R. 2010. A.” Progress In Electromagnetics Research. 2011. “Time domain sensors for radiated impulsive measurements. 8. Nov. Abramowitz. Symposium. Dniepropetrovsk. Bharata. A.” IEEE Ant. A. 5. 1991. Kanda. Cylindrical Antennas and Arrays. Nov. M. Fikioris. Suksmono.. P. “Input admittances arising from explicit solutions to integral equations for inﬁnitelength dipole antennas. 998. & Prop.” Proceedings of the IEEE Ant.

2010.” Progress In Electromagnetics Research B. Mar. A. “Wideband antennas. https://ipo. (see also Corrections.nabfastroad. Lawrence Livermore National Laboratory. “Central force optimization and NEOs — First . (ii) Unoﬃcial Numerical Electromagnetic Code (NEC) Archives. California. http://www. doi:10.org/pierb/pier. 5. E. 22. Formato. 2010. J. J. 24. N. 17. 1997. 405–425. R. 18. UCID-19934. Livermore.pdf. “Highperformance indoor VHF-UHF antennas: Technology update report. online at http://www. Livermore.ict. Oct. Heidelberg. Merulla.nl/∼arivoors/. No.” Parts I. Inc. Antennas & Propagation. National Association of Radio and Telecommunications Engineers.” UCRL-MA-109338. No.llnl.” Proceeding ICIC’10 Proceedings of the 6th International Conference on Advanced Intelligent Computing Theories and Applications: Intelligent Computing . W.net/swindex. Communications Quarterly Magazine.1049/iet-map. R.2009. Voors.. 2010. J. Berlin. 11.. Jan.jpier. Part I: User’s manual and Part II: Program description — Theory. 19.silist. USA. 1993– Jan. Formato. 21. 23. A. R.. Technology Advocacy Program.org/NABHighperformanceIndoorTVant ennaRpt. 202. Formato. Jan.0147. Vol. Vol. 2011 287 15. and R.. 309–318. 4. 16. doi: 10. “Numerical electromagnetics code — NEC-4. Poggio.. A. Springer-Verlag. A. 1992. Lawrence Livermore National Laboratory. Burke. 29. http://www. G. “Parameter-free deterministic global search with simpliﬁed central force optimization.php?paper=09112309. FASTROAD (Flexible Advanced Services for Television and Radio On All Devices). USA. A. May 15. Formato. Formato. A. 1981.. 5. California. R.” Electronics World Magazine. 583–592. Fall 1997).html. “Maximizing the bandwidth of monopole antennas.gov/technology/software/softwaretitles/nec.. 20. Qubati. I. G. 1994.Progress In Electromagnetics Research B. Vol. M. Formato. “Antenna benchmark performance and array synthesis using central force optimization. Burke. method of moments. A. Summer 1997. II and III. G. Formato. “Numerical electromagnetics code (NEC) — Method of moments. Dib. M. 2010. available online at http://home. 4. Cross. and A.” NARTE News. R. 27. (i) 4nec2 Antenna Modeling Freeware by A. and R.” National Association of Broadcasters (NAB).php.” IET (UK) Microwaves. “Improved CFO algorithm for antenna optimization... 93.2528/PIERB09112309.

Formato. 217–238.. Xie.” Nature Inspired Cooperative Strategies for Optimization (NICSO 2007).. http://arXiv. Vol. Bio-inspired Computation. http://arXiv. R. and R. 28.org.0221. G. 27. Vol. 32. Heidelberg. http://arXiv. Krasnogor... A. 34. A. cousins?. N. Formato. No. R.1504/IJBIC. A. . L.org. “Central force optimisation: A new gradient-like metaheuristic for multidimensional search and optimisation. A.” Progress In Electromagnetics Research. “Are near earth objects the key to optimization theory?” 2009. Nicosia. Formato. Formato.02. arXiv:1001.” Journal of Multiple-valued Logic and Soft Computing.2011.amc.2009. “Central force optimization: A new computational framework for multidimensional search and optimization.1007/s12597-009-0003-4. Vol. J.” Journal of the Operations Research Society of India.1016/j.org/10.5629. doi: 10. Vol. Formato.024721. R.5490.. Vol.” Applied Mathematics and Computation. 2011. Springer-Verlag. 30. online at http://dx. 2009. doi: 10. No. Pelta. Formato. 2009.” Mar.. 25–51. 425–491. 2008. and D. 29.org. 4. R..” Int. M. 33. “Central force optimization: A new deterministic gradient-like optimization metaheuristic. “Central force optimization applied to the PBM suite of antenna benchmarks. “Pseudorandomness in central force optimization. “Issues in antenna optimization — A monopole case study.” 2010. Eds. Formato. Formato.doi.0317. arXiv:1005. A. A. 129. Zeng. R.” 2010. 31. 46. 547–565. 2007.288 Formato 25. A. 77. http://arXiv. 1.. A. 1. A. Formato. Pavone. 26.1394. 2010. “Central force optimization: A new metaheuristic with applications in applied electromagnetics. 2011. “Convergence analysis and performance of the extended artiﬁcial physics optimization algorithm. arXiv:1003. arXiv:0912.062. R... R.” 2010. http://arXiv. J. R. Formato. “On the utility of directional information for repositioning errant probes in central force optimization. 16.. R. A.org.org/abs/1103.

- Microwave Circuit Design
- Reasons of Refusal 010513.pdf
- Ch1
- HDR Scholarship Expression of Interest 2016
- Reasons of Refusal 010513
- swiss phd
- Danish grades
- How to Make Polio Certificate
- How to Make Police Character Certificate
- Goipg 2016 Guide for Applicants
- APYA-S-15-02657
- Agency Number+Universites Name and Embassy Name
- 26- 01229898 Onthe Relationship Between Fractal Dimension and The
- 25- 04531791 CPW-fed Koch Fractal Slot
- 24- 06497485 UWB CPW-fed
- 23- 05766703 Novel Modified Pythagorean Tree Fractal Monopole Antennas
- 22- 1-s2.0-S1434841111001464-Main on the Design of Inscribed Triangle Circular Fractal Antenna for
- 21- 09205071%2E2013%2E852486 Multiple Fractal-shaped Slots-based UWB Antenna With
- 20- 01421156 Frequency Notched Ultra-wideband Microstrip Slot
- 19- 01189650_5 an Overview of Fractal Antenna Engineering Research
- 18- 05.08111907 a Rhombic Patch Monopole
- 16- 01.10040507 UWB Circular Slot Antenna Provided With an Inverted-l Notch Filter for the 5
- 15- 06166848 a Novel UWB Antenna With Dual Notched Bands for WiMAX and WLAN Applications
- 14- Mop28087 a Compact CPW‐Fed Circular Slot Antenna With Reconfigurable Dual Band‐Notch
- 13- 17.14010805 Band-notched UWB Monopole Antenna

- 11 Optimization of Resonance Frequency of Circular Patch Antenna at 5 Ghz Using Particle Swarm Optimization Copyright Ijaetby IJAET Journal
- Beatriz Maria Vidal and Zbynek Raida- Synthesizing Sierpinski Antenna by Genetic Algorithm and Swarm Optimizationby Irokk
- 6-Krishan Sherdia -Desing of Microstrip Antenna Withby Phong Trần
- IJCTEE_1111_22by Cha Ran

- Equivalent Ckt
- Genetic_SA_Antennas.pdf
- Aces Paper March2005
- 16.09071510
- 11 Optimization of Resonance Frequency of Circular Patch Antenna at 5 Ghz Using Particle Swarm Optimization Copyright Ijaet
- Beatriz Maria Vidal and Zbynek Raida- Synthesizing Sierpinski Antenna by Genetic Algorithm and Swarm Optimization
- 6-Krishan Sherdia -Desing of Microstrip Antenna With
- IJCTEE_1111_22
- AN-14
- Printed UWB Circular and Modified Circular Disc Monopole Antennas
- Microwave Planar Antenna Design
- 2011_59_02
- 389418.pdf
- Chpt18
- A Reconfigurable Stacked Microstrip Patch Antenna for Satellite A
- 05062526
- Compact Quasi Self-complementary Elliptical Disc
- Chap 04
- Low Cost Prime Focus Reflector Feed
- 19.11100703
- 06.07030904.Mahmoud.EIBZ
- 08-09_Greño,Cano_Ana Isabel
- 2P7_0197
- 4P3_1122
- 01629614
- Antenna Impedance Measurements
- 39005
- Em_ Talk - HFSS Tutorial 1_ Microstrip Patch Antenna
- paper on antennas
- jghv
- 14.11021904

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd