Supported

By

IEEE UP SECTION



A
CONFERENCE
On

“SIGNAL PROCESSING and REAL TIME OPERATING SYSTEM (SPRTOS)”
MARCH 26-27, 2011



Organized by





Department of Electronics Engineering
Harcourt Butler Technological Institute, Kanpur, (U.P) 208 002

Editor
Dr. Krishna Raj






About the Conference:

The program will intend to promote high standards in Technical Education by way of extending
opportunities to the academicians and working professionals by providing a forum for sharing
their knowledge, experiences, innovations and inventions in various fields of Signal Processing
and Real Time Operating System
The conference will cover the following fields:
 Signal collection/Storage/representation
 Data communication
 Data (signal) processing
 Monitoring of Data (signal)
 Multimedia signal processing and coding.
 Audio-Video Processing Techniques, Compression & Coding Standards
 Signal Processing in Network Analysis and VLSI architectures
 Speaker identification
 Computer Network Analysis and Design
 Signal detection and Spectrum estimation
 Application of DSP in Remote Sensors
 Digital and Multirate Signal Processing
 Digital Signal Processors
 Multidimensional and Multimodal signal processing.
 Sensor array and multi-channel signal processing
 Integrated Services Digital Networks (ISDN)
 Architecture, implementation, and applications of digital filters
 Signal Processing in Encrypted Domain
 Real-Time Signal/ Image/Video Processing
 Embedded systems
 Signal & Image Processing in Neural Networks and Artificial Intelligence
 Application of VxWorks


About H.B.T.I. Kanpur
The Harcourt Butler Technological Institute, Kanpur, one of the premier institutes of the country
in the heart of “ THE MANCHESTER CITY OF NORTH INDIA KANPUR”, founded on 25
th

November 1921, has retained its name and reputation ever since. It has fulfilled the
responsibilities of lead institute in Technical Education Quality Improvement Program (TEQIP)
scheme of Word Bank project. It has significantly contributed towards manpower development,
research and training. The Institute runs 13 under graduate and 6 post graduate level courses and
doctoral program in all emerging disciplines of science and technology.

About the Department
Department of Electronics Engineering offers B. Tech in Electronics engineering and M. tech in
Electronics &Communication Engineering. The department is guided by a team of qualified,
dedicated and competent faculty members. A highly experienced and dedicated faculty imparts
quality training to student, with great emphasis being laid on understanding the fundamentals and
applying it to solve the challenging problems. The department has excellent infrastructure to
explore technological advancements in the relevant fields. All the labs in the department are well
equipped with the state of the art equipment enabling to undertake research/industry oriented
projects both at under-graduate and post graduate level.

SPONSORS
The conference is supported by IEEE UP Section, Sponsored by Wind River Inc. and also
supported by IE, & QUARBZ and other organizations.





Advisory Committee

1.Prof. K.P. Singh, Director, IT BHU, Varanasi,U.P.
2.Prof. Y.N.Singh, IIT, Kanpur,U.P.
3.Prof. S.N.Singh, IIT, Kanpur,U.P.
4.Prof. B.P.Singh, MIST Laxmangarh, Seekar, (Raj.)
5.Prof. J.P.Saini, Principal MMMEC, Gorakhpur,U.P.
6.Prof. D.R.Bhaskar, JMI, Jamianagar, New Delhi
7.Prof. U.S. Triar, Dean, NIT Patna, Bihar
8.Prof. M.M.Sharma, Deptt. of E&C, NIT Jaipur, (Raj.)
9.Dr. Nishchal Verma, IIT, Kanpur, U.P.
10.Prof. Ajit Kumar Panda, NIST, Berhampur,Odisha
11. Prof. J.P. Pandey, KNIT, Sultanpur, U.P.
12.Prof. G.S.Tomar, Principal, M.I.T.M. Gwalior, M.P.
13.Prof. Sanjeev Jain, Director M.I.T.S. Gwalior, M.P.
14.Mr. Venkatesh Kumaran, Country Manager(India), Wind River Inc.

Organizing Committee:
Patron:
Prof. R. K. Khitoliya, Director, H.B.T.I. Kanpur

Convener: Dr. Krishna Raj, H.B.T.I. Kanpur

Members:
1. Sri G.P.Bagaria, H.B.T.I. Kanpur
2. Dr. Rachna Asthana, H.B.T.I. Kanpur
3. Smt. Rajani Bisht, H.B.T.I. Kanpur
4. Sri M.K.Shukla, H.B.T.I. Kanpur
5. Sri A.K.Shankhwar, H.B.T.I. Kanpur
6. Sri Ashutosh Singh, H.B.T.I. Kanpur
7. Sri R.C.S. Chauhan, H.B.T.I. Kanpur

Address for Communication
Dr. Krishna Raj
Head of the Department,
Department of Electronics Engineering,
Harcourt Butler Technological Institute,
Kanpur, (U.P.) 208002
e-mail: kraj_biet@yahoo.com; sprtos@rediffmail.com
Ph.05122534001-5 Ext 208, Fax.05122533812
Mob.9721456087, 9450607840


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 1 of 15

COMMUNICATION ENGINEERING RELATED PAPERS
S.NO


PAPER
NO.
PAPER TITLE AUTHOR(S) NAME EMAIL-ID
1 COM0101 Comparison of LFSR and
CA as Pseudorandom
Number Generator

Ashutosh Mishra*, Dr. Harsh
Vikram Singh**, S.P.
Gangwar**
*Student (M.Tech), ** Astt.
Prof. Dept. of Electronics
Engineering, KNIT
SULTANPUR
ashutosh.suryasurya@gmail.com
2 COM0102 Cost-effective markov
based segmentation
S.Robinson
Sobitha Raj, Mrs.V.S.Bindhu

robins.singh@yahoo.co.in
bindhuvs2003@gmail.com
3 COM0104 Performance Comparison
of VANET Routing
Protocols.
NITIN SHARMA
Department of Computer
Science and Engineering
MNIT, Allahabad, India
nsharma283@gmail.com
Mobile: +91 8861 6365 11
POOJA RANI
Department of Information
Technology
ITM, Sector-23A, Gurgaon,
India
me_pooja@msn.com
Mobile: +91 9899 7470 33
nsharma283@gmail.com
me_pooja@msn.com


4 COM0105 Greedy Grid Scheduling
Algorithm in Static
Environment.
.
POOJA RANI
Department of Information
Technology
ITM, Sector-23 A, Gurgaon,
India
E-mail: me_pooja@msn.com
Mobile: +91 9899 7470 33
NITIN SHARMA
Department of Computer
Science & Engineering
MNNIT, Allahbad, India
E-mail: cs0802@mnnit.ac.in
Mobile: +91 8861 6365 11
me_pooja@msn.com
nsharma283@gmail.com


5 COM0106 A comparative study of ip
and mpls based network
Sandeep singh
M.Altamash Sheikh
School of Information and
Communication Technology,
Gautam Buddha University
er.sandeep_vhdl@yahoo.com

altsheikh@gmail.com

6 COM0107 INTEGRATED
SERVICES DIGITAL
NETWORK
AJAY BHATIA
7 COM0108 PERFORMANCE
ANALYSIS OF WIRED
NETWORKS UNDER
VARYING NETWORK
ATTRIBUTES
Vaibhav Nijhawan

Department of Electronics &
Communication Engineering
Allenhouse Institute of
Technology, Kanpur
vaibhav.nijhawan@gmail.com
8 COM0109 OFDM PAPR reduction
based on nonlinear
functions without BER
Degradation and out of
band emission
Neelam
Suchita
neelamdewangan@csitdurg.in
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 2 of 15

9 COM0110 Analysis on Placement of
Wavelength Converters in
WDM p-Cycle Network

Rupali Agarwal
Deptt. Of EC, BBD, Lucknow
Dr. Rachna Asthana
Asso. Professor(HBTI
Kanpur)
roopali.ipec@gmail.com

10

COM0201 COMPARATIVE
ANALYSIS OF
WIRELESS NETWORKS
USING OPNET
Yaduvir Singh
1
, Vikas Gupta
2
,
Rahul Malhotra
3

rlyadava@yahoo.com
11 COM0202 Comparative Analysis of
Core Migration
Techniques in Wireless Ad
Hoc Networks
Rahul malhotra
(ECE Bhai Maha Singh
College
of Engineering, Sri Muktsar
Sahib by Er. Rahul Malhotra
HOD ,ECE .)
Reena aggarwal
rlyadava@rediffmail.com
12 COM0203 Performance comparision
of
RI & TBI based system in
fading envoirement using
MRC scheme
M. Shukla
Department of Electronics
Engineering
Harcourt Butler
Technological Institute
Kanpur, India
manojs@hbti.ac.in
Aasheesh Shukla, Rohit,
Sankalp
Department of Electronics &
Communication Engineering
GLA University
Mathura, India

aasheeshshukla@gmail.com
13 COM0205 Security Constraints in
Peer-to-Peer Network on
Overlay Networks
Raj Kumar Gaur
Divya Gupta

rajkumargaur@gmail.com
guptadivya28@gmail.com

14 COM0206 ENERGY
CONSERVATION
THEORY IN CONTEXT
AWARE COMPUTING
Umesh Chaudhary ,
Chandrika Prasad Department
of Computer Science, Banaras
Hindu University Varanasi-
221005
umesh9794@gmail.com,
chandrika.bhumca08@gmail.com
15 COM0207 AUTHENTICATE
SECURE ROUTING
PROTOCOL

ABHAY SHUKLA (M.Tech.,
persuing P.hd.)
Asociate t. Prof.
Department of Computer
Science and Technology
SSAIET, Kanpur*
9235651921

abhay002@sify.com
16 COM0209 Improvement of Hata
Propagation Prediction
Model for Suburban Area

Himani Kaur*, Er. Ravneet
Singh **

sapna.katiyar@abes.in

17 COM0210
Design of Stepped-
Impedance Microstrip Line
Low Pass Filter for
Wireless Communication

Navita Singh(Sr. Lecturer(EC
Deptt.)
KIET, Ghaziabad),
Saurabh Dhiman, Prerna Jain,
Shrestha, Tanmay Bhardwaj
singh.navita@rediffmail.com

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 3 of 15

18 COM0211 End Coupled Band Pass
Filter for Wireless
Communication

Avnish Kumar ,Manohar
Kumar, Ashish Kumar Verma,
Navita Singh
Final Year (ECE)
KIET, Ghaziabad


avi.k.prajapati@gmail.com

19 COM0212 Security of electronic
money
(RFID CREDIT CARD)
PERFORMANCE
MR. Pankaj singh
Mr. Sanyam agarwal
Mr. Rohit sharma
Singhpankaj76@gmail.com
20 COM0213 High temperature mems
based transmitter for
wireless sensor and
communication network.
Ruchi gupta
Asst. prof JPIET

ruchi.siet@gmail.com
21 COM0214 An overview on load
balancing routing
protocols for mobile adhoc
networks
Bhawana bharadwaj
Sanjay kumar
bhavana.bhardwaj@gmail.com

22 COM0215 Localization of Wireless
Sensor Network using
Geographical routing

Saurabh Dixit
1
and Arun
Kumar Singh
2

1.Member IEEE, IETE
Department of Electronics and
Communication
Northern India Engineering
College, Lucknow

saurabh2911@ieee.org
aksingh_uptu@rediffmail.com
23 COM0216 Timestamp based
loadbalancing in aodev in
mobile adhoc networks
Bhawana bharadwaj
Sanjay kumar
bhavana.bhardwaj@gmail.com

24 COM0302 Performance of Cross-
Shaped Junction
Microstrip Antenna using
IE3D

Jyoti Gupta, Babita Singh, D.
C. Dhubkarya
Department of Electronics &
Communication Engineering,
B.I.E.T., Jhansi (U.P.) INDIA

gold.jyoti@gmail.com,
babitasngh20@gmail.com,dcd3580@
yahoo.com


25 COM0303 Broadband Slot Loaded
Microstrip Patch Antenna

A. Ansaari , kamakshi, Ram
Brij Ram, A.Singh, Anurag
Mishra, N.P.Yadav
Department of Electronics and
Communication
University of Allahabad
kamakshi_iert_el@yahoo.co.in,
mishraanurag31@gmail.com
26 COM0304 Microstrip bandpass filter
at 6 ghz for wimax
Vanish kumar,
Navita singh
Manohar kumar
Ashutosh kr. Varun

avi.k.prajapati@gmail.com
27 COM0305 Antenna selection
algorithm for MIMO BC
channel using partial side
information

Swadhin kumar mishra
Prabina pattanayak
A.k. panda
swadhin.mishra@gmail.com
28 COM0306 HILBERT ANTENNAS
IN MOBILE , RADAR
AND LIGHT COMBAT
AIR CRAFT
TECHNOLOGY
KUMAR ASHOK
[1]
,
D.SUNITA
[2]
,SINGH
PUSHPENDRA
[3]


akprasad212@yahoo.co.in
sunitadeviiet@yahoo.co.in
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 4 of 15

29 COM0401 PERFORMANCE of
SOFT DECODING
VERSES HARD
DECODING
Archana Sharma, Mohd
Murtaja, SCRIET,
CCS.University Campus,
Meerut
Prof. (Dr.) Rajeev Kapoor,
H.O.D. of EC, Delhi
Technical University, Delhi

Archna.mtech@gmail.com
,murtaza_amu@yahoo.com,
rajeevkapoor@dce.ac.in
30 COM0402 PERFORMANCE
EVALUATION OF
VARIOUS ERROR
CORRECTING CODES
ON PLC SYSTEM
Priyank Sharma
Ravikant Saini
sh_priyank@yahoo.com
31 COM0421 Design and Analysis of
Microstrip Antennas for
W-LAN and Ku-band
Applications

1
Ravindra Kumar Yadav and
2
Ram Lal Yadava
1
Department of Electronics
and Communication
Engineering,
I.T.S. Engineering College,
Greater Noida, Uttar
Pradesh, India
2
Department of Electronics
and Communication
Engineering,
Galgotia’s College of
Engineering & Technology,
Greater Noida, Uttar Pradesh
ravipusad@gmail.com
32 COM0422 A Proposed Design of One
Dimensional Optical
Orthogonal Coding
Scheme and Analysis with
Others
1
Ratnesh Kumar,
2
R.C.S.
Chauhan,
3
G.P.Bagaria
1
Department of Electronics
Engineering, H.B.T.I.,
Kanpur, U.P., India
iam_ratnesh@yahoo.co.in

33 COM0431 Error modelling of BFSK
over Rayleigh Fading
Channel – A Statistical
Frame Work

Vicky Singh
1
, Amit Sehgal
2
,
Rajeev Agrawal
3

1,2,3
ECE Department, G.L.
Bajaj Institute of Technology
and Management, Gr. Noida
vicky.glbitm@gmail.com,
amitsehgal09@gmail.com,
rajkecd@gmail.com


COMPUTER RELATED PAPERS
Sr.No Paper Code Paper Title Author(s) Name Email ID
1 COP0101 Critical Analysis on
Prosthetic Arm
Anil kumar,
Dr. Monika Jain


anil00699@gmail.com

monika_jain24@rediffmail.com

2 COP0102 Analysis of Bio-
inspired
Algorithms– BFO,
PSO and FFA
Vipul Singhal

Vipul.vanshi@gmail.com

3 COP0104 Biometric
authentication using
image processing
tools
Neha mittal
Ruchi gupta
Pratibha singh

Nehamittal2008@gmail.com
4 COP0105 Medical image
enhancement in
spatial domain
:statical analysis
Sanjeev kumar gupta
,Sonika singh ,Rahul
singh

s.gupta@alumni.iitg.ernet.in

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 5 of 15

5 COP0301 AN INTELLIGENT
ALGORITHM FOR
TASK
ALLOCATION IN
DISTRIBUTED
NETWORK USING
STATIC
APPROACH
1
Kapil
1
Kapil Govil,
2
Dr. Avanish Kumar
and
3
Rakesh Kumar
Dwivedi

drkapilgovil@gmail.com
dravanishkumar@yahoo.com,
r_dwivedi2000@yahoo.com
6 COP0302 Improvement in
Dijkstra algorithm

Ashutosh Singh,
Sandeep Rathour


singhash@iitk.ac.in
ersandeep.hbti@gmail.com

7 COP0303
Executed Wrist
Movement
Classification For
Brain Computer
Interface


Mariam Siddiqui,
Ayman Nidal
Khalidi, Omar
Farooq,Yusuf
U Khan

ankhalidi@gmail.com

8 COP0304 Color Images
Enhancement Based
on Piecewise Linear
Mapping
Anubhav Kumar,
Awanish Kumar
Kaushik,R.L
Yadav,Divya Sexena
rajput.anubhav@gmail.com
9 COP0305
Information Fusion
Based on
Multiplicity of Data
Preprocessing
Improves AdaBoost
Classification
Efficiency
Divya Somvanshi
and R. D. S. Yadava


divya.jrf@gmail.com
10 COP0306 Real-time Hand
Gesture Recognition
Techniques for
Human-Computer
Interaction
ANSHUL
SHARMA

getanshulsharma@gmail.com


11 COP0307 Forensic Research
by Face Reading
Systems

Pramod Kumar Pal pramod_pal123@rediffmail.com

12 COP0401 An Automatic
Machine Learning
and Particle
Filtering Based
Approach to Real
Time Human
Tracking in Videos

Chandra Mani
Sharma
1
, Alok
Kumar Singh
Kushwaha
1
, Ashish
Khare
2
and Sanjay
Tanwani
1

chandramani.sharma@yahoo.com,
alokkumarsingh.jk@gmail.com,
akhare@allduniv.ac.in,
sanjay_tanwani@hotmail.com
13 COP0402 An Efficient Image
Segmentation
Method Based on
Normalized Cut
Vaibhav Sharma vaibhavshar@gmail.com

14 COP0501 OPTIMIZATION
OF WIRED
NETWORKS
USING LOAD
BALANCERS
Yaduvir Singh
1
,
Vikas Gupta
2
, Rahul
Malhotra
3

rlyadava@yahoo.com

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 6 of 15

COMPUTATION & COMPUTER OPERATING SYSTEM RELATED PAPERS
S.NO PAPER
NUMBER
PAPER TITLE AUTHOR(S) NAME EMAIL-ID
1 COS0101 TEMPORAL
DIFFERENCES IN
PROCESSING OF
SYMMETRIC OBJECTS
IN HUMAN BRAIN
Taslima Ahmed,


Umesh Kumar
3
Prasant kumar

"ahmed.taslima@gmail.c
om"

2 COS0102 Modelling Neuron for
Biomedical Applications:
A Review

Taslima Ahmed ,
Dr Jiten Ch Dutta


jitend@tezu.ernet.in

3 COS0103 Organ Failure Assessment
in Malarial Patients
Using Artificial neural
networks

Mr. Jay Kumar Pandey
Shri Ramswaroop
Memorial College of
Engineering and
Management , Lucknow ,
U.P

er.jay11@gmail.com

4 COS0201 AN INTELLIGENT
CONTROL METHOD TO
REDUCE BRAKE NOISE

Pramod Kumar Pandey
daddapramodpandey@gm
ail.com
Department of Electrical
Instrumentation
Engineering Thapar
University Patiala Punjab
-147004
Yaduvir Singh

dryaduvirsingh@gmail.co
m
Department of Electrical
Instrumentation
Engineering Thapar
University Patiala Punjab
-147004

daddapramodpandey@g
mail.com

5 COS0202 APPLICATION OF
FUZZY LOGIC IN
AUTOMATIC BOTTLE
FILLING SYSTEM
Nikita Agarwal
Ritika Srivastava

Preeti Dhiman



6 COS0203 Real Time Control of
Rotary Flexible Joint with
LQR and Fuzzy Controller
Mr. Jay Kumar Pandey
Shri Ramswaroop
Memorial College of
Engineering and
Management , Lucknow ,
U.P
er.jay11@gmail.com


7 COS0204 Fuzzy Logic Control for a
Speed Control of
Induction Motor using
Pulse Width Modulation
Mr. Jay Kumar Pandey
Shri Ramswaroop
Memorial College of
Engineering and
Management , Lucknow ,
U.P

er.jay11@gmail.com

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 7 of 15

8 COS0205 Optimal Placement
techniques of Facts
Controllers in multi
machine power system
environments: A literature
Suvey
Bindeshwar Singh
N.K.Sharma,A.N.Tiwari
Department of Electrical
Engineering
Kamla Nehru Institute of
Technology
Email:bindeshwar.singh2
025@gmail.com

bindeshwar_singh2006@
rediffmail.com
9 COS0206 Fuzzy logic and ruled
based system: research
issues and challanges
Praveen kumar shukla
Surya prakash tripathi
farhad_aei@yahoo.com

10 COS0301 A comparative analysis of
controllers controlling
uncertain factors affecting
the robust position control
of DC motor

Farhad Aslam
Gagandeep Kaur
Deptt. of
Electrical &
Instrumentation
Engg.
Thapar University,
Patiala.

farhad_aei@yahoo.com

11 COS0302 A SURVEY ON
CURRENT CONVEYOR:
NOVEL UNIVERSAL
ACTIVE BLOCK

Indu Prabha Singh and
2

Dr. Kalyan Singh
1
Deptt. of Electronics and
Comm. Engg. SITM,
UNNAO-209859, India
2
Dept. of Physics and
Electronics Engg. , Dr.
RML AVADH
UNIVERSITY,
FAIZABAD, India
induprabhasingh@gmail.
com,
ksshrinet@yahoo.co.in
12 COS0401 A comparative study of
genetic algorithm and the
particle swarm
optimization
Sapna Katiyar, Deepika
Pandey, Sakshi Chhabra,
Vaishali Gupta
Sapna.katiyar@abes.in

SIGNAL PROCESSING & RELATED PAPERS

S.NO
PAPER
NUMBE
R
PAPER TITLE AUTHOR(S) NAME EMAIL-ID
1

SIP0101

MEMS Based
MicroResonator Design
& Simulation Based On
Comb-Drive Structure

Mr. Prashant Gupta
Ideal Institute of
Technology, Ghaziabad
prashant_iit@ieee.org

2 SIP0102 Different Look-Ahead
Algorithm for Pipelined
Implementation of
Recursive Digital Filters
VivekanandYadav
Krishna Raj
vivekanand.hbti@gmail.com

3 SIP0103 Study of MC-EZBC and
H.264 Video Codec
Agha Asim1Husain and
Agha Imran2 Husain
1Deptt of Electronics &
Comm. Engg, ITS Engg
College, 201301, India
2Deptt of Computer
Science & Engg, MRCE,
121004, India

aghahusain@gmail.com,
aghaimrank2k@yahoo.com
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 8 of 15

4 SIP0105 FPGA BASAED
IMPLEMENTATION
OF
IIR FILTERS
Anup Saha, Saikat Karak,
Surajit Kangsabanik, and
Joyita RoyChowdhury
anupsaha.0733@gmail.com,
anupsaha.0733@gmail.com
karaksaikat@yahoo.com,
surajitkangsabanik@gmail.com,
joyitaroychowdhory@yahoo.in
5

SIP0106 Enhanced Clocking Rule
for A5/1 Encryption
Algorithm
Rosepreet Kaur Bhogal,
ECE Dept.,
Nikesh Bajaj, Asst. Prof.,
ECE Dept.
Lovely Professional
University -India
rosepreetbhogal155-2006@lpu.in
nikesh.14730@lpu.co.in
6 SIP0107 An Application of
Kalman Filter in State
Estimation of a Dynamic
System
Vishal Awasthi
1
Krishna Raj
2

awasthiv@rediffmail.com
7 SIP0108 Wideband Direction of
Arrival Estimation by
using
Minimum Variance and
Robust Maximum
Likelihood Steered
Beamformers: A Review

SANDEEP SANTOSH
1
,
O.P.SAHU
2
, MONIKA
AGGARWAL
3

Astt. Prof., Department of
Electronics and
Communication
Engineering
1
,
Associate Prof.,
Department of Electronics
and Communication
Engineering
2
,
National Institute of
Technology , Kurukshetra,

Associate Prof., Centre
For Applied Research in
Electronics (CARE)
3
,
Indian Institute of
Technology, New Delhi.
3
INDIA
profsandeepkkr@gmail.com
8 SIP0109 Electricity Generation by
People Walk through
Piezoelectric Shoe: An
Analysis

1.
Dr. Monika Jain,
2.
Ms. Usha Tiwari,
3.
Mohit
Gupta,
4.
Magandeep singh Bedi
1.
Member IEEE, IETE,
Professor-Dept of Electronics &
Instrumentation Engg,
Galgotias College of Engineering
& Technology,Greater Noida,UP,
INDIA
2.
Assistant Professor-Dept of
Electronics & Instrumentation
Galgotias College of Engineering
& Technology,Greater Noida,UP,
INDIA
.3&4.
B.Tech, 4
th
year student Dept
of Electronics & Instrumentation
Galgotias College of Engineering
& Technology,Greater Noida,UP,
INDIA

1
monika_jain24@rediffmail.com
2
usha.pant@rediffmail.com.
3
mohit_luvkush@yahoo.co.in
4.
magandeep.bedi@gmail.com


9 SIP0110 Spectral and Cepstral
analysis Using Modified
Bartlett Hanning
Window
Rohit Pandey
1
, Rohit
Kumar Agrawal
1
, Sneha
Shree
1

rohitpandey2108@gmail.com,
rohitagrawal07162@gmail.com

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 9 of 15

10 SIP0111 A 3d approach to face
expression recognition


Akshay gupta
Garima chandel

akshaygupta.abes@gmail.com
11 SIP0112 Performance evaluation
of signal selective DOA
tracking for wideband
cyclostationary sources
Sandeep Santosh, O P
Sahu, Monika Aggarwal
profsandeepkkr@gmail.com
12

SIP0113 BARTLETT
WINDOWED FAST
COMPUTATION OF
DISCRETE
TRIGNOMETRIC
TRANSFORMS FOR
REAL TIME DATA
PROCESSING
ABHIJIT KHARE
SHUBHAM VARSHNEY
VIKRAM KARWAL
khareabhijit14@gmail.com
shubham7502909dece@gmail.co
m
vikram.karwal@jiit.ac.in


13 SIP0114 Lossless compression
scheme based on
prediction for bayer
color filter
Anita patil
Sudhirkumar D. sawarkar
Nareshkumar harale
patilanitau@gmail.com
14 SIP0201 OPTIMAL RECEIVER
FILTER DESIGN
Vivek kumar
Dr. k. raj
Vivekhbti07@gmail.com
kraj_biet@yahoo.com
15 SIP0202 Signal Acquisition and
Analysis System Using
LabVIEW
Subhransu Padhee,
Yaduvir Singh
subhransu_padhee@yahoo.com,
dryaduvirsingh@gmail.com
16 SIP0203 METHODS OF
INTERCARRIER
INTERFERENCE
CANCELLATION
FOR ORTHOGONAL
FREQUNCY DIVISION
MULTIPLEXING



Dr.R.L.Yadav
Mrs.Dipti Sharma

rlyadava@rediffmail.com
dipti_sharma510@yahoo.co.in

17 SIP0204 OBJECT DETECTION
BASED ON CROSS-
CORRELATION
USING PARTICLE
SWARM
OPTIMIZATION


Sudhakar Singh
Yaduvir Singh

sudhakarsingh86@gmail.com

18 SIP0205 Multisegmentation
through wavelets:
Comparing the efficacy
of Daubechies vs
Coiflets
Madhur
Member, IEEE,
Yashwant Yashu, Member,
IETE, Satish K.
Member, IEEE
madhur.manas@gmail.com
yashwantyashu.jiet@gmail.com
satish432002@gmail.com

19

SIP0206 Analysis of Signals in
Fractional Fourier
Domain
Ajmer Singh, Student of
Lovely Professional
University(LPU)-India,
Nikesh Bajaj, Asst. Prof.,
ECE Dept.(LPU

ajmersingh155-2006@lpu.in
, nikesh.14730@lpu.co.in
20 SIP0207 Parzen-Cos
6
(πt)
combinational window
family based QMF bank
Narendra Singh
(*)
and
Rajiv Saxena,
Jaypee University of
Engineering and
Technology, Raghogarh,
Guna (MP)
narendra_biet@rediffmail.com ;
narendra.singh@jiet.ac.in

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 10 of 15

21

SIP0208

Performance Analysis of
Sub Carrier Spacing
Offset in Orthogonal
Frequency Division
Multiplexing System

Shivaji Sinha, Member
IETE, Rachna Bhati,
Dinesh Chandra, Member
IEEE & IETE

Department of Electronics
& Communication
Engineering, JSSATE
Noida
shivaji2006@gmail.com,
rachna.bhati1988@gmail.com
22 SIP0301 A Comparative Analysis
of ECG Data
Compression
Techniques
Sugandha Agarwal
Amity School of
Engineering and
Technology
Amity University Uttar
Pradesh,
Lucknow

Sugandhaa7@gmail.com
23 SIP0303 Biologically inspired
Cryptanalysis- A Review






Ashutosh Mishra*, Dr.
Harsh Vikram Singh**,
S.P. Gangwar**
*Student (M.Tech), **
Astt. Prof. Dept. of
Electronics Engineering,
KNIT SULTANPUR
ashutosh.suryasurya@gmail.com
24 SIP0304 Eye Based Cursor
movement using EEG in
brain computer interface.
Tariq S khan, Mudassir
Ali, Omar Farook,Yusuf U
Khan
tariq252@gmail.com

25 SIP 0333
AN INTERNET
BASED
INTELLIGENT
TELEDIAGNOSIS
SYSTEM FOR
ARRHYTHMIA

K.A.Sunitha

N.Senthil kumar
K.Prema
3

Sandeep Kotikalapudi

sunithasrm@yahoo.in

nsk_vnr@yahoo.com
premsb4u@gmail.com
sandeep_yours18@yahoo.com


26

SIP0401 Projected View &
Novel Application of
Context Based Image
Retrieval Techniques

Shivam agarwal
Rajeev Singh Chauhan
*
,
Vivek Vyas
shivam.agrawal@live.com ,
101.vivek@gmail.com
rajeev0507@hotmail.com

27 SIP0402 Recursive Algorithm and
Systolic Architecture for
the Discrete Sine
Transform

M.N. Murty,
Department of Physics
NIST, Berhampur-761008,
Orissa, India

sudhansusekharnayak@yahoo.co
m
mnarayanamurty@rediffmail.co
m
28 SIP0403 Multiscale Edge
Detection Based on
Wavelet Transform



Divesh Kumar, Dr.
Yaduvir Singh

kambojdivesh@gmail.com,
dryaduvirsingh@gmail.com

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 11 of 15

29 SIP0404 Color Image
Enhancement by Scaling
Luminance and
Chromatic Components

1
Satyabrata Das,
2
Sukanti
Pal and
3
Ajit Kumar Panda

satyabratadas.m@gmail.com,
sukanti.nist@gmail.com,
akpanda62@hotmail.com
30 SIP0405 A Tutorial on Image
Compression
Techniques

1
Vedvrat,
2
Krishna Raj
1
Department of Electronics
& Communication
Engineering A.I.T,
Kanpur, U.P., India
2
Department of Electronics
Engineering H.B.T.I.,
Kanpur, U.P., India

1
r.ved.hbti@gmail.com,
2
kraj_biet@yahoo.com
31

SIP0406 Comparative Study of
Lifting –based Discrete
Wavelet transform
Architectures

Vidyadhar Gupta, Krishna
Raj

Department of electronics
engineering

Harcourt Butler
Technological Institute,
Kanpur

vidyadhar.gupta@gmail.com

32 SIP0407
A Novel Approach in
Image De-noising for
Salt & Pepper Noise

J S Bhat

B N Jagadale
Lakshminarayan H K

jsbhat@kud.ac.in
basujagadale@gmail.com;
lakkysharmahk@yahoo.co.in
33 SIP0427 Content Based Image
Retrieval System for
Medical Images


Prof.K.Narayanan
1
,
Shaista Khan
2
1
Asst.Professor, Fr.Agnel
College of Engg.,
University of Mumbai,
India
unniknarayan@gmail.com
khan_shaista26@yahoo.co.in


34 SIP0501 AUDIO + Abhay Kumar
Research Scholar at
Associated Electronics
Research Foundation,
Phase-II Noida (U.P.)

abhay.2t@gmail.com
35 SIP0502 Speaker Identification

Prerana & Aditi
Choudhary
M.tech (DC)
DIT DEHRADUN

prerana6289@yahoo.com

36 SIP0503 Modeling of FBAR
Resonator and
Simulation using
APLAC

Deepak kumar, Navaid
Z.Rizvi,Rajesh Mishra
Gautam Buddha
University,Greater Noida

dkumar.gbu@gmail.com


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 12 of 15

37 SIP0506 Role of Speech
Scrambling and
Encryption in Secure
Voice Communication

Himanshu Gupta
Prof. (Dr.) Vinod Kumar
Sharma

himanshu_gupta4@yahoo.co.in
vks_sun@ymail.com

VLSI & RELATED PAPERS
S.No. Paper
Number
Paper Title Author(s) Name E-mail ID
1. VLP0101 Low power On-Chip
Amplifier for CCD
Array
Er. Rahul Malhotra*, Er.
Amit Kumar, Bhai Maha
Simgh College of
Engineering, Sri Muktsar
Sahib, India
blessurahul@gmail.com,
amitkumar_sgnr@yahoo.co.in
2. VLP0102 Programmable Input
Output Resistances
of FET Amplifier

Mrs. Meena Singh , Dr. B. P.
Singh

, Mr. Arun Kumar Singh
, Madan Mohan Malaviya
Engg. College, Gorakhpur
meena71_singh@rediffmail.co
m, singh16.arun@gmail.com,
bpsingh@ieee.org
3. VLP0103 RELIABILITY
PREDICTION FOR
IGBT BASED
INVERTERS
UNDER
DIFFERENT
SWITCHING
PATTERNS
Fuzail Ahmad, S.K.Singh,
Amit Kumar Verma,
CENTRE
GORAKHPUR,INDIA,
DOEACC CENTRE
GORAKHPUR,INDIA
IBM GURGAON,INDIA

1
er.fuzailahmad@hotmail.com
3amitkverma@in.ibm.com
sksingh@doeaccgkp.edu.in

4. VLP0104 Carry Ripple Adder
based on Charge
Recycling
for Lower Energy
MTCMOS
Arvind Kumar, Member,
IEEE , Sanjeev Rai, Sarad
Shrestha, ECED,
MNNIT,Allahabad
sarad693@gmail.com
arvindk@mnnit.ac.in
5. VLP0105 Forthcoming CMOS
Technology in
Nanoscale Era
Shashank Mishra, Kshitij
Bhargava, Rohit Tripathi,
Piyush Jain

Jaypee Institute of
Information Technology,
Noida


1
shashankmishra05EI44@gmail
.com
2
kbhargava3@gmail.com
3
rohit.tri2008@gmail.com
piyush.oct@rediffmail.com
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 13 of 15

6. VLP0106 On Demand
Simulation of Input
and Output
Resistances of
MOSFET Amplifier
Mrs. Meena Singh , Dr. B. P.
Singh

, Mr. Arun Kumar Singh
, Madan Mohan Malaviya
Engg. College, Gorakhpur
meena71_singh@rediffmail.co
m, singh16.arun@gmail.com,
bpsingh@ieee.org
7. VLP0107 Performance
Analysis and
Comparison of
PFSCL and MCML
Kirti Gupta , Ranjana
Sridhar, Jaya Chaudhary
DTU
ranjanasridhar@gmail.com

8. VLP0108 Comparative Study
of Fast Adders
using VHDL and
FPGA
Nishi Chandra
Rajani Bisht,
H.B.T.I.,Kanpur
lognishi@gmail.com
9. VLP0109 Organic Thin Film
Transistor:
Materials,
Structures and
Operational
Parameters

Poornima Mittal
1
, Brijesh
Kumar
2
, B. K. Kaushik
3
, Y.
S. Negi
4
and Krishna Raj
5

1
Electronics and
Communication
Engineering, Graphic Era
University, Dehradun,
INDIA
3
Electronics and Computer
Engineering, Indian Institute
of Technology, Roorkee,
INDIA
2,4
Polymer Science and
Technology Group, DPT,
Indian Institute of
Technology, Roorkee,
INDIA
5
Department of Electronics
Engineering, H.B.T.I.,
Kanpur, INDIA
poornima2822@gmail.com,
brijesh2228@gmail.com,
bkk23fec@iitr.ernet.in,
ynegifpt@iitr.ernet.in,
kraj_biet@yahoo.com

10. VLP0110 CHARACTERIZAT
ION OF 4T SRAM
CELL
Setu Garg ,GCET, Greater
Noida
Prof.S.N.Sharan, GNIT,
Greater Noida
, Garima Chandel
Member IEEE, Hridesh
Verma
ABES IT Ghaziabad, India
gargsetu06@gmail.com,
snathsharan@yahoo.com,
garimachandel@rediffmail.com,
hridesh.verma@gmail.com



CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 14 of 15

11. VLP0111 Quantitative
Analysis and
Optimization
Techniques
for On-Chip Cache
Leakage Power
Vikas Tiwari
M.Tech (VLSI Design
ITM, Gwalior, India
Shyam Akashe ,Associate
Professor. ITM, Gwalior
Rajkumar Rajoriya Assist.
Professor ITM, Gwalior
vikas.esvd@gmail.com
vlsi.shyam@gmail.com
12. VLP0112 Fault Tolerant
Design for analog
and digital circuits
Dr. Anand Mohan,Akhilesh
Pathak, Tarang
Agarwal,Trailokya nath
sasama(IIT Bhu)
Pathak.akhilesh@gmail.com
Sasamal.trailokyanath@gmail.c
om
13. VLP0113 Floating Point
Arithmetic
Operations Using
VHDL

S.C. Yadav, S. S. Chauhan
1
,
A. R. Khan
2

Electronics &
Communication Engg.,
1,2
Graphic Era University
566/6 Bell Road, Dehradun
(India)
subhash.yadav775@gmail.com,
sudakarnith@gmail.com
1
,
abdulamu@gmail.com
2

14.
a
VLP0114 A Complete CMOS
Based Low Power
Supply Bandgap
Voltage Reference
Circuit
Implemented On
TSMC 0.35-μm
Process

Kshitij Bhargava
#1
,
Kirmender Singh
*2

ECE Department
(Microelectronics And
Embedded Technology)
Jaypee Institute Of
Information Technology
University, Noida
India
1
kbhargava3@gmail.com

1
kshitijbhargava.09305160@gm
ail.com
2
kirmender.singh@jiit.ac.in
15. VLP0115 Performance
analysis of carbon
nanotube FET
Harish kumar mishra
S.P gangwar
Dr. Harsh v.singh
Knit sultanpur


rblharish@gmail.com
harshvikram@gmail.com
16.

VLP0201 POWER AWARE
PHYSICAL
MODEL FOR
EMBEDDED
SYSTEMS

Asstt Prof Yasmeen Hasan
Mtech(Electronic Circuits
&Systems (VLSI))
DEPT OF ECE, INTEGRAL
UNIVERSITY, LUCKNOW
yasmeen.hasan9@gmail.com
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

LIST OF PAPERS

Page 15 of 15

17. VLP0202 Efficient Power
Utilisation By
Controlling
Industrial And Home
Appliances Using
GSM and
Microcontroller
Raj Singh Yadav and Nidhi
Mishra
Krishna Institute of
Engineering and Technology
Electronics and
Communication Department
Ghaziabad India
yadav.raj31@gmail.com
nidhi.mishra9@gmail.com

18. VLP0301
Design and
Implementation of
Radix-2 & Radix-4
Booth Multipliers
Using VHDL
S. S. Chauhan, S.C. Yadav,
A. R. Khan
Graphic Era University
sudakarnith@gmail.com,
subhash.yadav775@gmail.com,
abdulamu@gmail.com


19. VLP0302
A Novel Approach
to Design of a
Multiplier Using
Reversible Logic
Gates
S. S. Chauhan, S.C. Yadav,
A. R. Khan
Graphic Era University
sudakarnith@gmail.com,
subhash.yadav775@gmail.com,
abdulamu@gmail.com


20. VLP0401 OTRA based
Grounded Inductor
and its application

Rajeshwari Pandey(member
IEEE),Neeta
Pandey(member IEEE),Ajay
Singh, B.Sriram,
Kaushalendra Trivedi
Delhi Technological
University, Delhi
rajeshwaripandey@gmail.com
21. VLP0402 OTRA based
Precision Full Wave
Rectifier

Rajeshwari Pandey (member
IEEE)

, Ajay Singh,
B.Sriram, Kaushalendra
Trivedi
rajeshwaripandey@gmail.com
22. VLP0403 GaN-based HEMTs
for Communication
Circuits

T R Lenka
1
and A K Panda
2

National Institute of Science
and Technology
Palur Hills, Berhampur,
Odisha
trlenka@gmail.com,
akpanda62@hotmail.com
23. VLP0405
Ternary logic in
digital
communication for
high speed and
performance









vinay.ic01@gmail.com,
vinay.ic1@rediffmail.com

CONFERENCE ON ―SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)‖ MARCH 26-27 2011
COMO1O1-1

Comparison of LFSR and CA as Pseudorandom
Number Generator
Ashutosh Mishra*, Dr. Harsh Vikram Singh**, S.P. Gangwar**
*Student (M.Tech), ** Astt. Prof. Dept. of Electronics Engineering, KNIT SULTANPUR

Abstract

Pseudorandom number generator (PRNG) is
a Key element in stream cipher. This paper
compares two techniques: Linear Feedback
Shift Register (LFSR) and Cellular
Automata (CA), used for pseudorandom
number generation. Both LFSR and CA are
analyzed based on their construction and
characteristics. A comparison of LFSR and
CA is presented to demonstrate their
shortfalls and suitability to certain
applications.

Introduction

In cryptography, a stream cipher combines
the plaintext bits with a pseudorandom
cipher bit stream (key stream), typically by
an exclusive-or (XOR) operation. In a
stream cipher the plaintext digits are
encrypted one at a time, and the
transformation of successive digits varies
during the encryption. In practice, the digits
are typically single bits or bytes.
A stream cipher makes use of much smaller
and more convenient key— 128 bits, for
example. Based on this key, it generates a
pseudorandom key stream which can be
combined with the plaintext digits in a
similar fashion to the one-time pad.
However, this comes at a cost: because the
key stream is now pseudorandom, and not
truly random, the proof of security
associated with the one-time pad no longer
holds: it is quite possible for a stream cipher
to be completely insecure. Binary stream
ciphers are often constructed using PRNG
such as linear feedback shift register (LFSR)
or additive cellular automat (ACA) in
today‘s cryptographic scenario.
LFSRs can be implemented in hardware,
and this makes them useful in applications
that require very fast generation of a pseudo-
random sequence, such as direct-sequence
spread spectrum radio. LFSRs have also
been used for generating an approximation
of white noise in various programmable
sound generators.
An ACA is a cellular automaton whose rule
is compatible with an addition of states.
Typically, this addition is derived from
modular arithmetic. Additive rules allow the
evolution for different initial conditions to
be computed independently, then the results
combined by simply adding. The results for
arbitrary starting conditions can therefore be
computed very efficiently by convolving the
evolution of a single cell with an appropriate
convolution kernel (which, in the case of
two-color automata, would correspond to the
set of initially "active" cells).
Linear feedback shift register
A linear feedback shift register (LFSR) is a
shift register whose input bit is a linear
function of its previous state.
The only linear function of single bits is
XOR, thus it is a shift register whose input
CONFERENCE ON ―SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)‖ MARCH 26-27 2011
COMO1O1-2

bit is driven by the exclusive-or (XOR) of
some bits of the overall shift register value.
The initial value of the LFSR is called the
seed, and because the operation of the
register is deterministic, the stream of values
produced by the register is completely
determined by its current (or previous) state.
Likewise, because the register has a finite
number of possible states, it must eventually
enter a repeating cycle. However, an LFSR
with a well-chosen feedback function can
produce a sequence of bits which appears
random and which has a very long cycle.
Additive cellular automata
Cellular automata evolve in step and the
value of node depends on the value of
neighbors. An additive Automata (ACA)
consists of a collection of cells/nodes
formed by flip-flops which are logically
related to their nearest neighbors using
XOR/XNOR gates[5]. When the value of a
node is deter-mined only by neighboring
cells the ACA is known as one-dimensional
linear CA. The logical relations which relate
a node to its neighbors are known as rules
and they define the characteristics of an
ACA. There are many rules which can be
used to construct an ACA register, the most
popular being rules 90 and 150 illustrated in
figure .





= (Rule 90)








(Rule 150)

The next state (t+1) of the node X
i
is
determined by

the current state X (t) of
neighboring nodesX
i-1
and X
i+1
for rule 90
and nodes X
i
, X
i-1
and X
i+1
for rule 150. All
the nodes of a CA register do not have to be
implemented with the same rule, different
nodes can employ different rules. The first
and the last nodes of a CA register have only
one neighbor unlike all other nodes which
have two, hence normal rules cannot be
applied here [5]. One solution is to assume
that the missing neighbor is fixed at logic ‗0‘
(null boundary condition). The other
solution assumes the last and first nodes to
be neighbors and is connected using normal
rules (cyclic condition). Connection
between the end nodes (first and last nodes)
introduces a feedback loop in the cyclic
boundary condition; this makes null
boundary condition a better choice.
Comparison

Building on the results of Serra, et.al [3] the
consequences of the similarity
transformation between cellular automata
and LFSRs has been explored. By definition
the LFSR obtained by a similarity
transformation of the transition matrix of a
CA has the same characteristic polynomial
as the original CA. It has been shown that
the characteristic equation determines the
recursion relation among the bits in the
D Q
X
i-1

D Q
X
i


D Q
X
i+1
D Q
X
i-1

D Q
X
i


D Q
X
i+1
CONFERENCE ON ―SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)‖ MARCH 26-27 2011
COMO1O1-3

output bit sequence of a CA. This implies
that the same linear dependencies exist in
the output bit stream of a CA as in the
output of the similar LFSR.

LFSR and CA are characterized by their
transition matrices, the analysis of these
matrices along with simulations give the
measure of the randomness in the patterns
generated; these measures show the higher
randomness of patterns produced by CAs.
Parallel patterns generated by LFSRs (using
outputs from different nodes of an LFSR)
have a strong correlation between each other
due to the shifting of data. Pattern
generation in CAs does not involve shifting
of data.

There is greater probability of an error in
LFSR by aliasing compared to ACA due to
shifting of data in LFSR [1]. In case of CA
each node value is a function of the
neighboring nodes resulting in a lower
probability of an error. The presence of
XOR gates in the feedback path of an
External Feedback LFSR and lack of a
feedback path in a null boundary condition
results in higher operating speed for CAs.

LFSR have a feedback from their end nodes;
this means a redesign of the LFSR is needed
if the pattern length has to be changed. This
is not the case with ACA. ACA is logically
connected to their only to their neighbors
and there is no feed-back for an ACA
employing the null boundary condition [5].
Therefore, the pattern length generated by
CAs can be easily changed by cascading the
nodes. The regular structure of the nodes for
CA makes them ideal for CAD tools by
providing the much needed flexibility in
design. However, it is difficult to construct a
maximum length sequence CA as compared
to an LFSR which can be constructed using
the primitive polynomials which are very
well documented.
An LFSR can be implemented using only a
few XOR gates whereas a CA requires at
least one XOR gate for each node. This fact
brings up an obvious draw-back of CA:
Higher area overhead involved in
implementation of CA compared to an
LFSR. So, the designer has to pay a penalty
on the area over-head by choosing CA over
LFSR.

In ACA the communication is generally
local, being restricted to the nearest
neighboring cells and cells are regular and
topologically equivalent to one another but
in the case of LFSR these property does not
exist [2].

Following table shows the summary of
comparison
Characteristics LFSR ACA
Performance

Very good
in case of
internal
feedback,
poor for
external
feedback
Good- no
feedback
path and
maximum
on XOR
gate
between
node
Randomness of
generated
pattern
Low-
shifting of
bit causes
correlation
between
pattern
High – no
shifting of
bit
CAD Friendly No-
requires
redesign for
change in
pattern
length
Yes- node
can be
cascaded
easily
Speed Lower than
CA
Higher than
LFSR
Error
Probability
Greater- by
aliasing due
to shifting
Lower- no
shifting of
bit
CONFERENCE ON ―SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)‖ MARCH 26-27 2011
COMO1O1-4

of bit
Conclusion

LFSR has been well researched and provide
a compact and simpler circuit for application
using polynomial division and generation of
patterns. Their popularity is owed to simple
and compact design for cryptographic
application they find use in pseudo random
number generation. But research has shown
some of the shortfalls of the LFSR can be
overcome by the use of Cellular Automata.
ACA use larger nodes as compared to the
LFSR but provide much more random
patterns and can be easily cascaded for
design flexibility.

References

[1] M. Serra, T. Slater, J. C. Muzio & D. M.
Miller, ―The Analysis of One Dimensional
Linear Cellular Automata and Their
Aliasing Properties,‖ IEEE trans. On CAD,
pp. 767-778, 1990.

[2] Ph. Tsalides, T.A. York, A. Thanailakis,
―Pseudorandom number generators for VLSI
systems based on linear cellular automata,‖
IEE Proceedings-e, Vol. 138, no. 4, July
1991.

[3] M. Serra, D. M. Miller, J. C. Muzio,
―Linear cellular automata and LFSRs are
isomorphic,‖ Proc. Third tech. Workshop
on new dir. For IC Testing, Halifax N.S. Pp.
213-223, oct. 1988.

[4] Pino Caballero-Gil, Amparo Fúster-
Sabater, Oscar Delgado-Mohatar, ―Linear
Cellular Automata as Discrete Models for
Generating Cryptographic Sequences,‖
Journal Of Research And Practice In
Information Technology, Vol.40, No.4,
November 2008.

[5] S. Nandi, B. K. Kar, and P. Pal
Chaudhuri, ―Theory and Applications of
Cellular Automata in Cryptography,‖ IEEE
Transactions on Computers, vol. 43, no. 12,
December 1994.

[6] S. Zhangt, R. Byrnel, J.C. Mutiot, D.M.
Miller, ―Why Cellular Automata Are Better
Than LFSRs as Built-In Self-Test
Generators for Sequential-Type Faults,‖
IEEE International Symposium on Circuits
and Systems, Vol. 1, 1994.

[7] N. M. Thamrin, G. Witjaksono, A.
Nuruddin, M. S. Abdullah, ―An Enhanced
Hardware-based Hybrid Random Number
Generator for Cryptosystem,‖ International
Conference on Information Management
and Engineering, 2009.

[8] Paul H. Bardell, ―Analysis of Cellular
Automata Used as Pseudorandom Pattern
Generators,‖ IEEE International Test
Conference 1990.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O2-1


COST-EFFECTIVE MARKOV BASED SEGMENTATION
S.Robinson Sobitha Raj Mrs.V.S.Bindhu
M.E., Applied Electronics, Asst. Professor, EEE dept
Noorul Islam University, Kumaracoil. Noorul Islam University, Kumaracoil
robins.singh@yahoo.co.in bindhuvs2003@gmail.com
.
Abstract—Image segmentation is an important
preprocessing step in a sophisticated and complex
image processing algorithm. In this paper we are going
to introduce a procedure to minimize the
misclassification cost with class-dependent cost. This
procedure assumes the hidden Markov model (HMM)
which has been popularly used for image segmentation
in recent years. We represent all feasible HMM-based
segmenters (or classifiers) as a set of points in the
receiver operating characteristic (ROC) space. Then,
the optimal segmenter (or classifier) is found by
computing the tangential point between the iso-cost line
with given slope and the convex hull of the feasible set
in the ROC space. We illustrate the procedure by
segmenting aerial images with different selection of
misclassification costs.
Index terms --- Convex hull, hidden Markov models,
image segmentation, iso-cost line, ROC convex analysis,
ROC curve.

I.INTRODUCTION

Image segmentation extracts explicit information
about content, and it allows human observers to understand
images clearly by focusing specific regions of interest. For
this reason, it is often used as an initial procedure to
simplify a sophisticated and complex image processing
system. Segmentation often requires smooth boundary
between regions for different classes, and hidden Markov
model (HMM) is possibly one of the most popular models
for it [4], [8], [11]. The HMM assumes that the true hidden
class Markovian dependency and, thus, has smooth
boundary between segmented regions. The popularly used
Markov models have two parameters, which we denote by
and β, where indicates the popularity of each class over
images and the other β indicates the strength of spatial
coherence. The parameters are estimated in segmentation
procedures or pre-decided by an expert.
In real examples, the cost of misclassification can
depend on the classes. For example, in cancer diagnosis,
misclassifying tumor cells pays much higher cost than
doing normal cells to tumor cells; or, segmenting and
detecting military targets, mistakenly detecting targets to
non targets may cost more than the other type of
misclassification. However, all existing segmentation
procedures do not take into account the cost of
misclassification, particularly the unequal cost that depends
on the classes.
The main theme of this paper is to design a
classifier to minimize the expected cost of misclassification
with class- dependent misclassification cost. We denote it
as a cost-effective classifier. The cost-effective classifiers
studied much with the name “ROC convex analysis” in
machine learning literature [2], [3], [5]-[7] [9], [10], where
targets to be classified are often independent to each other.
However, it has rarely been studied in the frame work of
image segmentation, in which targets are dependent from
the Markovian model. In this paper, we introduce the ROC
convex analysis procedure for HMM-based image
segmentation to consider class-dependent misclassification
cost.
The ROC convex analysis draws great attention
in the machine learning society. In a two-class problem
(positive and negative class), the ROC curve (or set) is the
plot of the probability of false positive decision (false
positive rate, FPR) and that of true positive decision (true
positive rate, TPR). The ROC convex hull analysis finds an
optimum point in an ROC space to minimize the
misclassification cost of classifier τ, which is defined as
(1)
Where implies false negative rate which is equal
to .Once misclassification cost is given, we
have a family of iso-cost lines with slope

On which costs of classifiers are the same. Then, the ROC
convex analysis finds the optimal classifier as the classifier
whose (FPR, TPR) pair is the tangential point between
these iso - cost line and the convex hull of the ROC curve.
The ROC convex analysis requires the entire set of feasible
classifiers. In the HMM based segmentation, the model has
two unknown parameters which are estimated in
segmenting the image. Fixing these two parameters, not
estimating, provides the set of classifier C (γ, β) s which are
indexed by . The ROC set is the set of error rates
s of classifier . We denote its convex
hull as .
The set should be characterized to find the
most cost effective classifier, but it s not computationally
feasible. In this paper, we provide an alternative way which
does not compute the entire set , but we approximate its
boundary. We find that the tangential point between the iso
– cost line with given slope and exists on the boundary
of , which we denote as . Also, we find that is the
convex hull of the boundary of . Hence, we suggest
to compute or approximate . To approximate the
boundary of , we fix the parameter γ but estimate β in
segmenting the target images. We compute
the of classifiers in s.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O2-2

We apply the proposed procedure to segment
aerial images with two classes, “man- made region” and
“natural region”, which could be used for target recognition
and tracking. Here, the cost of misclassifying targets (or
man-made regions) is higher than that of misclassifying
non targets (or natural regions) to targets. The results show
that, as the misclassification cost of man-made regions
increases, the HMM based segmenter tends to classify both
man-made and natural regions to man-made ones. Thus, the
rate of falsely classifying into man-made regions increases.
The remainder of this paper is organized as follows. We
introduce ROC convex analysis in section II. We study its
application to HMM based segmentation procedure in
section III. In section IV, We apply the proposed
procedure to segment an aerial image. Section V concludes
this paper.

II. ROC CONVEX ANALYSIS

In a binary classifier, the ROC curve (or space)
plots two accuracy measures of a classifier, FPR and TPR.
Suppose we use a continuous classification score X to
diagnose a certain disease, and we classify a subject into a
disease (positive) group if his/her score X is higher than a
given critical point ; otherwise, we classify him/her into
non disease group (negative) group. The ROC curve plots a
series of (FPR, TPR) pairs produced from different choices
of .
The optimal ROC curve is the one produced by
the classifiers which has the maximum TPR given FPR.
The optimal ROC curve has several geometrical properties
including convexity. Suppose it is not convex on an
interval , where a and b correspond to critical
values and in the way that the FPR at and is a
and b, respectively. Then, for the diagnosis with critical
values , we find a better diagnostic system
(classifier) which has the same FPR but higher TPR by
randomly choosing between two diagnoses with critical
values and . Thus, the convex hull of the observed
ROC curves represents the ROC curve of the set of
potentially optimal classifiers. We let in ROC space be
the set of s of all classifiers we consider, and
be the set of their random mixtures. Then, is the
smallest convex region which contains .
The goal of this paper is to find the classifier
which minimizes he cost (1) among classifiers in .
Since ,

– (2)

We represent the cost functions s in the
ROC space. Then, the intercept of the iso-cost line (2) is
– and, as in Fig. 2.1, the point to minimize the
cost is the tangential one between the line with slope and
the convex area .
In summary, proposed procedure to find the cost-
effective classifier has the following steps:
S1) we define a set of classifiers, say C, to be considered.
In case of HMM-based segmentation, this becomes a set of
classifiers corresponding to each of two parameters Ising
model prior. We will see details on HMM-based
segmentation in the next section.
S2) we compute FPRs and TPRs of classifiers in C and plot
them. This defines the region in the ROC space. Further,
we get the ROC convex hull region of .
S3) we find the tangential point between the line with slope
in (2) and the ROC convex hull region . The classifier
corresponding to the found tangential point is the most
cost-effective classifier.

Fig.2.1. Illustration of ROC convex hull analysis

III. COST MINIMIZATION IN HMM-BASED
IMAGE SEGMENTATION

In this section, we find the HMM-based classifier
to minimize the (expected) cost of misclassification (1)
(ECM).
To get better understanding of the problem, we
begin with HMM without spatial coherence such as GMM
[1]. The model assumes that the testing image is composed
of many independent sub-blocks, say X s for k = 1, 2, …,
K. The model has unknown parameter γ which indicates the
prevalence of each class. We let γ be the ratio of prior
probability of positive class (class P) to that of negative
class (class N). Then, given γ, the optimal classifier to
minimize the ECM among all classifiers is the maximum a
posteriori (MAP) classifier that assigns x to class N, for
each k, if

(3)

Where fn (x) and fp(x) is the probability density
function of class N and class P, respectively. In GMM, both
fn(x) and fp(x) are density functions of the GMM.
Suppose we denote the classifier in (3) given γ as C (γ), and
their collection as F. We further find that this collection D
is invariant to the cost ∞. In other words, we always have
same collection regardless of what ∞ we choose.
In practice, to find the optimal classifier, we set to be an
arbitrary fixed number, and we consider the set of
conditionally optimal classifiers which minimize the cost of
misclassification given γ > 0. We let D be the set of their
FPRs and TPRs in the ROC space. Following the steps in
the previous section, we get D and find the tangent point
and γ between the line (2) and D.
Now, we move to the HMM-based segmenter.
The HMM based segmenter assumes that the hidden
process is from the Markov random field having two
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O2-3

unknown parameters γ and β; the parameter γ is the
parameter that represents the magnitude of magnetization
of the random field which implies the dominance of class P
against class N in common words; it is also related to the
ratio of the prior probability of P to N; β is the parameter to
measure the strength of spatial coherence. For example, the
HMGMM model [11] uses the generalized bond
percolation (BP) model. Let Z = {} with Z = – 1 or 1: Z = 1
and Z = – 1 implies the class P and N, respectively. The
generalized BP assumes that the probability of Z is
(4)
Where (Z) ( (Z)) is the number of concordant (dis-
concordant) adjacent pairs which are neighbors to each
other. Here, ) is the partition function that is



Fig.3.1. ROC convex Hull for HMGMM Classifiers

As in the spatially uncorrelated model, we consider the
collection F of MAP classifier C , which, given γ and
β, assigns the observed image i,j = 1, 2, …, n} to
. We let C be the MAP
classifier, given γ and β, and F be the collection
of . We further define D and similarly. To find
the optimal classifier, again, we compute the tangential
point between the line (2) and , the convex hull of D in the
ROC space.
In HMM-based segmenter, evaluation of D is
computationally quite intensive in practice. Here, we
introduce a suboptimal but fast algorithm for it. Many
algorithms are studied from deterministic annealing to
simulated annealing to Markov chain Monte Carlo method
to find the MAP. They often assume γ=0 and estimate β,
the parameter of spatial coherence in finding the MAP. For
example, in [11], the parameter β is estimated using the
maximum likelihood estimate (MLE) along with the Gibbs
sampler. The MLE is (Z)|X}/total no of
edges), and we approximate the right hand side of the
equation using Monte Carlo method to get the estimate.
We let be the estimate of β given γ, and we
approximate the boundary points of with FPRs and TPRs
of by moving γ. As stated before, knowing
boundary is sufficient to find the tangential point
between iso-cost lines and . The approximated boundary,
denoted by , is a curve from (0, 0) to (1, 1). Finally,
is approximated by the convex hull of , and the optimal
classifier is found using the procedure in Section II.

IV. EXPERIMENTAL RESULTS
In this section, we apply the procedure in Section
III to segment the aerial image with HMGMM with
generalized BP model in (4). The aerial image is composed
of many sub-blocks which are classified into “natural”
regions or “man-made” regions. We call the “natural” sub-
block as “negative” and “man-made” one as “positive.”
The HMGMM model has two parameters γ and β which
reflects the overall portion of man-made and spatial
coherence between adjacent sub-blocks, respectively. Each
introduces a classifier, say , and a point of
(FPR, TPR) in the ROC space. We let be D the collection
of all these s. The input image taken is
illustrated below.


Fig.4.1. Input image

In the experiment, we vary from 0.35 to
2.85 by 0.1, and we get the empirical FPRs and TPRs from
the segmented results . We then compute the convex hull
of ( s of s to approximate the convex
hull of D. We denote the convex envelope as . Fig. 2
plots s and their convex envelope in the ROC
space. We find the optimal classifier to minimize the
misclassification cost (1) is the one that corresponds to the
tangential point between the (2) and .The is
piecewise linear function with 11 supporting points (points
that have changes in slope). Each point on is the
optimal classifier to minimize the cost for a specific choice
of .


Fig.4.2. Segmented image

Now, we report some details of the analysis for = 0.5,1,
and 2. The noise mixed image is given by,
segmented image
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O2-4


. Fig.4.3. Noised image

Here, the high value of implies high cost of
misclassification of “manmade” sub-blocks (to “natural”
ones). Thus, as .increases, the segmenter tends to classify
both “natural” and “man-made” blocks to “man-made”
ones. The enhanced image is given by,

Fig.4.4. Enhanced image

The analysis graph is given below:


Fig.4.5. Analysis graph

Finally, the original HMGMM in [11] is the classifier
with . Its FPR and TPR is 0.22 and 0.97,
respectively, which is close to the (approximated) boundary
of . Some computation shows that the HMGMM is close
to the optimal for between 0.16 and 0.22.



V. CONCLUSION

We propose a procedure to find the optimal
classifier to minimize the misclassification cost based on
HMM. The cost function to minimize allows unequal cost
that depends on the true class. The optimal classifier is
given as a classifier whose error rate pair (FPR, TPR) is the
tangential point between the iso-cost line (2) and the
convex hull of feasible set of error rates by HMM-based
segmenters. We apply the procedure to segment an aerial
image with different selection of unequal misclassification
cost.
The procedure in this paper does not depend on a
specific type of image but is applicable to a wider class of
images and HMM-based classifiers for them. The same
procedure can be extended to other applications such as
cancer diagnosis based on MRI imaging or 3-D temporal
images. A possible difficulty in these extension is from
obtaining , the boundary of the ROC convex space. The
current suboptimal algorithm could be more exhaustive as
the number of HMM parameters increases, and it might
provide rough approximation if the size of training data sets
is not enough. This would be our next step to move.

VI.REFERENCES
[1] A. Aiyer, K. Pyun, Y. Huang, D. B. O’Brien, and R. M.
Gray, “Lloyd clustering of Gauss mixture models for image
compression and classification,” Signal Process.: Image
Commun., vol. 20, pp. 459–485, Jun. 2005.
[2] H. Blockeel and J. Struyf, “Deriving biased classifiers
for improved ROC performance,” Informatica, vol. 26, pp.
77–84, 2002.
[3] R. Caruana and A. Niculesen-Mizil, “An empirical
comparison of supervised learning algorithms,” in Proc.
23rd Int. Conf. Machine Learning, 2006, pp. 161–168.
[4] P. A. Devijver, “Segmentation of binary images using
third order Markov mesh image models,” in Proc. 8th Int.
Conf. Pattern Recognition,
1986, pp. 259–261.
[6] P. A. Flach and S. Wu, “Repairing concavities in roc
curves,” in Proc. 19th Int. Joint Conf. Artificial
Intelligence, 2005, pp. 702–707.
[7] D. M. Gavrila, J. Giebel, and S. Munder, “Vision-based
pedestrian detection: The protector system,” in Proc. IEEE
Intelligent Vehicles Symp., 2004, vol. 13–18.
[8] J. Li, A. Najmi, and R. M. Gray, “Image classification
by a two dimensional hidden Markov model,” in Proc. Int.
Conf. Acoustics, Speech, and Signal Processing, 1999, pp.
3313–3316.
[9] R. C. Prati and P. A. Flach, “Roccer: An algorithm for
rule learning based on ROC analysis,” in Proc. 19th Int.
Joint Conf. Artificial Interlligence, 2005, pp. 823–828.
[10] F. Provost and T. Fawcett, “Robust classification for
imprecise environments,” Mach. Learn., vol. 42, pp. 203–
231, 2001.
[11] K. Pyun, J. Lim, C. S. Won, and R. M. Gray, “Image
segmentation using hidden Markov Gauss mixture
models,” IEEE Trans. Image Process., vol. 16, no. 7, pp.
1902–1911, Jul. 2007.
noised image
enhanced image
Performance Comparison of VANET Routing Protocols
NITIN SHARMA
Department of Computer Science and Engineering
MNIT, Allahabad, India
nsharma283@gmail.com
Mobile: +91 8861 6365 11
POOJA RANI
Department of Information Technology
ITM, Sector-23A, Gurgaon, India
me_pooja@msn.com
Mobile: +91 9899 7470 33
ABSTRACT
Vehicular Ad-hoc Network (VANET) is a
collection of communication vehicles, moving
in different directions. The vehicles form a
communication group to disseminate desired
information. Various routing protocols are
implemented in a VANET, each having benefits
and shortcoming in the domain of
implementation. In this paper performance of
three routing protocols, namely Ad hoc On-
Demand Distance Vector Routing (AODV),
Destination Sequenced Distance Vector (DSDV)
and Dynamic Source Routing (DSR) is
compared for different parameters. The
protocols are simulated on Network Simulator-
2(ns-2).
Index Terms
Routing protocols, Simulation, VANET, Vehicle-
to-Vehicle Communication, WAVE.
I - INTRODUCTION
VANET is an ad-hoc network formed between
vehicles as per their need of communication. In
order to develop a VANET every participating
vehicle must be capable of transmitting and
receiving wireless signals up to range of three
hundred meters. VANET communication range is
restricted up to one thousand meters in various
implementation [1]. The performance of a
VANET remains optimum within one thousand
meters and beyond that it is not feasible to
communicate among vehicles because of high
packet loss rate [2],[3]. Finding optimum path is
typical task for dynamic protocols as
management of vehicle movement is quite
complex. There is a need to update entries in the
route finding node [4]. VANET is not restricted
up to Vehicle-to-Vehicle communication, it takes
benefits of road side infrastructure that can also
participate in communication between vehicle
[5], but in this paper the main focus is on
Vehicle-to-Vehicle communication.
There are various challenges for VANET
such as high speed of vehicle, dynamic route
finding, building, reflecting objects, roadside
objects, other obstacles in path of radio
communication, different direction of vehicles,
concern about privacy, authorization of vehicle,
security of data and sharing of multimedia
services. High speed of vehicle requires regular
update of routing table whereas dynamic route
finding would result into high time loss before
static communication [6].
Various user group among VANET are
getting popular, mostly it is used in traffic
management agencies, highway safety agencies,
law enforcement agencies and emergency
services.
The main service provided through
VANET are GPS navigation system, electronic
payment of toll tax, authenticity of vehicle
without human intervention, traffic message.
Internet access, broadcasting of traffic scenario
and multimedia streaming.
As most of the properties of a Mobile Ad-
hoc Network (MANET) are common with the
VANET, various MANET routing protocols are
used in VANET [7]. Basic difference between
MANET and VANET is that under VANET
PERFORMANCE COMPARISON OF VANET ROUTING PROTOCOLS 1
movement of vehicle is at high speed and less
random as compared to MANET, existing
MANET routing protocols are not compatible
with VANET. VANET routing protocols are
broadly divided into two categories Table Driven
protocols and Source Initiated on Demand
protocol. In Table Driven protocol each vehicle
maintains a table of neighborhood vehicles
within its communication range and any change
in vehicle position is updated regularly. In Source
Initiated on Demand protocol, firstly source
vehicle broadcasts a query to find route up to
node gets a route up to the destination node. The
destination route is replied back to the source
node via same path. In this paper, the
performance of following three routing protocols
are compared:
Ad Hoc on-Demand Distance Vector Routing (AODV)
It is a Source Initiated on Demand routing
protocols used in VANET. In this protocol every
vehicle maintains a route information of every
vehicle. It uses sequence number concept to
acknowledge the entry update time and time
stamp based concept for table entry. If a table
entry is not used within a certain time limit, it
will be deleted from table and if there is any
breakage in linking with a vehicle to another
vehicle, route error (RERR) packet is forwarded
so that vehicle route is effectively updated in the
routing table.
Destination Sequenced Distance Vector (DSDV)
It is Table Driven routing protocol which is used
in VANET and is based on classical Bellman-
Ford algorithm. Initially every vehicle broadcasts
its own route tables to its neighbor vehicles. The
neighbor vehicles update routing table with the
help of two type of packets- Full Dump Packet
and Incremental Packet. Full Dump Packet
contains information about every participating
vehicle in the VANET. These packets are
transmitted periodically after a long time interval.
Incremental Packet contains updated change in
vehicle position since last Full Dump Packet.
These packets are transmitted periodically in
short interval of time and are stored in additional
table. Routes are selected with the latest entry in
the table. DSDV is good for networks where
location of nodes are less dynamic. If position of
a vehicle changes very often, its performance
goes down because more Full Dump Packets are
needed to sent in the network, resulting into
wastage of bandwidth.
Dynamic Source Routing (DSR)
It is Source Initiated on Demand routing protocol
used in VANET and is based on link state routing
algorithm. When a vehicle wants to communicate
data to another vehicle, firstly it finds route up to
that vehicle. For route discovery, source vehicle
initiates a route request (RREQ) packet in the
network and other nodes forward the RREQ by
changing their name as sender. Finally when
RREQ packet reaches to the destination vehicle
or to a vehicle having path to the destination
vehicle, a route reply (RREP) packet is unicasted
to the sender node. If the reply is not received,
the source vehicle restarts aggressive discovery
of route up to the destination vehicle.
The simulation setup is discussed in the next
section and results are presented in section III.
Section IV describes conclusion and future scope.
II - SIMULATION SETUP
The Protocol stack of VANET consists of two
combination of standards Dedicated Short Range
Communication (DSRC) and Wireless Access in
Vehicular Environment (WAVE). DSRC
contains three layer – Physical Layer, Media
Access Control (MAC) Layer, and Logical Link
Control (LLC) Layer. The layers make
communication possible in wireless environment.
WAVE lies on top of the DSRC layer. WAVE is
also known as IEEE P1609 and is used as a
standard for communication, which is further
classified as follows:
• IEEE P1609.1: It is standard for resource
manager, which defines the key resources for
data-flow and other resources as well as the
command messages.
PERFORMANCE COMPARISON OF VANET ROUTING PROTOCOLS 2
• IEEE P1609.2: It is standard for security
services in application and controlling
messages, encoding and decoding of
messages.
• IEEE P1609.3: It is the standard for network
services and involves the key features of
network and transport layer such as IP
addressing, information passing and security
while switching services.
• IEEE P1609.4: It is the standard for MAC
layer support and increases communication
capacity of node. Also it provides
multichannel operation to maintain Quality of
Service (QoS) with high speed.
The ns-2 is used to evaluate performance of
routing protocols in VANET. It is a network
simulator and is also widely used for VANET
related simulation work [8]. Ns-2 provides a
standard for IEEE 802.11p simulation in the form
of tcl scripts. Also various types of
communication patterns, traffic scenarios and
resources are available with ns-2.
The simulation grid is 500m x 500m. in which
the initial position of every vehicle is specified at
the starting time of simulation by calling C++
object and the destination of every vehicle is set
after a certain time interval. The vehicle approach
their destination with variable speed. In
simulation, three input parameters are provided
namely, speed of vehicle vary between 10MPH to
100MPH, number of vehicles vary between 20 to
100 and distance between vehicles very between
20m to 100m. for the traffic purpose, Constant
Bit Rate (CBR) traffic with a fixed packet size is
used. The vehicle antenna is omni directional and
its transmission range is 150m. The channel data
rate is 2Mbps. IEEE 802.11p standard is used for
vehicle-to-vehicle communication, which is
available is ns-2 are wireless support [9]. It
follows most of the WAVE standard, Nakagami
model is used for radio propagation, which is the
best model for WAVE environment [10]. While
implementation a routing protocol in ns-2 the
behavior of queue and its adaptability criteria are
carefully considered.
In the simulation work, the available
standard and data units are followed. Ns-2
produces output in the form of trace file which is
further processed by shell scripting to calculate
the desired parameter. Shell scripting and awk
are widely used to process trace files.
III - SIMULATION RESULT
The comparison of routing protocols are done on
the basis of following parameters:
Average Number of Hop Count:
It is the number of vehicles running between
source and destination and it signifies error in the
network. Time to Live (TTL) is decided on the
basis of Hop Count, which helps in avoiding the
congestion in the network.
Average Jitter Rate:
It is the delay between two consecutive packet
delivery at a node. Quality of Service (QoS) of
the network is measured by Average Jitter Rate.
Packet Delivery Ration:
The fractional of total packets received at a node
to the packets sent by the node. It is associated
with the QoS and bandwidth utilization in the
network.
Routing Overhead:
It is the ration of number of MAC packets and
total number of packets sent, Routing Overhead
increases with increases in vehicle density.
Throughput:
Throughput is the sum of data to all the nodes in
the system during a period. In a time interval the
throughput reflects the bandwidth utilization.
For each of the above parameters, various traffic
scenarios are simulated by changing the number
of vehicles, distance between vehicles and
vehicle speed. The effect of these changes in the
routing protocols are analyzed and results are
shown below.
PERFORMANCE COMPARISON OF VANET ROUTING PROTOCOLS 3

Fig 1: Jitter Rate with various vehicle distances

Fig 2: Packet Delivery Ratio with various vehicle distances
Figure 1 shows the Jitter Rate variation at
different distances between vehicles. It is
observed that DSR Jitter Rate remains lowest
throughput the interval of observation, while
DSDV and AODV show comparatively higher
Jitter rate.
Figure 2 shows the Packet Delivery Ratio at
various distances between vehicles. It is observed
that DSR Packet Delivery Ratio remains high
during entire duration of observation. AODV
Packet Delivery Ratio in between other two and
DSDV Packet Delivery Ratio is lowest among
three.
Fig 3: Routing Overhead with various vehicle distances
Fig 4: Throughput with various vehicle distances
Figure 3 shows the Routing Overhead at various
distances between vehicles. It is observed that
DSR Routing Overhead is lowest, AODV and
DSDV Routing Overhead remain similar from
20m to 100m but after 100m AODV Routing
Overhead is going down.
Figure 4 shows the Throughput with various
vehicle distances is shown. Clear inference from
graph DSR Throughput is highest. AODV
Throughput remains in between other two and
DSDV Throughput is lowest among all three.
PERFORMANCE COMPARISON OF VANET ROUTING PROTOCOLS 4
In the simulation work, we compared the
performance of AODV, DSDV and DSR on
fifteen criteria as shown in Table I. It is observed
that no protocol performs good in all the
conditions. One protocol performs well in
specific domain area and others in different
domain area.
DSR had better performance than other two
protocols because of its dynamic route finding
nature that decreases Hop Count as every time it
increases portability of finding optimum route. It
also increases Packet Delivery Ratio. Routing
Overhead is also decrease because in dynamic
network it is found that most of the maintain any
routing table that benefits in Hop Count, it is
increase Throughput of network by reducing
regular update packet.
TABLE I
PARAMETER COMPARISION OF ROUTING PROTCOLS
Parameters Variables Routing Protocols
AOD
V
DSDV DSR
HOP COUNT
Dist. b/w Vehicles L M H
HOP COUNT
No of Vehicles L H M
HOP COUNT
Speed of Vehicles L M H
Jitter Rate Dist. b/w Vehicles M H L
Jitter Rate No of Vehicles M H L
Jitter Rate Speed of Vehicles M H L
Packet Del Ratio
Dist. b/w Vehicles M L H
Packet Del Ratio
No of Vehicles H L M
Packet Del Ratio
Speed of Vehicles H L M
Routing
Overhead
Dist. b/w Vehicles H M L
Routing
Overhead
No of Vehicles H M L
Routing
Overhead
Speed of Vehicles H M L
Throughput Dist. b/w Vehicles M L H
Throughput No of Vehicles H L M
Throughput Speed of Vehicles H L M
Abbreviation: L- Low Value, M- Medium Value & H- High Value
AODV performs better than DSDV in most of the
parameters because of its time limit in table entry
usages. That benefits in Throughput, Packet
Delivery Ration and some times that also benefits
in Throughput Routing Overhead. DSDV
Performance is worst among three protocols, it
uses updation of table and depends on table the
unnecessary decreases Packet Delivery Ratio,
increases number of Hop Counts and decreases
Throughput.
IV - CONCLUSTION AND FUTURE WORK
On the basis of simulation results presented in
Table I for different traffic scenario and for three
protocols, performance of DSR has been found to
be better than that of AODV and DSDV. The
performance results are summarized for
Highway, Urban and Freeway traffic scenarios in
Table II.
From the parameter values characterizing the
three traffic scenarios. DSR is found suitable for
Highway and Freeway traffics, whereas AODV is
suitable for Urban traffic scenario.
Thus it can be concluded that a single protocol
doesn’t give best performance in all traffic
scenarios. Since traffic scenarios change
throughput the day, some part of hybrid adaptive
protocol would give better performance.
TABLE II
DIFFERENT TRAFFIC SCENARIO
Traffic Scenario Freeway Highway Urban
Distance b/w Vehicles Small Large small
Density of Vehicles Low Low High
High Speed of Vehicles No Yes No
Suggested Protocol DSR DSR AODV
PERFORMANCE COMPARISON OF VANET ROUTING PROTOCOLS 5
REFERENCES
[1] Yi wang. Akram Ahmad, Bhaskar Krishnamachari
and Konstantinos Posunis, IEEE 802.11p Performance
Evaluation and Protocol Enhancement. IEEE
International Conference on Vehicular Electronics and
Safety, 2008.
[2] Bo Xu, Aris Ouksel and Ouri Wolfson,
Opportunistic Resouce Exchange in Inter-Vehicle Ad-
hoc Networks, International Conference on Mobile
Data Management, 2004.
[3] Lin Yand, Jindua Guo and Ying Wu, Piggyback
Cooperative Repetition for Reliable Broadcasting of
Safety Message in VANET’s IEEE Internatitional
Conference on Consumer Communication and
Networking, 2009, pp-1-5
[4] Katrin Bilstrup, Elisabeth Uhlemann, Erik
G.Strom, and Urban Bilstrup, On the Ability of the
802.11p MAC Method and STDMA to support Real-
Time Vehicle-to-Vehicle Communication. EURASIP
Journel on Wireless Communication and Networking
2009.
[5] Rainer Baumann, Simon Heimilicher and Martin
May, Towards Realistic Mobility Models for Vehicular
Ad-hoc Networks. IEEE International Conference of
Mobility Networking for Vehicular Environment,
2007.
[6] Valery Naumov, Rainder Baumann and Thomas
Gross, An Evaluation of Inter-Vehicle Ad-hoc
Networking Based on Realistic Vehicular races, ACM
International Symposium on Mobile AdHoc
Networking and Computing, 2006.
[7] D. Rajini Girinath and Dr. S. Selvan, Data
Dissemination to regulate vehicular traffic using
HVRP in urban mobility model, International Journel
of Recent Trends in Engineering, 2009
[8] Victor Cabrea, Francisco J. Ros and Pedro M.
Ruiz, Simulation based Study of common Issues on
VANET Routing Protocols, IEEE Vehicular
Technology Conference, 2009.
[9] Djamel Djenouri, Wassim Soualhi and Elmalik
Nekka, VANET’s Mobility and Overtaking: An
Overview, International Conference on Information
and Communication Technologies, 2008, pp1-6
[10] Christoph Sommer, Isbel Dietrich and Falko
Dressler, Realistic Simulation of Network Protocols in
VANET Scenarios. International Conference on Mobile
Networking for Vehicular Environments, 2007
PERFORMANCE COMPARISON OF VANET ROUTING PROTOCOLS 6
Greedy Grid Scheduling Algorithm in Static Environment
POOJA RANI
Department of Information Technology
ITM, Sector-23 A, Gurgaon, India
E-mail: me_pooja@msn.com
Mobile: +91 9899 7470 33
NITIN SHARMA
Department of Computer Science & Engineering
MNNIT, Allahbad, India
E-mail: cs0802@mnnit.ac.in
Mobile: +91 8861 6365 11
Abstract— Grid scheduling is a technique by which the
user demands are met and the resources are efficiently
utilized. The scheduling algorithms are used to minimize
the jobs waiting time and completion time. Most of the
minimization algorithms are implemented in homogeneous
resource environment. In this paper the presented
algorithm minimize average turnaround time in
heterogeneous resource environment. This algorithm is
based on greedy approach which is used in static job
submission environment where all the jobs are submitted
at same time. Taken all jobs independent the turnaround
time of each job is minimized to minimize the average
turnaround time of all submitted jobs.
Keywords- greedy, grid, heterogeneous, high performance
computing, scheduling.
I. INTRODUCTION
Using the distributed resources to solve the applications
involving large volume of data is known as grid computing
[1], [2]. There exists many tools to submit jobs on the
resources which have different computational power and are
connected via Local Area Network (LAN) or Virtual Private
Network (VPN). The main challenge in grid computing is
efficient resource utilization and minimization of turnaround
time. The existing system model consists of the web based
grid network platform with different management policies,
forming a heterogeneous system where the computing cost
and computing performance become significant at each node
[3], [4].
In grid computing environment, applications are submitted for
use of grid resources by users from their terminals. The
resources include computing power, communication power
and storage. An application consists of number of jobs; users
want to execute these jobs in an efficient manner [5]. There
are two possibilities of submission of jobs/data on resources;
in one of them, job is submitted on the resources where the
input data is available and in the other, on the basis of specific
criteria, resource is selected on which both job and input data
are transferred. This paper uses second approach, wherein the
job is submitted on a scheduler and data on a resource
identified by the scheduler. A resource in existing algorithms
is selected randomly, sequentially or according to its
processing power [2], [6], [7]. In this paper the proposed
algorithm chooses a resource on the basis of processing
power, job requirement and time to start at that resource.
The next section describes details about system model.
Section 3 describes the proposed scheduling algorithm in
static job submission environment. In section 4 the
experimental details and the results of experiments are
presented with comparison of some existing algorithms. In
section 5 conclusions and suggestions for future
improvements are proposed.
II. SYSTEM MODEL
A grid is considered as the combination of multiple layers. In
our model the whole system is composed of three layers
(Fig.1). The first layer is the user application layer in which
the user authentication is done and jobs are submitted to the
scheduler by the user. The second layer contains scheduler
and GIS. The scheduler schedules jobs among various
resources after taking resource status information from GIS.
The second layer is connected through a VPN to user. VPN
provides additional security and only authorized users can
access services. All the resources reside in third layer where
user's jobs are executed which are also connected through
VPN.
Fig. 1: Layered architecture.
III. PROPOSED ALGORITHM
The existing grid scheduling algorithms are based on the
speed of resources [6], [7]. Each resource of layer 3 (Figure.
1) has different processing power and all the resources of
layer 3 are connected via homogeneous communication
environment in which the communication delay between
1
scheduler and resources is assumed constant, also the jobs are
assumed to submitted on layer 1 having different job
requirement.
An algorithm is proposed in this paper which is suitable for
static job submission in heterogeneous resource environment
connected to the scheduler through homogeneous
communication environment. Greedy approach is used to
solve the job scheduling problem. According to the greedy
approach “A greedy algorithm always makes the choice that
looks best at that moment. That is, it makes a locally optimal
choice in the hope that this choice will lead to a globally
optimal solution" [8]. The proposed algorithm uses the similar
approach; it takes every job as independent of each other and
each of them is scheduled on a resource to give minimum
turnaround time for that job.The overall turnaround time of all
the jobs is thus minimized. The parameters used in this
algorithm are as follows:
A set of resources, R = {R
1
, R
2
, R
3
,........, R
n
}.
Ji = The submitted i
th
job.
Arr_timei = Arrival time of job J
i
.
Proc_powerj = Processing power of resource R
j
.
Strt_timej = Estimated time at which a job starts execution at
resource R
j
.
Job_reqi = Length of job J
i
.
Schd_valueij = Expected turnaround time of i
th
job at j
th
resource.
Min = The minimum of Schd_value
ij
among all resources.
Res_id = Current selected resource id having optimum
turnaround time.
The algorithm used to schedule a job is given as follows:
GREEDY_SCHEDULE
/*The users submit their jobs on the scheduler.*/
For all resource R
j

/*Initialize the start time at resources.*/
Strt_time
j
=0.0
End For
/*The jobs are stored in a queue Q.*/
Insert all the jobs J
i
in Q
While Q is not empty
do
Delete the job J
i
SUBMIT_NEW_JOB
UPDATE_STATUS
Advance the Q pointer
End While
End GREEDY_SCHEDULE
The scheduler uses SUBMIT_NEW_JOB algorithm to find
the best suited resource that minimizes the turnaround time.
The turnaround time is calculated on the basis of expected
completion time of a job. The detailed SUBMIT_NEW_JOB
algorithm is as follows:
SUBMIT_JOB
Min = ∞
For every resource R
j
/* Calculating the expected turnaround time*/
Schd_value
ij
= Strt_time
j
+ (Job_req
i
/Proc_power
j
)
If Min is greater than Schd_value
ij

Then
Min = Schd_value
ij
Res_id = R
j
End If
End For
Submit the job J
i
to Res_id resource
Submit the input data of J
i
job to Res_id resource
End SUBMIT_NEW_JOB
Once the scheduler submits a job to a resource, the resource
will remains for some time in processing of that job. The
UPDATE_STATUS algorithm is used to find out when the
resource will be available to process a new job. The
UPDATE_STATUS algorithm is given below:
UPDATE_STATUS
/* Res_id is the resource on which the job Ji is submitted. j
is the index of resource on which the job Ji is submitted
and Rj = Res_id*/
Strt_time
j
= Schd_value
ij
End UPDATE_STATUS
The above presented algorithm has the time complexity of
O(n) for each job, where n is the number of resources. The
above algorithm required additional space to store the resorces
current status for availability.
IV. EXPERIMENTAL RESULTS
The GridSim simulator [6] is used to simulate the algorithms.
The GridSim toolkit is used to simulate heterogeneous
resource environment and the communication environment.
The experiments are performed with three algorithms. The
algorithms are Random Resource Selection and Equal Job
Distribution and Proposed Algorithm. The input data is taken
to be the same for all the three algorithms. The simulation is
conducted with three resources which are shown in Table 1.
TABLE 1: Resources with their architecture and processing power.
Resource R0 R1 R2
Architecture Sun Ultra Sun Ultra Sun Ultra
OS Unix AIX Unix
Proc_power(in MIPS) 48000 43000 54000
The scheduler submits these jobs on resources according to
these algorithms. The algorithms are presented one by one
with their simulation results.
A. Random Resource Selection
In this algorithm the scheduler contacts GIS to obtain the
resource information and then it chooses a resource randomly
[7]. The job is submitted on this chosen resource. This
algorithm is very simple to implement and has less overhead
on the scheduler. The bar chart (Fig. 2) shows the turnaround
time of different jobs. The completion time is a time at which
2
the result of a job is available. After simulation the average
turnaround time is found to be 20105.65 seconds and all the
jobs are completed at the 64420.25
th
second.
Fig. 2: Jobs and turnaround time using Random
Resource Selection.
B. Equal Job Distribution
In Equal Job Distribution we firstly calculate the total length
of all the jobs and
then distribute these length equally on every resource. The
main notations which are used in the formula are as follows:
L = Total length of all the jobs taken together.
Proc poweri = Processing power of resource R
j
.
tProc power = Total processing power of all resources.
Loadi = Load assigned on resource R
j
.
The formula used to calculate the job distribution is given
bellow:
Loadi = L* (Proc poweri /tProc power)
The turnaround time of each job is shown by the bar chart in
Fig. 3. Experimental results show that the average turnaround
time is 17968.55 seconds and the last result is outputed at
39000.22
th
second. Equal Job Distribution reduces the average
turnaround time by 10.62% and it takes less time in
comparison to the Random Resource Selection to give all the
results.
Fig 3: Jobs and turnaround time using Equal Job
Distribution.
C. Proposed Algorithm
In Proposed Algorithm, the scheduler finds the resource
information with the help of GIS and calculates the
approximate completion time of this job on every resource.
Using these values the scheduler chooses a resource which has
the minimum of completion time and submits that job on this
resource. The turnaround time of each job is shown in bar
chart in Fig. 4. Through this algorithm the average turnaround
time of these jobs is 17208.77 seconds and all the jobs are
completed at 41840.88
th
second. The Proposed Algorithm
further reduces the average turnaround time by 4.22% as
compared with Equal Job Distribution. The completion time
of all jobs takes some more time than Equal Job Distribution
algorithm.
Fig. 4: Jobs and turnaround time using Proposed Algorithm.
V. CONCLUSION AND FUTURE WORK
The proposed scheduling algorithm reduces the average
turnaround time of all submitted jobs. The considered
environment executed the jobs on different resources which
are geographically distributed. It is observed that the Proposed
Algorithm reduces the average turnaround by 4.22% with
Equal Job Distribution (as shown in Table 2). The algorithm
uses meta-scheduler where resource failure is not considered.
TABLE 2: Algorithms with their average turnaround time and
completion time.
Algorithms Average
Turnaround
Time
(In Seconds)
Completion
Time
(In Seconds)
Random Resource
Selection
20105.65 64420.25
Equal Job Distribution 17968.55 39000.22
Proposed Algorithm 17208.77 41840.88

3
REFERENCES
[1] Ammar H. Alhusaini, Viktor K. Prasanna,C.S.
Raghavendra, “Unified Resource Scheduling Framework for
Heterogeneous Computing Environments", in Proceedings of
the Eighth Heterogeneous Computing Workshop, San Juan,
Puerto Rico, pp. 156-165, 1999.
[2] N. Muthuvelu, J. Liu, N. L. Soe, S.r Venugopal, A.
Sulistio and R. Buyya, “A Dynamic Job Grouping-Based
Scheduling for Deploying Applications with Fine-Grained
Tasks on Global Grids", Proceedings of the 3rd Australasian
Workshop on Grid Computing and e-Research (AusGrid
2005), Newcastle, Australia, 41-48, January 30 - February 4,
2005.
[3] I. Foster, C Kesselman, “The Grid: Blueprint for a new
computing infrastructure", Morgan Kaufmann Publishers, San
Francisco, USA, 1999.
[4] R. Buyya, D. Abramson, J.Giddy, “Nimrod/G: An
Architecture for a Resource Management and Scheduling
System in a Global Computation Grid", International
Conference on High Performance Computing in Asia-Pacific
Region (HPC Asia 2000), Beijing, China. IEEE Computer
Society Press, USA, 2000.
[5] Cong Liu, Sanjeev Baskiyar and Shuang Li, “A General
Distributed Scalable Peer to Peer for Mixed Tasks in Grids",
High Performance Computing – HiPC 2007, ISBN:978-3-
540-77219-4, 320-330, 2007.
[6] Rajkumar Buyya, Manzur Murshed, “GridSim: a toolkit
for the modeling and simulation of distributed resource
management and scheduling for Grid computing", Technical
Report, Monash University, Nov. 2001. To appear in the
Journal of Concurrency and Computation: Practice and
Experience (CCPE), pp. 1-32, Wiley Press, May 2002.
[7] Volker Hamscher, Uwe Schewiegelshohn, Achim Streit,
Ramin Yahyapour, “Evaluation of Job-Scheduling Strategies
for Grid Computing", in 1st IEEE/ACM International
Workshop on Grid Computing (Grid 2000), Berlin, Lecture
Notes in Computer Science (LNCS), Springer, Berlin,
Heidelberg, New York, pp. 191-202, 2000.[8] Cormen TH,
Leiserson CE, Rivest RL, “Introduction to algorithms 2nd
edition", MIT and McGraw-Hill Book Company, Boston
Massachusetts, cp. 16, 370-403, 2001.
4
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-1


A Comparative Study of IP and MPLS Based
Networks
Sandeep singh
School of Information and Communication
Technology,
Gautam Buddha University
Greater Noida, U.P.
er.sandeep_vhdl@yahoo.com
M.Altamash Sheikh
School of Information and Communication
Technology,
Gautam Buddha University
Greater Noida, U.P.
altsheikh@gmail.com

Abstract
In today’s world of networking IP routers
buildup and uses the routing tables and these
packets of data are routed on the IP prefixes
which are already stored in routing tables of
routers. These routers fallows the lookups to
find out the next destination of packets. This
process gives several overheads like delay ,
jitter and traffic drops. simple data can
accept much delay but the real time traffic
needs guarantied QoS with low jitter , delay
and packet drop etc. optimum performance
of IP based network can not be achieved by
just applying the QoS but MPLS can
provide better support to QoS . Here the
performance factor of the network can be
improved with MPLS in a better way as
compared to just IP routing.
Keywords: QOS ,TOS , MPLS , IP Based
routing , Traffic Engineering.
Introduction Whenever there is traffic to
flow normal routing faces several issues
related to congestion and delay . Router
needs to lookup there routing table every
time so this makes router very busy during
the processing. Routers decides the path
based on its IP prefixes and this path is
always the shortest path based on shortest
path first algorithm (SPF). This results the
unnecessary load on a particular link and
gives traffic drop. When we are dealing with
the real time traffic it needs guaranteed
service[4] , [5].
In general MPLS can provides a combined
and better environment for the network with
capability of supporting QOS and traffic
engineering. In the emerging and
developing trend of technology small , mid
sized and large companies are frequently
moving for changing there infrastructure .
These companies are migrating towards
Traffic engineering[5] , [6] , [7].
Related work :
Background of QOS : At the initial stage
when IP header was created there was the
space in that header called Type Of Service
(TOS) byte which was basically for future
perspective. It was well known that
technologies were growing very rapidly as
no one could think of it. This Type of
service(TOS) field was reserved for future
purpose of improving QOS[17].
In the early development of internet the
applications running over it did not take
much care about TOS because during
forwarding of IP datagram the devices like
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-2


router did not interpret this field until it
gives any kind of service parameters. In this
TOS byte three fields were reserved for IP
precedence and other three for type of
service[12] , [13]. Higher the IP precedence
data would treated more priority and lower
the IP precedence data would be dealt with
low priority .
During the initial stage of internet it was
not needed in the network because
commercialization was not so much but as it
grown up the need of different type of
service were also grown up[16].
Classification of traffic based on QOS
Classification of different flows of traffic
into classes may depend on several
parameters. Basically similar data packets
are considered in the same class[18]. The
most common way to classify the flows may
depend on header fields such as IP
precedence and DSCP fields[15]. One of the
header of TCP is also used for classification
of traffic by identifying the length of coming
packets or by checking the MAC address of
both senders and receivers address[14].
When the traffic is classified three main
classes comes out –
(i) High priority for sensitive traffic : The
traffic like real time requires some special
treatment like no delay and less jitter .This
comes in High Priority traffic and most
common example of such traffic is
VoIP[10].
(ii) Best effort traffic : When traffic does
not need any delay related guarantees . In
this case packets are send after high priority
traffic and the best example of this kind of
traffic is email[9] , [10].
(iii) Low priority traffic : This kind of
traffic does not need any guarantee , priority
or significance and it is not required to
deliver the traffic at its appropriate
destination . This kind of traffic includes the
codes generated by a hacker or spam kind of
mails[18].
MPLS:
Multi Protocol Label Switching is a
new technology for getting the fast and rapid
transfer of data. High speed convergence of
scalability is possible in MPLS based
network [1],[2],[3]. Multi Protocol Label
Switching has become the first choice for
packet transportation , which fulfills the
several requirements of next generation[19].
Several service providers are going to
deploy the MPLS on a common platform to
achieve convergence of existing
technologies like X.25 ; ATM / Frame relay
(FR) and best effort services. It provides the
transmission resources by providing the Diff
Serve to a network. MPLS provides the
intelligent routing and gives the important
improvements in the switching performance
of the network instead of what networks
architecture is. MPLS also provides
scalability like advantages for Virtual
Private Network (VPN) and maintain end-
to-end QOS [1] , [3] , [5].
In normal IP based networks IP
routers performs destination based routing.
when they have to sent traffic router always
uses the simple shortest path first algorithm
to compute the shortest path between
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-3


themselves and destination[8] . This distance
can be hop count for protocols such as
Routing information Protocol (RIP) or least
total metrics. They don’t care whether
alternate path exist in the network or not.
This creates a problem while delivering the
traffic because when the traffic exceeds the
limit of link it gives traffic drop[11]. But in
MPLS Traffic engineering it enables the
operator to build a traffic engineering
tunnel for safely delivery of data.
MPLS Architecture
MPLS is an IETF standard which provides
Label-swapping-based forwarding in the
presence of routing information[20]. It
basically consist of some principal
components :
(1) Control : Control component uses a
Label distribution protocol to maintain label
forwarding information for all destination in
MPLS networks.
(2) Forwarding : The forwarding
component switches packets by swapping
labels using the label information carried in
the packet.


Figure 1. MPLS Architecture
(3) Ingress label edge router : This is
basically an edge router of MPLS network
which classify packets and add MPLS labels
to the packets.
(4) Label switch router : It is a networks
core MPLS enabled router which forward
the packets based on layer 2 MPLS labels
and later removes the label before sending it
to the egress label edge router.
(5) Egress label edge router : Egress label
edge router switches the unlabeled packets
based on its destination IP address.
Problem Statement and Proposed
Solution :
The primary goal of this paper is to expose
how the MPLS supports QOS needs. It also
highlight the several issues and in depth
knowledge regarding it.

(i) How MPLS helps to provide QOS
demanding traffic .
(ii) Comparison of MPLS with traditional
technologies like IP routing.
(iii) Practical implementation of MPLS with
IP based routing in simulation environment.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-4



Figure 2. Network Diagram of MPLS
Domain
The packet drop rate is increased in the
best effort service because packet arrival is
not treated with any differentiation. Due to
fill the buffer , network devices like router
and switches are filled[24]. The packets
were going to drop so receiver requests for
retransmission which gives unwanted delay.
Real time traffic like voice and video
conferencing can not stay in a queue for
long time. Error recovery is also needed to
check the data to ensure whether the content
of packet is same as it was sent originally
during the transmission [11].
Quality of service for MPLS based
network is not very much differ from the
way as any other IP network. In this way
[22] the Quality of service strategy may
divided into four stages :-
(I) Classification of traffic is considered at
the initial stage in provision of QOS at the
edge of MPLS domain. The traffic is
classified on the basis of preference[25].
Edge routers are generally responsible for
such task but switches can also install at the
edge for classifying traffic.
(II) Polarization of traffic helps the
important traffic to be forward first based on
its sensitivity and importance. This is
achieved by class of service for a given
flow. MPLS provides different categories
based on priority like gold , silver , and
bronze[26].
(III) Management of bandwidth is another
big issue for accurate delivery of traffic. It is
assumed that an amount of bandwidth is
needed to pass the low priority traffic[25] ,
[26]. Real time traffic is always given
priority and some bandwidth is reserved for
its transmission in such a way that when it
exceeds the threshold it would be denied in
order to pass the rest of the traffic.
(IV) Traffic can be shaped in such a way
that it remains in the limits of available
bandwidth pool[22]. For this purpose traffic
may be compressed or decompressed.
Router itself perform this task of
compression and decompression. Resources
are allocated in the network by applying
admission control on all label switch
paths[21].
Simulation and Results :
The simulation environment is created
in laboratory where two scenarios are being
considered in order to test the working of
MPLS. In first scenario the simple and
general network is build with a traditional
IP routing then the same network is
configured with MPLS and results were
obtained. The obtained results shows that
the MPLS network gives better output in
terms of performance and throughput as
compared to traditional IP routing.
Policies for IP based Networks
Class Matching
Criteria
DSCP Precedence
Critical RTP 46 07
Web SMTP , HTTP 18 03
Interactive ICMP , OSPF ,
EIGRP
26 5
Default
Class
Any 00 00

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-5


Table 1. Policies for IP Network

Table 2. Precedence Policies of IP Network

The Quality of Service is introduced in
both scenarios. Normally when ordinary
traffic is to be transfer it doesn’t need of
any guarantee of service. Factors of
performance degradation like Jitter , delay
and dropping of packets did not harm this
data[23]. But real time traffic is transferred
in consideration of more sensitively for such
factors.
Policies for MPLS based networks
Class Matching Criteria MPLS EXP
Critical Precedence 7 5
Web Precedence 3 3
Interactive Precedence 5 4
Default Class Precedence 0 0

Table 3. Policies for MPLS Network





Table 4. Policies for MPLS Network
(Critical Class)

Results :
The above two scenarios produces the
results related to the performance of MPLS
and IP routing. following statistics were
obtained by issuing “pkt-seq-drop-stats,
show delaystats,show jitter-stats”
commands. Here we use a PAGENT router.
PAGENT router contains the Cisco IOS which
is able to generate the automatic traffic.

Statistics of IP Networks
Sent Receive Dropped
Drop 688 688 000
Minimum Maximum Average
Jitter 0.000002 0.030636 0.000342
Delay 0.017585 0.048495 0.017951

Table 5.O/P Statics of IP Network
Statistics of MPLS based networks
Sent Received Dropped
Drop 755 755 000
Minimum Maximum Average
Jitter 0.000005 0.004931 0.000222
Delay 0.017692 0.022794 0.017928

Policies for IP based Networks
Class Matching
Criteria
Band
width
%
Average Rate
Prec7 Precedence
7
50 240000/120000
Prec5 Precedence
5
35 160000/80000
Prec3 Precedence
3
15 60000/30000
Default
Class
Precedence
0
-- 800000/400000
Policies for MPLS based networks
Class Matching Criteria Bandwidth %
MPLS EXP 5 60 %
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-6


Table 6. O/P Statics of MPLS Network
(1) Packet forwarding in MPLS is depends
on labels rather then IP prefixes , which
reduces overhead of processing of routers.
(2) It has been observed that features of load
balancing of MPLS removes the unwanted
congestion from the network.
(3) Practical scenario proved that QOS
works better in MPLS than IP routing.
(4) Large routing tables were being
generated by the routers (IP) which makes
IP router slower.
(5) In IP routing , the interior gateway
protocol (IGP) used the shortest path to
forward the packets between each hop, but
in MPLS tunneling removes such overhead
of load balancing
(6) In MPLS dropping of packets , delay
related problems and issues of jitter are very
nominal in comparison to IP based routing.
Conclusion
This research focus on practical comparison
of MPLS and IP based networking which
helps to understand the concepts and
working of MPLS. Guaranteed quality of
service is needed to transfer the real time
traffic over the network. In MPLS based
network QOS gives better results as
compared to IP networks regarding surety of
data transfer , delay and jitter etc. Delay ,
jitter and results related to packet drops were
obtained by issuing commands. Then the
network was reconfigured for MPLS
network and again some parameters were
calculated for MPLS based networks.
The statistics collected from both the
networks and MPLS gives better result than
normal IP routed networks. Packet of data
faces longer delays in IP based networks and
other factors were also improved in MPLS
based networks. All the necessary
configurations of routers have been recorded
and two different scenarios were compared.
Then routing tables of IP based networks
and Cisco Express Forwarding (CEF) tables
of MPLS networks have been obtained.
Acknowledgement
This research work was supported by
Prof(Dr.) Sanjay Jasola under the School of
Information and Communication
Technology in Gautam Buddha University,
Greater Noida.

References
[1]. Lawrence J, “Designing multiprotocol
label switching networks”, Communications
Magazine, IEEE, Volume 39, Issue 7, July
2001.

[2]. Kuribayashi,Tsumura, “Optimal LSP
selection method in MPLS Networks”,
Communications, Computers and Signal
Processing, IEEE acific Rim Conference on
22-24 Aug. 2007.

[3]. Mahesh K.P, Yadav S. V, Charhate
M.E, “Traffic Analysis of MPLS and Non
MPLS Network including MPLS Signaling
Protocols and Traffic distribution in OSPF
and MPLS”, First International Conference
on Emerging Trends in Engineering and
Technology 2008.

[4]. David Applegate, Mikkel Thorup,
“Load optimal MPLS routing with N + M
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-7


labels”, AT&T Labs–Research Shannon
Laboratory 180 Park Avenue Florham Park,
2003.

[5]. Luc De Ghein, “MPLS Fundamentals”,
Cisco Systems, Cisco Press 800 East 96th
Street Indianapolis, ISBN: 1-58705-197-4,
2007.

[6]. Benjamin Tang, Ahmet A. Akyamac,
Chi-Hung Kelvin Chu and Ramesh
Nagarajan, “MPLS Network Requirements
and Design for Carriers: Wireline and
Wireless Case Studies”,
Telecommunications Network strategy and
planning symposium, pages 1-6, Nov 2006.

[7]. Muzammil Ahmad Khan, “Quantitative
Analysis of Multiprotocol Label Switching
(MPLS)”, Student Confrence, Volume 1,
Issue 16-17, page(s): 56-65, Aug 2002.

[8]. Loa Andersson, Stewart Bryant, "The
IETF Multiprotocol Label Switching
Standard: The MPLS Transport Profile
Case," IEEE Internet Computing, vol. 12,
page(s):69-73, Aug, 2008.

[9]. Mahesh K.P, Njulata Yadav S. V,
Charhate M.E, “Traffic Analysis of MPLS
and Non MPLS Network including MPLS
Signaling Protocols and Traffic distribution
in OSPF and MPLS”, 2008.

[10]. Srinivas Vegesna Quality of Service
By Published by Cisco Press, 2001 ISBN
1578701163, 9781578701162

[11]. B.E. Nichols, “Differentiated services
in the Internet Carpenter”, Proceedings of
the IEEE, Volume 90, Issue 9,
page(s):1479–1494, Sept 2002.

[12]. Leonardo Balliache, Network QoS
Using Cisco HOWTO, April 2003 v0.1
Published on May 1, 2003, v0.2 May 15,
2003, v0.3 September 10, 2003

[13]. Traffic conditioning factors
http://www.cisco.com/en/US/tech/tk543/tk7
57/technologies_tech_note09186a00800949f
2.shtml http://en.wikipedia.org/ wiki/
IP_traceback# Packet_marking,
http://en.wikipedia.org/wiki/Traffic_policing
,http://en.wikipedia.org/wiki/Traffic_shapin
g# Traffic_Classification(Access Date
January 12, 2009)

[14]. Carter Horney for Nuntius Systems,
Quality of Service and MPLS (White
Paper), Inc.13700 Alton Pkwy., Suite 154-
266 Irvine, CA 92612 949.295.0475 voice
www.nuntius.com

[15]. P. Prabagaran & Joseph
B,“Experiences with Class of Service (CoS)
Translations in IP/MPLS Network”,
Proceedings of the 26th Annual IEEE
Conference on Local Computer Networks,
ISSN:0742-1303, page(s):243, 2001.

[16]. Ji-Feng Chiu, *Zuo-Po Huang, *Chi-
Wen Lo, *Wen-Shyang Hwang and Ce-
Kuen Shieh, “An Approach of End-to-End
DiffServ/MPLS QoS Context Transfer in
HMIPv6 Net”, Autonomous Decentralized
Systems, Eighth International Symposium,
ISBN: 0-7695-2804, page(s):245-254, Mar
2007.

[17]. Sundeep .B.Singh and Girish.P.Saraph,
“DiffServ over MPLS: Tuning QOS
parameters for Converged Traffic using
Linux Traffic Control”, Indian Institute of
Technology Bombay, Powai, Mumbai India.

[18]. Hongyun Man, Linying Xu, Zijian Li,
Lianfang Zhang, “END-TO-END QOS
IMPLEMENT BY DIFFSERV AND
MPLS”, Electrical and Computer
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-8


Engineering, volume 2, page(s):641-644,
May 2004.

[19]. International Workshop
NGNT,“DiffServ and MPLS Tímea
Dreilinger”, Budapest University of
Technology and Economics, Department of
Telecommunication and Telematics, High
Speed Networks Laboratory.

[20]. F. Le Faucheur, L.Wu, S. Davari, P.
Vaananen, R. Krishnan, P. Cheval,
J.heinanen, “Multi-Protocol Label Switching
(MPLS) Support of Differentiated Services”,
RFC 3270, May 2002.

[21]. John Bartlett and Rebecca Wetzel
BCR, “QOS over MPLS–the Complete
Story”, P2P for Communications: beyond
file sharing, Volume 36, Number 2,
February 2006.

[22]. Victoria Fineberg, “QoS Support in
MPLS Networks MPLS/Frame Relay
Alliance”, White Paper May 2003.

[23]. LI Ming-hui and Xiz Jing-bo,
“Research and Simulation on VPN
Networking Based on MPLS”, Wireless
Communication, Networking and mobile
Computing ,page(s): 1- 4, Oct 2008.

[24]. Hiroshi Yamada, “End-to-End
Performance Design Framework of MPLS
Virtual Private Network Service across
Autonomous System Boundaries”,
Telecommunications Network strategy and
planning Symposium , page(s): 1-6, Nov
2006

[25]. Label distribution:
“http://www.cisco.com/en/US/docs/ios/mpls
/configuration/guide/mp_mpls_overview_ps
6350_TSD_Products_Configuration_Guide_
Chapter.html” (last accessed date 10 April
2009).

[26]. Jing Wu, Delfin Y.Montuno, Hussein
T.Mouftah*, “Improving reliability of label
distribution protocol”, Proceedings of the
26th Annual IEEE Conference on Local
Computer Networks, ISSN: 0742-1303,
page(s) 236, 2001.


Sandeep Singh has received his
B. Tech. degree in Electronics and
Communication Engineering from
Chaudhary Charan Singh University, Meerut
in 2008. He is a M.Tech. student at Gautam
Buddha University, Greater Noida. He has
been worked as a faculty of Dr.K.N.Modi
Institute of Engineering and Technology,
Modinagar, Ghaziabad. He is working as an
Assistant Professor at Sobhasaria
Engineering College, Sikar, Rajasthan. He is
a life time member of IACSIT, Singapore
and member of CSI. He has published a
paper in “national conference on
communication network and security”,
Alwar, Rajasthan. His research interest
includes Ad-hoc wireless networks, Mobile
IP, Quality of Service, MPLS based
networks and Traffic Engineering.





CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O6-9


M.Altamash Sheikh has
received his B.Sc degree in
Electronics Engineering in
2007 from Annamalai
University, Tamil Nadu. He
received his M.Sc. degree in Electronics
Communication Engineering from Jamia
Milia Islamia, New Delhi in 2009. He is a
M.Tech. student at Gautam Buddha
University, Greater Noida. His research
interest includes , mobile IP, Sensor
Networks, Ad-Hoc Networks, & Bluetooth
Technology.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O7-1

INTEGRATED SERVICES DIGITAL NETWORK
AJAY BHATIA
(Dept. Electronics & Communication)
Punjab College of Engineering and Technology, Mohali

ABSTRACT
Integrated Services Digital Network(ISDN)
is a digital communications technology that
enables a small business or an individual to
connect directly to both the Internet and
other users. ISDN provides a standard
interface for voice, fax, video, graphics and
data- all on a single telephone line. ISDN is
a set of communications standards for
simultaneous digital transmission of voice,
video, data, and other network services over
the traditional circuits of the public switched
telephone network. It was first defined in
1988 in the CCITT red book. ISDN enables
the telephone network, which was built for
traditional analog calls, to carry information
digitally at higher speeds and without the
errors of the traditional analog system or
POTs (Plain Old Telephone Services) lines.
As a result, ISDN has the ability to deliver a
wide range of desirable applications in
education, healthcare, business, and
multimedia. In light of this substantially
increased functionality, ISDN can be the
basis for a new type of useful medium for
Internet access that is: reasonably priced,
readily available, and also allows for the
delivery of information at high speeds, and
provides for many types of new and
previously unavailable applications. The
technical design of ISDN is motivated by
the primary aim of modernizing traditional
telephone networks. The use of digital
switching and transmission technology
within the ISDN facilitates the transmission
of digital data besides the digitized voice.
KEYWORDS
ISDN, ISDN traffic, voice/data networks, value added services, video conferencing.


INTRODUCTION
Integrated Services Digital Network (ISDN)
is comprised of digital telephony and data-
transport services offered by regional
telephone carriers. ISDN involves the
digitalization of the telephone network,
which permits voice, data, text, graphics,
music, video, and other source material to be
transmitted over existing telephone. The
emergence of ISDN represents an effort to
standardize the subscriber services,
user/network interfaces, and network and
internetwork capabilities. ISDN applications
include high-speed image applications,
additional telephone lines in homes to serve
the telecommuting industry, high-speed file
transfer, and video conferencing. Voice
service is also an application for ISDN. This
paper summarizes the underlying
technologies and the services associated
with ISDN. Also, there are two significant
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O7-2

developments drive the commercialization
of ISDN: digital switching and digital
transmission. It is estimated that nearly all of
the access lines and trunk routes in the
United States will be digital within a few
years and a third of these will be fiber optic
cable. Pulse Code Modulation (PCM) and
Time Division Multiplexing (TDM) are
important related technical developments
enabling digital switching and transmission.

ISDN STANDARDS
The ISDN Basic Rate Interface (BRI)
service offers two B channels and one D
channel (2B+D). BRI B-channel service
operates at 64 kbps and is meant to carry
user data; BRI D-channel service operates at
16 kbps and is meant to carry control and
signaling information, although it can
support user data transmission under certain
circumstances. The D channel signaling
protocol comprises Layers 1 through 3 of
the OSI reference model. BRI also provides
for framing control and other overhead,
bringing its total bit rate to 192 kbps. The
BRI physical-layer specification is
International Telecommunication Union
Telecommunication Standardization Sector
(ITU-T) (formerly the Consultative
Committee for International Telegraph and
Telephone [CCITT]) I.430. ISDN Primary
Rate Interface (PRI) service offers 23 B
channels and one D channel in North
America and Japan, yielding a total bit rate
of 1.544 Mbps (the PRI D channel runs at 64
Kbps). ISDN PRI in Europe, Australia, and
other parts of the world provides 30 B
channels plus one 64-Kbps D channel and a
total interface rate of 2.048 Mbps. The PRI
physical-layer specification is ITU-T I.431.




ISDN SIGNALING
There are two different types of signaling
used in ISDN. For communicating with the
local phone company, ISDN uses the Digital
Subscriber Signaling System #1 (DSS1).
DSS1 defines what format the data goes in
on the D-channel, how it is addressed, etc. It
also defines message formats for a variety of
messages used for establishing, maintaining,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O7-3

and dropping calls, for instance SETUP
messages, SUSPEND and RESUME
messages, and DISCONNECT messages.
Once your DSS1 signal makes it the the
phone company, their own signaling system
takes over to pass the call information
within their system, and between other
phone companies. Signaling System #7
(SS7) is supposed to be used for this. SS7
defines a communications protocol, and
formats similar to DSS1, however SS7 is
designed in a broader, more general way.
DSS1 is specific to ISDN, however SS7 will
handle the signaling needs of ISDN as well
as other older signaling systems and
(hopefully) will adapt well to future needs.
One important feature of SS7 is providing
CCS. This makes it harder for malicious
users of the phone network to put one over
on the phone company. It also improves the
service, for instance by offering faster
connection establishment. Older equipment
still looks for the signaling information in
the same channel as the voice, in the eighth
bit of each piece of voice data.

ISDN SWITCHING
With pure ISDN, switching is the
department of the phone company.
Traditional phone services is Circuit
Switched Voice (CSV). The voice data goes
through several switches before reaching its
final destination. For point-to-point data
connections, you need Circuit Switched
Data (CSD) - the exact same thing with data
instead of voice. If CSV are used, they are
free to route your call through any type of
switch, even the old analog switches. The
digital channel may also be shared with
other channels, in the moments when there
is silence on the phone line. And the digital
parts of a CSV call can go through noisy
switches that might create an undetected
error. There are actually a standard set of
combinations defined for setting up BRIs.
These are called National ISDN Interface
Groups (NIIGs),so there will be a limited
menu of offerings available. One can get
both B-channels for data, or one for voice
and the other for data, or one for voice and
the other for either voice or data. In order to
facilitate this, North American phone
companies use an optional part of the ISDN
standard to identify each TE1 or TA you
use. The phone company assigns a Service
Profile Identifier (SPID) to each of these
devices, and one has to manually enter them
into each device use. The phone company
then stores this data somewhere, and when
one connect the machine to the network, it
sends its SPID to the nearest phone
company switch which identifies what type
of connection the device needs and
(therefore) how to route its calls.
Presumably, the SPIDs have to refer to a
configuration that matches one of the two B-
channels you have. By the way, the SPIDS
are arbitrary numbers that refer to data
stored by the phone company. The phone
Company often includes the phone number
in the SPID for their own convenience. One
older type of phone company switch, a
DMS-100, was improperly designed with
respect to the standards relating to SPIDs.
This switch misguidedly assigns one SPID
to each B-channel that is use, rather than to
each device. Therefore if the nearest switch
is a DMS-100, one will only be able to hook
up two devices to the CPI, rather than eight.
If one is only going to be hooking up a
single device to the ISDN (i.e. setting it up
in a point-to-point configuration, one might
not need a SPID at all, as the phone
company can identify the ISDN line as one
particular type, full time. This depends on
what equipment they have - the old DMS-
100 switch will still require to have a SPID.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O7-4

ISDN FRAME


 Protocol discriminator
The protocol used to encode the
remainder of the encoder.
 Length of call refrence value
Defines the length of the next field.the
call refrence may be one or two octets
long depending upon the size of the
being encoded.
 Flag
Set to zero for the messages sent by the
party that allocated the call refrence
value,otherwise set to one.
 Call refrence value
An arbitrary value that is allocated for
the duration of the specific session
which identifies the call between the
device maintaining the call ant the
ISDN switch.
 Message type
Defines the primary purpose of the
frame .the message type may be one
octet or two octets .when there is more
than one octet ,the first octet is coded as
eight zeroes.


ISDN AND VALUE ADDED SERVICES

The usual reference to an ISDN implies an
integration of services on the sense of the
transport of digitalized voice and data, the
former with telephonic signaling and the
latter either gaining access via telephonic
signaling and then employing data signaling
or via data signaling only. Circuit and
message, synchronous and asynchronous,
packet - all traffic types are to be
accommodated. Packet switching is
considered by some as a value-added
service, but by others as a transport
mechanism, which can be combined within


a communications node with circuit traffic.
The potential for combining the
telecommunications and the computer
technologies becomes apparent when the
delivered signal is not just a replica of the
input signal, but has been processed in a
useful manner. It can be debated on
operational, economical, technical and
political grounds whether such processing
should be included within the public and
private network itself, treated and offered as
a network service, or whether these
functions should be external to the network,
provided at its periphery.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O7-5

ISDN VIDEOCONFERENCING
Today, a modern organization is connected
via an IP network resembling a “hub”
between dedicated connections in multiple
remote offices. Employees use the IP
network for text-based communications,
meet regularly via telephone and frequently
travel to remote offices for face-to-face
meetings. When at their desks, employees
have both Intranet services and access to the
public Internet. Videoconferencing is being
explored as a technology that can improve
group and individual communications and
thereby affect both these strategic
objectives.

CONNECTIVITY OPTIONS
In the past, network options for
videoconferencing were simpler. Heavy
users placed endpoints on a common
“private network” of switched circuits
leased from telecommunications carrier.
Most companies preferred to connect
videoconferencing endpoints to a public
network via Integrated Switched Digital
Network (ISDN) connections obtained from
the local exchange carrier. As a result of the
improved ability to guarantee quality of
services in packet networks, the ubiquity of
switched local area Ethernet networks, and
the release of numerous products for
business quality H.323 videoconferencing,
IP-based networks are now firmly on the list
of videoconferencing connectivity options.

ISDN VIDEOCNFERENCING SECURITY
In order to properly secure video
conference meeting, one must be concerned
about three key areas: data storage, data
radiation, and data encryption.

Data Storage:-
There are two types of videoconferencing
systems, or codecs, commercially available
today; PC-based and appliance.
The PC-BASED VIDEOCONFERENCING
SYSTEMS that run PC operating systems
and utilize, at least in part, standard COTS
PC hardware. Since these systems include
internal hard drives and plug-and-play
connectivity for external storage devices,
one must take additional steps to secure
these systems. Specifically, the system must
be equipped with a removable hard-drive
setup. In addition, two different hard drives
must be used; one for secure calls and one
for non-secure calls.Because securing a PC-
based videoconferencing system requires
additional cost and work, these systems are
not very well suited for secure
videoconferencing.
APPLIANCE VIDEOCONFERENCING
SYSTEMS are not based on a PC platform.
These devices have been custom designed
and manufactured to provide only specific
functionality and typically do not utilize
standard PC- based hardware or software. In
addition, the storage capabilities of these
systems are usually limited to storing
address book information, usage data, and
configuration settings. Since these devices
do not provide data storage capabilities, they
are a good choice for secure
videoconferencing.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O7-6


Data Radiation and Encryption:-
Another area of concern is that confidential
information may be electrically radiated by
the conferencing equipment. When this
occurs, data is susceptible to monitoring and
eavesdropping by unauthorized personnel.
According to the F.A.S. (Federation of
American Scientists), this would allow a
person to view all of the contents of another
person’s computer monitor from over half a
mile away – without being detected.
Therefore, to provide a secure
videoconference environment, data radiation
must be controlled or eliminated.

BENEFITS OF VIDEOCONFERENCING

Cost savings.
Time savings.
Improved quality of life.
Gaining ad-hoc access to contacts.
Bandwidth on demand
Efficient and Flexible
Reliable Communication

ISDN BACKUP


BROADBAND ISDN (B-ISDN)
B-ISDN is not a computer network but the
telephone network. Specifically, it is the
worldwide digital telephone system
currently being installed in virtually every
developed country of the world. Once
completed, B-ISDN will permit its users to
communicate over high-quality, high-speed
digital communication channels. The
supported media include telex, fax, voice
telephone, video telephone, audio, high
definition TV (HDTV), and, of course,
computer networking.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O7-7


B-ISDN WITH ATM

B-ISDN is the so-called Information
Superhighway. Along with ATM,
Asynchronous Transfer Mode, is often
confused with B-ISDN. In fact, ATM is
simply a service that can run over B-ISDN.
It is literally little more than the
specification for a 48-byte packet or cell of
information with a five-byte header which
tells the telephone system where that packet
is going. It can run over a number of
different physical media, ranging from UTP
(unshielded twisted pair), to something
called TAXI, to the B-ISDN superhighway.

APPLICATIONS OF ISDN
Studio quality audio transmission for
broadcast.
Image archives(Real estate, medical
images, photographic image banks etc.).
Preparation of printed materials.
Electronic manuals online.
Stock quote for brokers.
Credit card authorization.

CONCLUSION

As ISDN is deployed, few people are
currently replacing their home phone system
with an ISDN phone network. The trend for
now seems to be providing an entire ISDN
network in a single box, with the NT1, TA,
and TE1 equipment all built in. An example
would be the Pipeline 25, from Ascend,
which provides ISDN to Ethernet
connections, using IP. It has an NT1 built in,
and provides two phone jacks for standard
POTS telephones.

REFERENCES

 www.radvision.com
 Computer networks ISDN systems.
 An ISDN approach to integrated
corporate networks.
 www.perey.com
 www.searchnetworking.com
 CCITT Red Book,
Recommendations on ISDN,
Geneva, 1985.
 M. Huet and P. Put, TRANSPAC
and Emerging ISDN, P.J. Kuehn,
ed., New Communication Services: A
Challenge to Computer Technology
(North-Holland, Amsterdam, 1986)
119-124.
 M. Intorella, M. Benedetti and G.
Gasparrone, Italy's Pilot ISDN
Service, in: P.J. Kuehn, ed., New
Communication Services: A
Challenge to Computer Technology.
 www.data .com

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-1

PERFORMANCE ANALYSIS OF WIRED NETWORKS
UNDER VARYING NETWORK ATTRIBUTES
Vaibhav Nijhawan

Department of Electronics & Communication Engineering
Allenhouse Institute of Technology, Kanpur
{vaibhav.nijhawan@gmail.com}

Abstract: Wired network is an interconnection
of remote nodes through physical
infrastructure of wires and cables. It provides
advantages of fast, reliable, secure and long
haul communication. The modeling and
analysis of a network provides understanding
of behavior of various network parameters
which can be further used for computation of
desired results and efficiency. In this paper
analysis of wired networks has been done to
evaluate the performance of a network under
varying bandwidth of transmission links and
number of workstations (nodes) attached to the
network.

Keywords: OPNET, Wired Networks,
Transmission links

I. INTRODUCTION
Data communication and networking is the fastest
growing technology in the modern era. Advanced
research projects agency network (ARPANET)
added fuel to the growth of wired network
industry and since then wired networks are
emerging as social networks which are linking
people, organizations and knowledge worldwide.
This radical change in the wired networking
domain over the last decades is the result of
convergence of internet coupled with engineering
advances in the field of information and
communication technologies. Wired computer
networks have bridged the gap between countries
and continents and brought up an era of
globalization.

II. WIRED NETWORKS
Wired network is an interconnection of remote
nodes through physical infrastructure of wires and
cables. It provides advantages of fast, reliable,
secure and long haul communications. The wired
networks are most suitable for environments with
fixed entities like home and office networks and
high data rates environments where dedicated
links are necessary as a link between keyboard
and central processing unit (CPU). The primary
parts of a wired network are network cables,
network adapters, hubs, switches, routers and
process governing software’s like transmission
control protocol/internet protocol (TCP/IP), token
ring, fiber distributed data interface (FDDI). The
processing capabilities and methodologies of the
computer hardware and networking components
may be slightly varied by using different network
processing software’s. Institution of electrical and
electronics engineering (IEEE) model IEEE 802.3
Ethernet working on transmission control
protocol/internet protocol (TCP/IP) model is most
famous of all the wired local area network models.
The most common Ethernet links used are
10BaseT, 100BaseT and 1000BaseX providing a
data transfer rate of 10 Mbps, 100 Mbps, and 1
Gbps respectively. Fiber optic variants of Ethernet
offer high performance, electrical isolation and
network span up to tens of kilometers. The
backbone of internetwork of interconnected
networks (INTERNET) relies on the huge
worldwide established wired network known as
public switch telephone network (PSTN). Some of
the limitations of wired networks include
immobility, uneasy upgradation, difficult fault
diagnosis, elaborate infrastructure, intricate
installation and its non availability at isolated
places.

III. COMPONENTS OF WIRED NETWORK
Components of wired network includes all those
devices that when combined together forms an
effective communication link. The network
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-2

components should facilitate performance,
reliability and security of the network. The
computer network component includes
transmission media, networking devices
(hardware) and network governing software’s.

Transmission Media
Transmission media is a physical link between the
network nodes that connects the nodes to form a
network. Physical links may vary from a common
twisted pair cable to an expensive fiber optic cable
according to the scale and performance criterion
of the network. Twisted pair cables consist of two
conductors (normally copper) each with its own
plastic insulation twisted together. The twisting of
cables together eliminates unwanted signals like
interference and crosstalk. The most common
twisted pair cable used in communication is
referred to as unshielded twisted pair (UTP). The
electronic industries association has developed
standards to classify unshielded twisted pair cable
into seven categories, defining the cable quality
with 1 as the lowest and 7 as the highest. Twisted
pair cables are used in telephone lines to provide
voice and data channels. Digital subscriber line
(DSL), Ethernet (10BaseT, 100BaseT) employs
twisted pair cables as their physical media.
Coaxial cables as transmission media are used to
carry signals of higher frequency ranges. These
cables carry a core conductor of copper wire
enclosed in an insulating sheath which in turn is
encased in an outer conductor of metal foil.
Coaxial cables are employed in television and
telephone networks.



Figure 1 Twisted pair cable and coaxial cable

A fiber optic cable is made of glass or plastic and
transmits signals in the form of light. It provides
enormous amount of bandwidth and transmits data
over large distances. Optical fiber cables are
mainly used in long haul communications and
high speed local area networks (10BaseF).

Networking Devices
Networking devices or hardware devices or
intermediate units (IU) are equipments that
mediate data in a computer network. Networking
devices play a vital role in describing the behavior
and characteristics of network parameters like
delay, transmission rate, throughput and
bandwidth. Common networking devices are
Network Interface Cards (NICs), modulator
demodulator (MODEM), Bridges, Hubs, Switches
and Routers.
Network interface cards (NIC) provides
physical access to a network media (link)
through the use of medium access control
(MAC) addresses. Network interface cards
(NICs) operates at a speed ratings of 11
Mbps, 54 Mbps and 100Mbps.
Repeater is an electronic device that
receives a signal cleans it of unnecessary
noise, regenerates it and transmits it at a
higher power level, so that the signal can
cover longer distances without
degradation. In most twisted pair Ethernet
configurations repeaters are required for
cable that runs longer than 100 meters.
Network hub or concentrator duplicates
the data packets received through one port
and makes it available to all the ports
attached to it allowing data sharing
between all devices connected to the hub.
The network switch is a device that
forwards datagram between ports based on
the MAC addresses in the data packets.
Switch is distinct from a hub in that it only
forwards the frames to the ports involved
in the communication rather than to all
ports connected.
A router is used to connect subnets to the
main network or two distinct networks
together, provided they are based on the
same network model. As router only
understands the addressing modes of
similar networks (networks that are
governed by same protocol suite).
Gateway is a device that is used to transfer
data between dissimilar networks
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-3

(networks that are governed by different
protocol suite).
Bridge is used in packet-switched
computer networks and unlike routers;
bridge makes no assumptions about where
in a network a particular address is
located. It depends on flooding and
examination of source addresses in
received packet headers to locate unknown
devices. Once a device has been located,
its location is recorded in a table where the
MAC address is stored. Their application
is limited to local area networks because of
its lengthy algorithms.


Figure 2. A wired computer network

IV. COMMUNICATION SOFTWARES
A network uses number of protocols to establish a
link between nodes for data and resource sharing.
The most common protocol that governs the
communication in almost every network is
transmission control protocol/ internet protocol
(TCP/IP) known as the internet protocol suite.
Some other important protocols working under or
over the internet protocol suite includes hypertext
transfer protocol (HTTP), file transfer protocol
(FTP), medium access protocol (MAC),address
resolution protocol (ARP), secure file transfer
protocol (SSH), internet control message protocol
(ICMP) and multi protocols label switching
(MPLS). The arrival of internet work of
interconnected networks (INTERNET) into daily
life has greatly influenced the relevance and
importance of networks. Nowadays more and
more applications rely on efficient and reliable
network operations such as World Wide Web
(WWW), electronic-mail, instant messaging,
voice over internet protocol (VoIP) and network
based audio and video services thus making
computer networks an important part of
infrastructures for home and offices. Therefore it
is often desirable to improve the efficiency and
designing of computer networks. Analytical
methods using queuing network theory and
computer simulation methods are extensively used
for network analysis and evaluation. Former is a
conventional approach used for smaller networks
and later is best suited for the analysis of complex
and large networks of nowadays.
The modeling and analysis of a network provides
the understanding of the behavior of the network
parameters which can be used for analysis of
desired results and higher efficiency. Moreover
the study helps in increasing the application
domain of the technology and sometimes may also
result in discovery of newer technologies.

V. INTRODUCTION TO PROBLEM
A comparative study of network parameters
delay-global (sec), delay-node (sec), load
(bits/sec) and traffic received (bits/sec) is made by
varying the physical characteristics and logical
conditions of computer networks. Physical
characteristics involved number of nodes and type
of cabling. The size of data packet is chosen from
logical domain. The delay may be measured as a
global parameter or as a nodal parameter. In
global sense, the delay may be defined as an
average time gap experienced by all the nodes in a
network between transmission and reception of
data packets or between request and response. The
delay in nodal sense is defined as the average time
taken by the data packets to reach a particular
node. It is measured in seconds. Load is the
amount of data packets transmitted by a particular
node, thereby adding load on network. It is a
nodal characteristic and is measured in bits per
second (bps). Traffic received is the amount of
throughput measured for the particular node. It is
measured in bits per second or packets per second.
The comparison of delay, traffic received and
load is computed by varying
Type of link which may be categorized
depending upon its speed of carrying data.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-4

Size of the network by varying number of
nodes connected to the network.
The analysis was done on an Ethernet based local
area network (LAN) having physical design of
star topology. The choice of Ethernet LAN having
star topology was made because Ethernet is the
most common local area network worldwide.

VI. NETWORK MODELING
Network 1: The LAN consisting of 5 Ethernet
stations each having a uniform data packet size of
1500 bits is modeled in the office
(100*100meters) environment of OPNET in three
scenarios using 10BaseT cabling, 100BaseT
cabling and 1000BaseX cabling.

Figure 3. Network 1 Model
Methodology: The network parameters: delay
(seconds) of the whole network, load (bits/sec) for
network node (node_0) and traffic received
(bits/sec) by node_0 are investigated. The
simulation time is set to 30 minutes.

Network 2: Ethernet LAN having a link type of
10BaseT and uniform packet size of 1500 bits is
tested by varying the number of nodes from 10, 20
to 30 nodes. The parameters tested are delay (sec)
for the whole network, delay (sec) for node:
testnode, load (bps) and traffic received (bps) for
node testnode.
Figure 4. Network 2

Methodology: The self comparison of network
parameters delay-global (sec), delay-node (sec),
load (bits/sec) for node (test_node) and traffic
received (bits/sec) for node (test_node) is made.

VII. SIMULATION RESULTS
The results obtained from the study of self
comparison of parameters delay-global (sec),
delay-node (sec), load (bits/sec) and traffic
received (bits/sec) under physical and logical
constraints of type of link & number of nodes in
computer networks simulated are presented.

Network 1: Results and analysis

Ethernet local area network (LAN) is established
by using link 10BaseT, 100BaseT and
1000BaseX.
 The delay in data transmission varies with
the type of cable that is used to connect
nodes in the network. The 10BaseT cable
having the minimum rate of data
transmission produces maximum delay,
whereas delay is manageable in 100BaseT
cable and is negligible in 1000BaseX
cable.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-5


Figure 5. Variation in delay-global

 On changing the cable from 10BaseT,
100BaseT to 1000BaseX, no variation in
load pattern is observed across the node
(node_0) as recorded by the simulation
results.


Figure 6 Variation in load

 The simulation result shows that there is
no change in traffic received by node
(node_0) on changing the cables from
10BaseT, 100BaseT to 1000BaseX.

Figure 7.Variation in traffic received

The results obtained by simulation of the network
verifies that the Ethernet cables are categorized on
the basis of data rates.10BaseT, 100BaseT and
1000BaseX categories of cables that provide
different data rates. 10BaseT is the slowest of two
and offers maximum delay, whereas 1000BaseX
is fastest and has negligible delay.

Table 1 Result Comparison: Network 1


Network 2 Result analysis

Local area network following the Ethernet
standard is established for 10 nodes, 20 nodes and
30 nodes for comparison of parameter
performance.
 The delay-global (sec) in network
increases on increasing the number of
nodes in network from 10, 20 to 30 nodes.
The simulation test for load (sec) analysis
lasted for 30 minutes.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-6


Figure 8. Variation in delay-global

 The load (bits/sec) by node (testnode)
increases with increase in number of nodes
from 10 nodes, 20 nodes to 30 nodes over
the time. During first hour of simulation
the network having least number of nodes
is affected by maximum per node load
(bps) but during intervals of 8 hours and
20 hours the network with 20 nodes and 30
nodes start contributing maximum per
node load (bps) respectively on the
network.

Figure 9. variation in load (bits/sec) for 8 Hrs and
20 Hrs

 The received traffic (bits/sec) by node
(testnode) increases linearly with increase
in number of nodes from 10 nodes, 20
nodes to 30 nodes over the time. The
simulation results show that the traffic
received by individual node (bps) is
maximum in a network having maximum
number of nodes. The simulation results
for 8 hour simulation and 20 hour
simulation is presented.

Figure 10 Variation in traffic received (bits/sec)
for 8 hrs and 20 hrs
The results obtained by simulation of the network
verifies the statements
 The delay-global (sec) and delay-node
(sec) increases with increase in number of
nodes in computer network as a node
contributes to the traffic (bps) and load
(bps). The traffic in network contributes to
collisions and data loss, which results in
delay of data packets reaching the
destination hence contributing to the
overall network delay. Therefore more the
number of nodes more will be the delay.
 The load (bps) contributed by an
individual node increases with the increase
in number of nodes in the network. More
the nodes in network more will be the
communication or transfer of data between
them, which in turn increases the load on
network links.
 Traffic received (bps) by an individual
node has maximum value for the
maximum number of nodes initially. But
as the number of nodes increases from 10
nodes, 20 nodes to 30 nodes the number of
collisions, the delay also increases in
Ethernet links. Therefore the traffic
received (bps) for minimum number of
nodes increases linearly from the network
having maximum number of nodes. But
after duration of 20 hours, when the
network stabilizes a linear relationship
develops between traffic received by
individual node of networks with number
of nodes varying from 10 nodes, 20 nodes
to 30 nodes. Network with 30 nodes
attains the maximum linear value whereas
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-7

network with 10 nodes attains the least
linear value.





VIII. CONCLUSION
The Ethernet cables are categorized on the basis
of data rates.10BaseT, 100BaseT and 1000BaseX
categories of cables that provide different data
rates. 10BaseT is the slowest of two and offers
maximum delay, whereas 1000BaseX is fastest
and has negligible delay. The delay and traffic
received increases with increase in the network
load.

REFERENCES
[1] S. Pierre, G. Legault, “A genetic algorithm
for designing distributed computer
network topologies”, IEEE transaction on
system, man and cybernetics, April 1998,
vol. 28, issue 2, pp 249-258
[2] F. Kamoun, L. Kleinrock, “Analysis of
shared finite storage in a computer
network node environment under general
traffic conditions”, IEEE transaction on
communications, July 1980, vol. 28, issue
7, pp 992-1003
[3] Yu-Chee Tseng, Chih-shu Hsu, Ten-
Yueng Hsieh, “Power saving protocols for
IEEE 802.11 based multi hop adhoc
networks”, computer networks magazine,
October 2003, vol.43, issue 3, pp 317-337
[4] B.C. Neuman, T. T’so, “Kerberos: an
authentication service for computer
networks “, IEEE communication
magazine, September 1994, vol.32, issue
9, pp 33-38
[5] C.S. Raghavendra, M. Gerla, A. Avizienis,
“Reliable loop topologies for large local
computer networks”, IEEE transactions on
computers, January 1985, vol. c-34, issue
1, pp 46-55
[6] Shenker, David Clark, Deborash Estin,
Shai Herzog, “Pricing in computer
networks: reshaping the research agenda”,
Telecommunication policy, April 1996,
vol.2, issue 3, pp 183-201
[7] Laurent Paquereau, Bjarne E. Helvik, “A
module based wireless node for NS2”,
Proceedings from the workshop NS2: The
IP network simulator, 2006, article 4
[8] Qichen, Felix Schnidt- Eisenlohr, Daniel
Jiang, Marc-Torrent-Moreno, Luca
Delgrossi, Hannes Hartenstein, “Overhaul
of IEEE 802.11 modeling and simulation
in NS2”, Proceedings of the 10
th
ACM
symposium on modeling, analysis and
simulation of wireless and mobile systems,
2007, pp 159-168
[9] Min-Seak Kang, Jaeshin Jang,
“Performance evaluation of IEEE 802.16d
ARQ algorithm with NS2 simulator”,
Communications magazine, August 2006,
pp 1-5
[10] D. Mahrenholz, S.Ivanov, “Real time
network emulation with NS2”, Eight IEEE
International symposiums on distributes
simulation and real time applications,
October 2004, pp 29-36
[11] S. Kurkowski, T. Camp, N. Mushell, M.
Colagrasso, “Visualization and analysis
tool for NS2 wireless simulations:
iNSpect”, 13 IEEE International
symposium on modeling, analysis and
simulation of computer and
telecommunications systems, September
2005, pp 503-506
[12] Li Huijie, Lin Xiaokang, “A network
based 3 tier network simulation
architecture”, 2005 IEEE International
symposium on communication and
information technology, October 2005, pp
793-796
[13] Yijinzhang, Pingping Xu, Guangguo Bi,
Sheng Bao, “Analysis of energy efficient
power saving in IEEE 802.15.4”, IEEE
NETWORK ATTRIBUTES NETWORK PARAMETERS
Type of
cable
No.
of
node
Packet
Size
Delay-
Global
(Sec)
Load
(bps)
Traffic
Received
(bps)
10BaseT 10 1500 19
millisecond
s
520
approx.
560
approx.
100Base
T
20 1500 16
millisecond
s
620
approx.
590
approx.
100Base
X
30 1500 15
millisecond
s
625
approx.
600 p
r
o
x
Table 2 Result comparison: Network 2


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO1O8-8

wireless communication and networking
conference, march 2007, pp 3330-3334
[14] Li- ming, Li- dong Zhu, Shi-qiwa, “Design
on the simulation platform for mobility
management in LEO satellite network
based on OPNET” Future generation
communication and networking, 2007, vol.
1, pp 198-202
[15] M. Foras, J. Mohorko, “ Estimating the
parameters of measured self similar traffic
modeling in OPNET”, Systems, signals
and image processing, June 2007, pp 78-
81
[16] N. Golmie, R.E Vandyck, A. soltanian,
“Interference of Bluetooth and IEEE
802.11 simulation, modeling and
performance evaluation”, International;
workshop on modeling analysis and
simulation of wireless and mobile systems,
2001, pp 11-18
[17] R. Pries, D. Staehle, D. Marsico,
“Performance evaluation of piggyback
requests in IEEE 802.16”, IEEE 66
th
vehicular technology conference,
September 2007, pp 1892-1896
[18] D. Stevanaic , N. Vlajc, “Performance of
IEEE 802.15.4 in wireless sensor networks
with a mobile sink implementing various
mobility strategies”, 33 IEEE conference
on local computer networks, October
2008, pp 680-688
[19] Guangwei Bai, C. Williamson,
“Simulation evaluation of wireless web
performance in an IEEE 802.11b
classroom area network”, 28 Annual IEEE
international conferences on local
computer networks, October 2003, pp 663-
672
[20] M.S. Hasan, Hangnian Yu, A. Griffiths,
T.C Yang, “Simulation of distributed
wireless networked control systems over
MANET using OPNET”, 2001 IEEE
International conference on networking,
sensing and control, April 2007, pp 699-
704
[21] Huazhu Zhu, Ming Li, I. Chlamtac, B.
Prabhakaran, “A survey of quality of
service in IEEE 802.11 networks”, IEEE
wireless communications, August 2004,
vol. 11, issue 4, pp 6-14
[22] S. Garg, M. Kapper, “An experimental
study of throughput for UDP and VoIP
traffic in IEEE 802.11b networks”, IEEE
conference on wireless communication
and networking, march 2003, vol.3, pp
1748-1753
[23] Mache Creeger, “Embracing wired
networks”, Queue archive, vol.5, issue 4,
June 2007, pp 12-15
[24] F. Cali, M.conti, E. Gregori, “IEEE 802.11
protocol: design and performance
evaluation of an adaptive backoff
mechanism”, IEEE journal on selected
areas in communications, September 2000,
vol. 18, issue 9, pp 1774-1786
[25] R. Bruno, M. Canti, E. Gregori, “Mesh
networks: commodity multihop adhoc
networks”, IEEE communications
magazine, March 2005, vol. 43, issue 3, pp
123-131
[26] S.I. E.G. Nikolovzou, I.S. Venieris, “QoS
issues in the converged 3G wireless and
wired networks”, IEEE communications
magazine, August 2002, vol. 40, issue 8,
pp 43-53
[27] C.S. Raghavendra, M. Gerla, A. Avizienis,
“Reliable loop topologies for large local
computer networks”, IEEE transactions on
computers, January 1985, vol.c-34, issue
1, pp 46-55
[28] Chee-Yee Chang, S.P. Kumar, Booz Allen
Hamilton, “Sensor networks: evolution,
opportunities and challenges”, Proceedings
of the IEEE, August 2003, vol. 91, issue 8,
pp 1247-1256
[29] A. Adya, P. Bahl, J. Padhye, A. Wolwan,
Lidong Zhou, “A multi-radio unification
protocol for IEEE 802.11 wireless
networks”, Proceedings of first
international conference on broadband
networks, October 2004, pp 344-354
[30] J. Comez, A.T. Campbell, M. Naghshinlh,
C. Bisdikian, “Conserving transmission
power in wireless adhoc networks”, Ninth
international conference on network
protocols, November 2001, pp 24-34

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO1O9-1

OFDM PAPR Reduction Based on Nonlinear Functions without
BER Degradation and Out-Of-Band Emission


Neelam
Dewangan &
Suchita
Chatterjee
Chattrapati Shivaji Institute Of Technology,Durg

MARCH
2011
Abstract

One of the challenging issues for Orthogonal
Frequency Division Multiplexing (OFDM)
system is its high Peak-to-Average Power Ra-
tio (PAPR) ,which makes system performance
very sensitive to non-linear distortions.In this
paer we reduce the OFDM PAPR by using
a non-linear function without degrading the
BER and without any OOB emisssion.Using
bussagang theorem ,we found that the uncor-
related component resulting in the non-linear
process of input signal amplitude is the
component for PAPR reduction.Based on
this result,we used the unused sub-carriers
embedded in the spectrum of multi-carrier
standards to carry the component for PAPR
reduction ,since the compoinent for PAPR
OFDM has been chosen for high data rate
communications and has been widely deployed
in many wireless communication standards
such as Digital Video Broadcasting (DVB)
and based mobile worldwide interoperability
for microwave access (mobile WiMAX) based
on OFDM access technology [3]. However,
still some challenging issues remain unresolved
in the design of the OFDM systems. a major
disadvantage of OFDM is the inherent high
peak-toaverage power ratio (PAPR) due
to its approximately Gaussian distributed
waveform generated by the summation of
many subcarrier-modulated signals. There-
fore, the OFDM receivers detection efficiency
is very sensitive to the nonlinear devices
used in its signal processing loop, such as
Digital-to-Analog Converter (DAC) and High
Power Amplifier (HPA), which may severely
impair system performance due to induced
spectral regrowth and detection efficiency
degradation. As one of characteristics of
the PAPR, the distribution of PAPR, which
bears stochastic characteristics in OFDM
systems, often can be expressed in terms
of Complementary Cumulative Distribution
Function (CCDF). Recently, some researchers
have reported on determination of the PAPR
distribution based on different theoretics and
hypotheses . Moreover, various approaches
also have been proposed to reduce the PAPR
is carried by unused sub-carriers there
BER degradation and OOB emission .
is no
I. INTRODUCTION
As an attractive technology for wireless
communications, Orthogonal Frequency Divi-
sion Multiplexing (OFDM), which is one of
multi-carrier modulation (MCM) techniques,
offers a considerable high spectral efficiency,
multipath delay spread tolerance, immunity
to the frequency selective fading channels
and power efficiency [1], [2]. As a result, including clipping, coding schemes ,phase op-
























CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO1O9-2

timization,nonlinear companding transforms
, Tone Reservation (TR) and Tone Injection,
constellation shaping , Partial Transmission
Sequence (PTS) and Selective Mapping (SLM)
and other techniques such as pre-scrambles.
An effective PAPR reduction technique should
be given the best tradeoff between the capacity
of PAPR reduction and transmission power,
data rate loss, implementation complexity and
Bit-Error-Ratio (BER) performance etc. In
this paper we focus on PAPR reduction tech-
niques based on nonlinear functions. Two well
known examples are clipping techniques which
use a clipping function for PAPR reduction
and companding techniques (a) which use a
compression function at the transmitter side
for PAPR reduction. This paper propose to
reduce the OFDM PAPR by using a nonlinear
function called function for PAPR reduction
without degrading the BER and without any
OOB power.
N −1
1 2π
s
n
= √
X
X
k
exp j nk, 0 ≤ n ≤ N L − 1
N L
N
k=0
(2)
When L = 1, the above equation reduces to
the Nyquist rate sampling case. Eq.(2) can
be implemented by using a length(NL) IFFT
operation with the input vector
X ext = {X
0
, ....., X
N/2−1
, 0, 0, 0, X
N/2
, ...., X
N −1
}
Thus, X
ext is extended from X by using
the so-called zeropadding scheme, i.e., by
inserting (L-1)N zeros in the middle of X.
Suppose that the input data streams is
statistically independent and identically dis-
tributed (i.i.d.), i.e. the real part <{s(t)}
and imaginary part ={s(t)} are uncorre-
lated and orthogonal. Therefore, based
on the central limit theorem, when N is
considerably large, the distribution of both
<{s(t)} and ={s(t)} approaches Gaussian
distribution with zero mean and variance
II. OFDM SYSTEM MODEL
Let a block of N symbols X = {X
k
, k =

σ
2
= E[| <{s(t)} |
2
| ={s(t)} |
2
] , where Es
0, 1, 2, ....., N − 1} is formed with each sym-
is the expected value of s. In other words,
bol modulating one of a set of subcarriers
{f
k
, k = 0, 1, 2, 3, ....., N − 1},where N is the
number of subcarriers. The N subcarriers are
chosen to be orthogonal, that is, f
k
= k/T
s

,and T
s
is the tome duration of the OFDM
symbol. Therefore, the baseband time-domain
signal s(t) is
OFDM signals with large N become Gaussian
distributed with Probability Density Function
(PDF) as
P (s(t)) =

1
exp −(s(t)
2
)/2σ
2

2πσ
where,σ
2
is the variance of s(t).
Moreover, the Rayleigh nature of original
OFDM signals amplitude can be gotten and
its PDF can be expressed as
N −1
1
X
s(t) = √
N
X
k
exp(j2πf
k
t), 0 ≤ t ≤ T
s

k=0
(1)
P (r) = 2r exp −r
2


where r is the amplitude of OFDM sig-
nals.
By sampling s(t) defined in Eq.(1) at
frequency fs = NL/Ts,where L is the oversam-
pling factor, the discrete-time OFDM symbol
can be written as
III. DEFINITION OF PAPR







CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO1O9-3



as,
1) Continuous-time PAPR
In general, the PAPR of OFDM signals is de-
2
max
0≤n≤N L−1
|s(n)|
P AP R[s] =
E|s(n)|
2

(4)
fined as the ratio between the maximum
stantaneous power and its average power
in-
where denotes the expectation operator.
2
IV. CHARACTERIZATION OF PAPR
max
t [0,T
s
]
|s(t)|

P
s

P AP R[s] = (3)
where P
s
is the average power of and it can
be computed in the frequency domain because
Inverse Fast Fourier Transform (IFFT) is a
(scaled) unitary transformation.
2) Discrete-time PAPR
The PAPR of the discrete time sequences
typically determines the complexity of the
digital circuitry in terms of the number of
bits necessary to achieve a desired signal to
quantization noise for both the digital oper-
ation and the DAC. However, we are often
more concerned with reducing the PAPR
of the continuous-time signals in practice,
since the cost and power dissipation of the
analog components often dominate. To better
approximate the PAPR of continuous-time
OFDM signals, the OFDM signals sam-
ples are obtained by times oversampling.
L-times oversampled time-domain samples
are NL-point IFFT of the data block with
(L-1)N zero-padding. Therefore, the oversam-
pled IFFT output can be expressed as in eq(2).
Figure 1: Distribution of PAPR of OFDM sig-
nal samples oversampled by different L
REDUCTION SCHEME BASED ON NON-
LINEAR FUNCTION
Fig. 1 shows the distribution of the PAPR
of the OFDM signals with N=256 and
L=1,2,4,16. As shown, the largest PAPR in-
crease happens from to L=1 to L=2. However,
the PAPR does not increase significantly after
L=4. It has shown that L ≥ 4 is sufficient
to get accurate PAPR results . The PAPR
computed from the L-times oversampled time
domain OFDM signal samples can be defined
A. Definition and Remark : Let f (.), be
a function and let s(t) = |s(t)| exp jϕ(t) be
a
is
a
single or multicarrier signal, where ϕ(t)
the phase of s(t). The function f (.) is
function of s(t) PAPR reduction if the
PAPR of s(t) is strictly inferior to the PAPR
of s˜(t) = f [|s(t)|] exp jϕ(t), i.e., it exists
∆P AP R > 0 such as



CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO1O9-4



P AP R[s˜] + ∆P AP R = P AP R[s](indB),
C. PAPR reduction techniques based on
Nonlinear Functions in adding signal context where ∆P AP R is the gain in PAPR re-
duction.Indeed, a signal s(t) and its
scaled version have the same PAPR, i.e.,
P AP R[s] = P AP R[αs], where α,is a scalar.
The adding signal techniques consist of
reducing the envelope of OFDM signal by
adding a peak-reducing signal[5] just before
the HPA as shown in Fig.1 B. Proposition
The uncorrelated component resulting in non-
linear process of the input signal amplitude is
the component for PAPR reduction. Proof:
Let f (.) be a function of s(t) PAPR reduction,
s(t) can be a single or multicarrier signal.
(i) According to the above definition ,it
exists ∆P AP R > 0 such us
P AP R[s˜] + ∆P AP R = P AP R[s](indB),
Figure 2: Adding signal scheme for PAPR re-
duction,where c(t) is the peak reducing signal
(ii) As f(.) is a nonlinear function (refer to
the remark), using the Bussgang decomposi-
tion[4], the signal s˜(t) = f [|s(t)|] exp jϕ(t)can
be written as The PAPR reduced signal is therefore
expressed by
s˜(t) = αs(t) + d(t),
s˜(t) = s(t) + c(t)
where α = R
s˜s
(0)/R
ss
(0). The compo-
nent d(t) is uncorrelated with the input signal
s(t), i.e.,R
sd
(τ ) = E[d(t + τ )s˜(t)] = 0.
from Eq and Eq , the peak reducing sig-
nal c(t) is expressed as
(iii) As, s(t) and its scaled version αs(t) have
the same PAPR, i.e., P AP R[s] = P AP R[αs],
and as
c(t) = (α − 1)s(t) + d(t)

from above Eq we can see that ,the peak
reducing signal depends on the uncorrelated
P AP R[αs+d]+∆P AP R = P AP R[s˜]+∆P AP R
component resulting in the nonlinear process
(5)
(6)

(7)
of the input signal amplitude.
= P AP R[s]

= P AP R[αs];
D. Properties of the peak-reducing sig-
nal
(iv) Therefore,
Using the statistical properties of uncor-
related component resulting in the bussgang
decomposition [6,7],we derive the statistical
properties of the peak-reducing signal.we have
P AP R[αs + d] < P AP R[αs]; (8)
so d(t) is the component for s(t)
reduction.
PAPR

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO1O9-5



the correlation function for c(t) is Here we use the unused subcarriers em-
bedded int he spectrum of multicarrier
standards to carry the peak-reducing signal
in order not to degrade the BER of system
and notto generate OOB distortions.Without
unused subcarriers of multicarrier standards
,PAPRcan be reeuced without BER degra-
dation but with OOB emission or without
OOB emission but with BER degradatoin. A
FFT/IFFT pair-based digital filtering to filter
the peak reducing signal in order to carry it
only on the unused subcarriers of multicarrier
standards. Let I and O be be the set of the
in band and out band indices respectively
R
cc
(τ ) = |α − 1|
2
R
ss
(τ ) + R
dd
(τ ) (9)
where ,R
dd
(τ ) is the correlation function for
the uncorrelated component d(t) and is given
by

R
dd
(τ ) =
X
b
n
[R
ss
(τ )/P
s
]
(
2n + 1)
n=1
(10)
and
r
2
1 1
Z
I
b
n
=
P

rf (r)p(r)L
n
(
p
)dr k

n + 1
k

s Dr s
and {i
0
, ...., i
N
r
−1
}
be the locations of the
and R the complement
(11)
c
unused subcarriers
whereD
r
= r : 0 ≤ r ≤ ∞ is the domain of
integration,p(r) is the probability density of
the input signal amplitude andL
I
(x) is the
of R in I. The forward FFT transforms c
n
back to the frequency domain.The discrete
frequency components of c
n
on the unused
subcarriers R are passed unchanged while the
data subcarriers R
c
and the OOB components
O are setted to zero, i.e,
(
C
k
, k <
n
languerre function expressed by
k
exp x
n!
x


L
I

n n+k
n
(x) =

(d/dx) (x exp −x)
(12)
The peak reducing signal power spectrum
density is expressed by

k
=
k (<
c
∪ O) 0,
bn
[S S
cc
(ϑ) = |α − 1|
2
S
ss
(ϑ) +
P

(ϑ) ⊗
ss 1 n=1
P
2n+1

s
.... ⊗
2n+1
S
ss
(ϑ)]
from above Eq it can be seen that c(t) is
distributed over a wider bandwidth than s(t).
In [6][7] it is shown that the uncorrelated
component resulting in bussgang decomposi-
tion is responsible to in-band and out-of-band
distortions. As c(t) = (α − 1)s(t) + d(t),the
peak reducing siganl is responsible for in-band
and out-of-band distortion. Thus a trade-off
must be done between PAPR reduction and
distortions generated.
Figure 3: Adding signal scheme for PAPR re-
duction including FFT/IRRT-based filter and
the iteration process ,where c(t) is the peak
reducing signal
V. PRINCIPLE OF PAPR REDUCTION
WITHOUT BER DEGRADATION AND
OOB EMISSION
The IFFT operation transforms
C
˜
k
, k = 0, ..., N L − 1 back to the time domain.
This results in the filtered peak-reducing





CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO1O9-6

signal c
n
at the output of the filter-based
FFT/IFFT. Because of R
T
R
c
= Φ, the BER
of the system is not degraded. Because of
amplitude of the signal. In [11],it is shown
that any form of clipping (in particular, the
classical clipping) can be formulated as an
adding signal technique as illustrated in Fig.1. C
˜
k
= 0 for k O, there is not OOB emission.
It is obvious that the filtering eliminates a
part of the peakreducing signal which causes
a peaks regrowth. In order to reduce as much
as possible the PAPR, the process of adding
signal computation followed by filtering must
be repeated several times. The new PAPR
reduction scheme including the filtering and
the iterations is illustrated in Fig.2 in discrete
time-domain.
B. The IEEE 802.11a/g standards based
WLAN system
In WLAN IEEE 802.11a/g standard IFFT
size (N) is 64. Out of these 64 subcarriers, 48
subcarriers are used for data, while 4 subcarri-
ers are used for pilots. The rest 12 subcarriers
are unused (null) subcarriers located at the
positions R = {0, 27, ......, 37} of the IFFT
input. Only these unused subcarriers shall be
utilized for WLAN PAPR reduction.
Fig.3 shows the IEEE 802.11a/g Standard
specifications: Subcarriers (data, pilots, un-
used) positions [9].
VI. ILLUSTRATION BY THE CLASSI-
CAL CLIPPING FUNCTION IN A WLAN
SYSTEM CONTEXT
There are many functions which can be
used for PAPR reduction such as clipping
,companding techniques and within compand-
ing there exist exponential companding , µ
-law compander,A-law compander, exponen-
tial companding and many more.In this paper
, we will discuss the classical clipping function
which is a function for PAPR reduction.This
function is used to reduce the PAPR of the
local-area-networks (WLAN) system based on
IEEE 802.11a/g standards under the condition
VI. CONCLUSION
OFDM is a very attractive technique for
wireless communication due to its spectrum
efficiency and channel robustness.One of the
serious drawback of OFDM system is that
the composite transmit signal can exhibit a
very high PAPR when the input sequences
are highly correlated.In this paper using
Bussagang theorem we have pointed out that
input signal when undergoes nonlinear process
results in an uncorrelated component which
is used as a component for PAPR reduction
and unused subcarriers in the spectrum of
uncorrelated component is used to carry
the component for PAPR reduction.Here
a conventional clipping technique has been
formulated as an adding signal technique
for PAPR reduction and has been used to
illustrate these assertions in a WLAN system
based on IEEE 802.11a/g standard.
of no BER degradation
emission.
and no OOB power
A. Clipping function
Using the conventional clipping technique [8]
to reduce OFDM PAPR, the output signal
s˜(t), in terms of the input signal s(t) is given
as follows: s˜(t) = f [|s(t)|] exp jϕ(t), where
ϕ(t) is the s(t) phase and f(.) is the clipping
function which is expressed as
(
r, r ≤ A
A, r ≥ A
f (r) =
REFERENCES
where A is the clipping threshold and r is the [1] Y.Wu and W. Y. Zou, ”Orthogonal frequency

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO1O9-7

division multiplexing: A multi-carrier modulation
scheme”, IEEE Trans. Consumer Electronics, vol.
41, no. 3, pp. 399, Aug. 1995.
[2] W. Y. Zou and Y. Wu, ”COFDM: An
overview, IEEE Trans. Broadcasting”, vol. 41, no.
1, pp. 18, Mar. 1995.
[3] T. Jiang, W. Xiang, H. H. Chen, and Q.
Ni, ”Multicast broadcasting services support in
OFDMA-based WiMAX systems”, IEEE Commu-
nications Magazine, vol. 45, no. 8, pp. 7886, Aug.
2007.
[4] J. Bussgang, ”Crosscorrelation function of
amplitude-distorted Gaussian signals”, Research
laboratory of electronics, Massachusetts Institute
of Technology, Cambridge. Technical Report 216,
(1952).
[5] D. Guel and J. Palicot, ”Clipping formu-
lated as an adding signal technique for OFDM
Peak Power Reduction”, IEEE Vehicular Technol-
ogy Conference, VTC-Spring, Barcelona, Spain,
26-29 April 2009.
[6] P. Banelli, G. Baruffa, and S. Cacopardi, ”Ef-
fects of HPA non linearity on frequency multiplexed
OFDM signals”, IEEE Trans. Broadcast., vol. 47,
pp. 123-136, June 2001.
[7] P. Banelli, and S. Cacopardi ”Theoretical
analysis and performance of OFDM signals in
nonlinear fading channels”, IEEE Transactions on
Wireless Communications, vol. 47, no. 2, pp. 284-
293, March 2003.
[8] X. Li and L. J. Cimini, Jr., ”Effects of clipping
and filtering on the performance of OFDM”, IEEE
Commun. Lett., vol. 2, pp. 131-133, May 1998.
[9] ”Wireless LAN Medium Access Control
(MAC) and Physical Layer (PHY)specifications”,
Adopted by the ISO/IEC and redesignated as
ISO/IEC 8802-11:1999/Amd 1:2000(E).


Analysis on Placement of Wavelength Converters in WDM p-Cycle Network






Abstract

The p-cycles can be described as preconfigured closed
protection paths in a mesh network. Like rings, the
protection path is preconnected in advance of any
failure and protection capacity is very fast and simple.
And like mesh p-cycle offers the high capacity
efficiency. The use of wavelength converter in p-cycle
enhances the capacity and rapid protection mechanism
for mesh-restorable networks. We can trade- off
between costs associated with the number of
wavelength converters required and the total spare
capacity needed for protection. The most important
thing is that the number of wave- length converters
can be greatly reduced, with placing wavelength
converters at the different points between a
transparent optical working path layer and a
corresponding set of single-wavelength p cycle
protection structures. Here we are comparing different
techniques named as sparse, partial and sparse-partial
wavelength conversion.

Index Terms - WDM Networks, Sparse Wavelength
conversion, Partial Wavelength conversion, p-cycles

1. Introduction

p-Cycles offer an attractive option for protection in
wavelength division multiplexed (WDM) transport
networks, by combining the benefits of both ring-
based protection and mesh-based restoration. In a
WDM network where wavelength conversion has
significant cost, however, the assignment of
wavelengths to working paths, and to protection
structures, also needs to be considered so that the
overall cost of wavelength conversions and capacity is
minimized. Without any (wavelength) converters,
each working path must have the same wavelength
from origin to destination node. In WDM networks
without wavelength conversion, situations can occur
where there are free channels, but blocking still occurs
because a common free wavelength cannot be found
for the entire path. Changing the wavelength along
the lightpath can prevent this. Deployment of
wavelength converters thus increases flexibility but
adds costs that can be quite high.
Therefore, the p-cycle concept in WDM
networks with the idea of requiring as few wavelength
converters as possible while avoiding any significant
penalty in required capacity due to such wavelength-
blocking effects is investigated. The new approaches
for an efficient configuration are developed and focus
on the two aspects of protection capacity efficiency
and the number of required wavelength converters is
given.

Here, the few terms used in the paper are defined as
follows:
Link: Individual wavelength channel between 2
adjacent nodes.
Span: Set of all links between 2 adjacent nodes.
Path: End to end connection described by a sequence
of links and nodes.
Wavelength path (WP): Path that consists of links
with the same wavelength and thus does not require
any wavelength converter.
Virtual Wavelength path (VWP): Path that includes
a wavelength converter at each node.

2. The p-cycle concept


Fig.2.1. Simple mesh network with one p-cycle

The p-cycle concept, first presented in [1], is a recent
strategy to recover from network failures. In fig.2.1
the dashed line represents a p-cycle of a single unit of
Rupali Agarwal
roopali.ipec@gmail.com
Deptt. Of EC, BBD, Lucknow
Rohit Kumar
rohit.hbti@rediffmail.com
Deptt. of EI, BBD, Lucknow

protection capacity. It provides a protection
mechanism for spans and for transiting traffic through
failed nodes. p-Cycles can be described as pre-
configured closed protection paths in a mesh network.
Cycle-oriented pre configuration remains
fundamentally a mesh restorable network technology
in terms of its capacity efficiency and in its functional
differences from self-healing rings. The concept is
based on the property of a ring to protect not only its
on cycle spans but also the straddling spans [1].


Fig.2.2. Failure of an on cycle scan


Fig.2.3. Failure of a straddling span

The p-cycle concept combines the merit of the two
basic protection strategies so far known: ring
protection and mesh restoration. On one hand, it
allows short recovery times of about 50-150 ms known
from protection ring structures. On the other hand it
has been repeatedly validated that under appropriate
design methods, it offers the capacity-efficiency which
is essentially as high as that of a span-restorable mesh
network.

3. Wavelength converters in p-cycle networks

When wavelength conversion is not available, a
lightpath must use the same wavelength on all the
links traversed in a WDM optical network. This
requirement is known as the wavelength continuity
constraint. On the other hand, if wavelength routers
are capable of wavelength conversion, an optical
signal may be converted from one wavelength to
another wavelength. In some previous work on -cycle
based protection, a path following the wavelength
continuity constraint is called a wavelength path (WP),
while a virtual wavelength path (VWP) is defined to be
a path that uses wavelength conversion at each node
on the path, and may have different wavelengths on
different links that the path traverses. Therefore, a WP
network has no wavelength conversion capabilities at
all, while a VWP network has full wavelength
conversion at every node, i.e., there are sufficient
converters at each node to convert any incoming
wavelength to any outgoing wavelength.

Previous research [3], [4] has shown that wavelength
conversion enables more efficient resource utilization,
and may reduce the lightpath blocking probability
significantly by resolving the wavelength conflicts of
lightpath routing. However, wavelength converters
should not be used arbitrarily due to their high costs,
and possible signal quality degradation incurred by
some types of converters. Therefore, a tradeoff
between the performance of a WDM network, and the
number of wavelength converters used exists whether
network protection provisioning is considered or not.

4. Wavelength conversion techniques

It has been demonstrated that a relatively small
number of converters is sufficient for networks to
achieve a certain level of acceptable blocking
performance. Such networks therefore only have
partial wavelength conversion capabilities. The
different techniques used for wavelength conversion
are as follows:
1. Partial wavelength conversion
2. Sparse wavelength conversion
3. Sparse partial wavelength conversion


4.1. Partial wavelength conversion

WDM networks add the aspect of wavelength
assignment to the p-cycle protection concept. Working
and p-cycle links may have different transmission
wavelengths. It is not sufficient to provide only
enough p-cycle protection paths for the working links
but its also have to consider the wavelengths of the
working paths failed in any given scenario and the
wavelength(s) on which a p-cycle is established. If the
protection path for a working link is allocated at a
different wavelength, the wavelength must be
converted to access the p-cycle in case of a failure.
Otherwise, all working link and protection path
arrangements must be coordinated to have the same
wavelengths.
In nodes with partial wavelength conversion,
only a limited number of incoming lightpaths can
change to a different wavelength on the outgoing link.
Fig 4.1(a) and 4.1(b) depict an optical cross-connect
node with a shared pool of C wavelength converters. If
there is no wavelength conversion required, an
incoming lightpath will be directed to the appropriate
output port of an outgoing fiber or to the local access.
Otherwise the lightpath can pass through an available
converter in the converter pool. As in a WP network
there are no converters, and in a highly equipped
VWP network each node has at least m*n converters,
where m is no. of input and output fibers in the optical
switch and n is the total no. of wavelengths in
wavelength mux and demux.
With this type of node architecture in mind, it
is considered to have two basic alternatives to provide
p-cycles in WDM networks with partial wavelength
conversion. One employs WP p-cycles with converters
used for WP working paths to access them, and the
other is based on VWP p-cycles. The first idea is that
in order to provide for protection without requiring a
set of p-cycles dedicated to every wavelength, but
while using as few converters as possible overall, it
may be efficient to associate converters only with the
p-cycle access points-leaving the working paths to be
implemented in a pure WP manner incurring no
converter costs. Although in practice the converters
may be available for working paths as well, it is
assumed for research comparisons on the basic idea
that the converters are used for protection paths only.
One reason for this is also that protection paths
typically need more flexibility in the wavelength
selection to be able to protect many working
wavelengths. A built-in advantage is also that because
a failed working link is replaced by a longer protection
path, the converters (which are typically o-e-o
regenerative circuits) can ensure that the signal quality
of affected paths is not degraded by the protection re-
routing. In this basic approach, after detecting a span
failure the failure-adjacent nodes switch the working
link to a predetermined protection path on a p-cycle
and, if necessary, convert the wavelength of the
working link to that of the p-cycle at the access point.


Fig.4.1 (a). A wavelength converter with full wavelength
conversion

Fig 4.1 (b): A wavelength converter with partial
wavelength conversion

Advantages of partial wavelength conversion:
1. No. of wavelength converters is reduced.
2. Blocking probability of calls is reduced i.e. success
probability is increased.

4.2. Sparse wavelength conversion


To achieve the full wavelength conversion capability,
i.e. the input wavelength can be converted into any
output wavelength; a simple method is to use a
converter per wavelength per port in a dedicated
manner. However, from commercial point of view this
method is not cost effective as the cost of wavelength
converters is quite high and also the quality of signal
is also degraded. So it is recommended to use the
minimum no of converters. To reduce the usage of
wavelength converters, there are two approaches. In
one approach, sharing of wavelength converters is
done through a switch as discussed in the aforesaid
approach. Or in another approach, wavelength
converters are allocated to only few of the nodes in the
network, i.e. some of the nodes possess the wavelength
conversion capability while others do not. This refers
to sparse wavelength conversion. Also the sparse
wavelength conversion fulfils the requirement of
wavelength conflict rule. According to wavelength
conflict rule, the number of wavelengths required in a
WDM network is at least equal to the maximal
number of channels over a fiber (called maximal link
load) in the network. By placing wavelength
converters at some nodes in the network (i.e. sparse
wavelength conversion), the number of wavelengths
needed can be made equal to the maximal link load.
The converter allocation principle is explained with
the help of figure 4,4(a) and 4(b). In figure 4 there are
two working paths A-B-C and B-A-E and the p-cycle
configuration has been completed in the network. In
the figure, the shaded nodes indicate the places where
converters are needed and the numbers associated with
each edge represent the wavelengths assigned to each
link passed through the working paths and p-cycles.


Fig 4: P-cycle configuration in two working paths network



When span (B − C) fails, as shown in the figure
4(a), two converters are needed at node B for the
working path (B − C) to access the p-cycle (converting
wavelength from 0 to 2 and vice versa) as the path is
bidirectional and one converter is needed in each
direction. If span (A −E) fails, similarly, we need not
only two converters at node E for the second working
path to access the p-cycle, but also two converters at
node B and C, respectively, to transmit the on-cycle
traffic (converting wavelength from 2 to 1 and then to
2), as shown in fig 4.2(b). In the example, node B
needs two converters for working path (A − C) to
access the p-cycle in the case of span (B − C) failure,
and two converters for on-cycle wavelength
conversion in the case of span (A −E) failure. Here it
may be noticed that to minimize the number of
converters, only two converters are needed at node B
to protect spans (B − C) and (A − E) since at most one
of the spans may fail at a time in the single failure
scenario. Thus in case of sparse partial wavelength
conversion, the phenomenon of converter sharing
comes in the picture. As per the simulation results [2],
the proposed approach significantly outperforms the
approach for WP networks in terms of protection cost
and can obtain the optimal performance as achieved
by the approach for VWP networks, but requires fewer
wavelength conversion sites and fewer wavelength
converters.

Fig 4.2(a): span BC fails




Fig 4.2(b): Span AE fails



3.3 Sparse partial wavelength conversion
A special case of partial wavelength conversion is
sparse partial wavelength conversion in which only a
subset of network nodes is having the wavelength
conversion capacity and also the nodes are not fully
wavelength convertible. This network architecture can
significantly save the number of wavelength
converters, yet achieving excellent blocking
performance. Theoretical and simulation results[7]
indicate that, the performance of a wavelength-routed
WDM network with only 1-5% of wavelength
conversion capability is very close to that with Full-
Complete Wavelength Conversion capability. Actually
the sparse partial wavelength conversion (SPWC)
technique combines the benefits of partial wavelength
conversion and sparse wavelength conversion. There
are two kinds of nodes in the network: common
wavelength routers without wavelength conversion
capability, and WCRs with partial wavelength
conversion capability. By using sparse conversion and
partial conversion together, only a small number of
wavelength converters are needed to achieve
comparable performance as full-complete wavelength
conversion. And it only requires that a small fraction
of wavelength routers be replaced with WCRs, which
is very flexible for the network carriers to migrate the
existing network to support wavelength conversion.
The SPWC technique can be understood by the
following explanation-




Upon arrival of a lightpath request, if there is any link
in the selected route has no free wavelength, we have
to block this request. Otherwise, we first try to find a
common free wavelength on all the links along the
selected path. If there is no common free wavelength,
we will check whether wavelength converters can
help[8]. A lightpath is divided into several segments
by the intermediate WCRs which currently have free
converters, as shown in Fig. 5. Notice that, a WCR
can not provide conversion if its wavelength
converters have all been allocated. Each segment still
suffers the wavelength continuity constraint because
there are no WCRs in a segment (except the two end
nodes of the segment). The lightpath can be set up
successfully if and only if every segment has common
free wavelength(s). So we have to check whether there
exist common free wavelengths for each segment
individually. Wavelength converters will be allocated
if necessary. Once the lightpath is terminated, the
allocated converters will also be released. One
advantage of the sparse-partial wavelength conversion
is its flexibility for the network carriers to install
WCRs gradually. From the performance analysis [9],
it is shown that the performance in terms of blocking
probability is very much similar for sparse partial
wavelength conversion and partial wavelength
conversion, if the total no of wavelength converters is
not very large.

5. Conclusion
Wavelength conversion has been shown as
one of the key techniques that can improve the
blocking performance in a wavelength-routed all-
optical network. Also we know that the wavelength
converters are very expensive. To make effective use
of the limited number of wavelength converters
different techniques are used. The different techniques
analysed in this paper are partial wavelength
conversion, sparse wavelength conversion and sparse
partial wavelength conversion. It has been observed
that wavelength conversion can decrease the blocking
probability by a large margin. The p-cycle concept is a
recent strategy to recover from network failures. So
the different techniques of wavelength conversion are
analysed in p-cycle network. Among all the three, the
partial wavelength conversion technique is best as it
comprises of the benefits of both partial wavelength
conversion and sparse wavelength conversion. But if
we add more wavelength converters into the
wavelength convertible routes, the performance of
full-partial wavelength conversion can be the same as
full-complete wavelength conversion.

6. References

[1]. W. D. Grover and D. Stamatelakis, “Cycle-
oriented distributed preconfiguration: ring-like speed
with mesh-like capacity for self-planning network
restoration,” in Proc. IEEE International Conf.
Commun.(ICC’98), Atlanta, GA, June 1998, pp. 537–
543.
[2] D.A. Schupke, C.G. Gruber, and A. Autenrieth.
“Optimal configuration of p-cycles in WDM
networks”. Proc. of IEEE International Conference on
Communications (ICC), New York, USA, 5:2761–
2765, April/May 2002.
[3] D.A. Schupke, M.C. Scheffel, and W.D. Grover.
“Configuration of p-cycles in WDM networks with
partial wavelength conversion”.
Photonic Network Communications, pages 239–252,
June 2003.
[4] D. A. Schupke, C. G. Gruber, and A. Autenrieth,
“Optimal configuration of p-cycles in WDM
networks,” in Proc. of IEEE International Conference
on Communications (ICC), New York, April/May
2002, vol. 5, pp. 2761–2765.
[5] Tianjian Li, Student member, IEEE, and Bin
Wang, member IEEE, “ An Optimal p-cycle based
protection in WDM optical networks with sparse
partial wavelength conversion”.
[6]. Tianjian Li, Student member, IEEE, and Bin
Wang, member IEEE, “Optimal Configuration of p-
Cycles in WDM Optical Networks with Sparse
Wavelength Conversion”
[7]. C. M. Assi, A. A. Shami, M. A. Ali, Z. Zhang,
and X. Liu, “Impact of Wavelength Converters on the
Performance of Optical Networks,” SPIE/Kluwer
Optical Networks Magazine, Vol. 3, No. 2, pages 22-
30, March/April 2002.
[8]. X. Chu, B. Li, and I. Chlamtac, “Wavelength
Converter Placement under Different RWA Algorithm
in Wavelength-Routed All-Optical Networks,” IEEE
Transactions on Communications, Vol. 51, No. 4,
pages 607-617, April 2003.
[9]. Xiaowen Chu, Jiangchuan Liu, Zhensheng Zhang
“Analysis of Sparse-Partial Wavelength Conversion in
Wavelength-Routed WDM Networks” IEEE
INFOCOM 2004

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-1

COMPARATIVE ANALYSIS OF WIRELESS
NETWORKS USING OPNET
Yaduvir Singh
1
, Vikas Gupta
2
, Rahul Malhotra
3

1
Thapar University, Patiala, Punjab

2,3
Adesh Group of Institutions, Sri Muktsar Sahib
1
yad_pra@yahoo.com
3
blessurahul@gmail.com

Abstract: - The traditional wired transmission medium poses
constraints like mobility and extensive cabling. As the networks are
being upgraded from scratch all over the world, network planning
is becoming all the more important. Computing the viability and
performance of computer networks in real can be very expensive
and painstaking task. To ease and comfort the process of
estimating and predicting a network design, simulation and
modeling techniques are widely used and put into practice. Among
the variety of simulation tools available OPNET was used for the
purpose of modeling and simulation. In wireless networks, the
metrics like delay, retransmission attempts and throughput have
been estimated with varying number of nodes, physical
characteristic, buffer size, data rate, RTS threshold and
Fragmentation thresholds.
Keywords: OPNET, Buffer size, RTS, FTS, Data rate.
I. INTRODUCTION
Due to the various drawbacks of wired LANs like
extensive cabling and immobility etc., the wireless technology
gained momentum. Wireless communication technologies
employ infrared, spread spectrum and microwave radio
transmission techniques with varying data rates. The demand of
wireless LAN has increased over a span of time because of its
comparative simplicity, flexibility, high rate access and low
cost. The wireless network infrastructure is useful to provide
accessibility in rough terrains and even rural areas where
establishing wired infrastructure is difficult.

1.1 Issues in Wireless Network
Wireless technology provides convenience and advantages like
easy of mobility, scalability and flexibility but it has certain
downfalls like:
Speed: The speed of the wireless networks (ranging from
2Mbps to more than 100 Mbps according to the IEEE 802.11n
standards) is comparatively less than the wired LANs (available
in Gbps). The data rate decreases with the increase in the
number of nodes.
Range: The devices operate within a limited distance from an
access point. The distance between the devices is determined by
the standard used. It is distance between the buildings and other
obstacles between the access point and the user.
Cost: The wireless connecting devices are more costly as
compared to connecting devices used in wired networks.
Reliability: The wireless networks are subject to interference
and can thus pose a problem in the administration of Wireless
Infrastructure.
Security: Technically wired LANs are more secure than
WLANs. Since wireless signals are transmitted through the air,
they can be captured by devices outside the network. However,
the majority of wireless LANs today protects their data with the
Wired Equivalent Privacy (WEP) encryption standard or Wi-Fi
Protected Access (WAP) which makes wireless communications
almost as safe as wired ones in homes.
Bit Error Rate (BER): The wireless network’s media is error
prone hence its BER is higher than the wired LANs.
Carrier Sensing: Carrier sensing is difficult in wireless
networks because a station is incapable of listening to its own
transmissions.
Hidden Terminal Problem: The hidden terminals decrease the
performance of the wireless LANs.
II. PERFORMANCE ANALYSIS
Here, the simulation scenario of the wireless LAN has
been described in figure 2.1. A snapshot of the wireless LAN
models is shown here:

Figure 1 Network Model of the Wireless LAN
For performance analysis of wireless networks, following
attributes can be considered:
 RTS Threshold (bytes)
 Physical Characteristics
 Data rate
 Fragmentation Threshold (bytes)
 Buffer Size
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-2

These attributes have been described below:
RTS/CTS control enable/disable: To reduce the throughput
reduction owing to hidden stations RTS/CTS exchange is used.
This mechanism is used for large packets to improve the system
performance.
Physical Characteristics: Based on the value of this attribute,
which determines the physical layer technology in use, the
WLAN MAC will configure the values of the protocol
parameters.
Data Rate: Specifies the data rate that will be used by the MAC
for the transmission of the data frames via physical layer. The
set of supported data rates depending on the deployed physical
layer technology are specified in IEEE's 802.11, 802.11a,
802.11b and 802.11g standards. The value of the sibling
attribute "Physical Characteristics" determines the deployed
physical layer technology, and consequently, the set of the data
rate values that can be configured under this attribute.
Fragmentation Threshold: The fragmentation allows a large
MSDU to be divided into several smaller data frames.
Fragmentation allows early detection of errors and hence
reduces the retransmissions and the duration of time for which
channel is occupied when fragments collide.
Buffer Size (bits): Specifies the maximum size of the higher
layer data buffer in bits. Once the buffer limit is reached, the
data packets arrived from higher layer will be discarded until
some packets are removed from the buffer so that the buffer has
some free space to store these new packets.
The rest of the parameters can be set to constant or standard
values.
Case I: Performance Analysis for varying Physical
characteristics:
Table 1 Performance Analysis scenarios for varying physical characteristics
Parameters Case 1.1 Case 1. 2 Case 1.3 Case 1.4
RTS Threshold None None None None
Fragmentation
Threshold
None None None None
Data Rate 2 Mbps 2 Mbps 2 Mbps 2 Mbps
Physical
Characteristics
Frequency
Hopping
Direct
Sequence
Infra
Red
ERP
(802.11g)
Buffer Size 256000 256000 256000 256000
The scenarios mentioned in Table 1 presents a
performance analysis under influence of the physical layer type
on the IEEE 802.11 network performance.

Figure 2 Media Access Delay for varying physical characteristics
During the investigation IEEE 802.11 network
performance using DSSS, FHSS, IR and ERP layers has been
compared. The Figures 2 showing delay characteristics, Figure 3
showing retransmission characteristics and Figure 4 showing
throughput are used for investigation of scenarios with varying
type of Physical characteristics. The Physical characteristics
used are Frequency Hopping, Direct Sequence, Infra Red and
Extended rate PHY (802.11g). Based on the value of this
attribute, which determines the physical layer technology in use,
the WLAN MAC will configure the values of the following
protocols parameters as indicated in the IEEE 802.11 WLAN
standard: SIFS time, SLOT time, Minimum contention window
size, Maximum contention window size and any other parameter
value derived from the values of these parameters (like DIFS).
Table 2 The selected protocol parameter values for physical layers [IEEE
802.11, 2003]
Parameter
/ Physical
layer type
IR DSSS FHSS
ERP
802.11g
DIFS 23 µs 50 µs
128
µs
28/50
µs
PIFS 15 µs 30 µs 78 µs
19/30
µs
SIFS 6 µs 10 µs 28 µs 10 µs
Slot time 8 µs 20 µs 50 µs 9/20 µs
Length of
physical
layer
preamble
16 µs(1Mbps)
20 µs (2 Mbps)
144
bits
96 µs 72 bits
Minimum
number of
slots- CWmin
63 31 15 67.5/15
Maximum
number of
slots-
CWmax
1023 1023 1023 1023
The inter-frame spaces (xIFS) are compared. The
smallest values are defined for the IR layer. The longest times
are proposed for the FHSS layer. The smallest values are
defined for the IR layer. The longest times are proposed for the
FHSS layer. The figure 2 shows lower delays for IR and ERP as
compared to the higher delays for DSSS and FHSS. For each
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-3

packet the delay is recorded when the packet is sent to the
physical layer for the first time. The smaller Slot Time and SIFS
values can decrease the average media access delay, and hence
improve the performance of the wireless network. The increase
in delay leads to retransmissions as shown in Figure 3. Total
number of retransmission attempts by all WLAN MACs in the
network occurs until either packet is successfully transmitted or
it is discarded as a result of reaching the short or long retry limit.
The retransmission attempt counts recorded under this statistic
also include retry count increments due to internal collisions. A
higher level of offered load brings a large number of collisions.

Figure 3 Retransmission Attempts for varying physical characteristics
The lack of the RTS/CTS mechanism causes a high
level of losses that arises from a large number of collisions of
relative large data frames. When there is a collision, the
contention window size is doubled, until a maximum value:
CW
max
is reached. Figure 4 represents the total number of bits
forwarded from wireless LAN layers to higher layers in all
nodes of the wireless network.

Figure 4 Throughput (bits/sec) for varying physical characteristics
The investigations show that the network attains the
maximum throughput using IR layer. The worst results are
achieved when IEEE 802.11 protocol uses FHSS layer. But an
important thing to focus on is, that the throughput may vary
according to the type of the network modelled, the network
objects variation may occur in terms of number of stations, data
rate and type of network load too among certain other
parameters. The phenomenon of retransmissions as shown in
Figure 3 is due to heavy congestion in the network or failures,
which lead to an excess in maximum TCP retransmission
timeouts. The effect of change in physical characteristics on
throughput i.e. on the bit rate sent to the higher layer is shown in
Figure 4.
Case 2: Performance Analysis for varying number of
workstations
This section presents simulation analysis under the
influence of offered load, for ERP- 802.11g, in form of
increased number of stations. Throughput and Delay as a
function of offered load for different number of stations have
been realized. The network has been configured to 11 Mbps
medium capacity. All other parameters are according to the
standard specifications.
Table 3 Performance Analysis scenarios for varying number of workstations
Parameters Case 2.1 Case 2.2 Case 2.3
Number of
Workstations
3 6 12
RTS Threshold(bytes) None None None
Fragmentation
Threshold(bytes)
None None None
Data Rate (bps) 11 Mbps 11 Mbps 11Mbps
Physical Characteristics
Extended Rate
PHY
(802.11g)
Extended Rate
PHY (802.11g)
Extended Rate
PHY (802.11g)
Buffer Size (bits) 256000 256000 256000
The results of obtained simulations are presented in
Figure 5 – Figure 7. The average media access delay as a
function of offered load for different number of stations is
shown in Figure 5. The realized throughput as a function of
offered load for number of stations is shown in Figure 7.
A very low increase in the average delay for low load
offered network is shown in the Figure 7. The average media
access delay is in milliseconds. The Figure 8 also shows that the
increase in the number of stations has a large influence on the
transmission delay. The variations in the delay are notable as the
offered load increases with the increase in number of
workstations. The largest value of average media access delays
obtained is approximately 0.25s (for 12 workstations),
approximately 0.07s (for 6 workstations) and 0.03s (for 3
workstations).
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-4


Figure 5 Media Access Delay for varying number of workstations
The throughput/delay can vary due to collisions of data
packets. When a collision occurs, the received packets are
discarded without recovering any data. New transmissions of the
same packets must follow, possibly inducing secondary
collisions, so that more slots are used than transmitted packets.
Therefore, throughput decreases and delay can become
excessive. High delay variability is also due to retransmissions
at the link level. One of the reasons the retransmission occurs is,
when ACK are not available in time, due to heavy delay or
congestion in the network. Link-layer retransmissions induce
delays which do not conform to the assumptions on which the
transport protocol is based. This causes undesired TCP control
actions which reduce throughput. Through simulations, it has
been found that the effective throughput of wireless networks
decreases as the retransmission limit increases as shown in
Figure 5.

Figure 6 Retransmission Attempts for varying number of workstations
An analysis of the presented results in figure 5 (shows
node statistics for throughput) allows drawing the conclusions
regarding the impact of number of stations on throughput.

Figure 7 Throughput for varying number of workstations (node level)
The analysis showed that the throughput (bits/sec) at
the client node reduces with the increase in the number of
workstations in the network. The statistic Throughput (bits/sec)
–a Node Statistics: means the total data traffic in bits/sec
successfully received and forwarded to the higher layer by the
WLAN MAC. This statistic does not include the data frames
that are 1) unicast frames addressed to another MAC, 2)
duplicates of previously received frames, and 3) incomplete,
meaning that not all the fragments of the frame were received
within a certain time, so that the received fragments had to be
discarded without fully reassembling the higher layer packet.
The network load applied increases with increase in number of
stations. The retransmission attempts and hence the delay in the
increased number of nodes is also more. The reduction of the
maximum of realized throughput with the increased number of
contending stations is very characteristic. A higher level of
offered load brings a large number of collisions. The lack of
RTS/CTS mechanism causes a high level of losses that arise
from a large number of relative large data frames. The graph in
Figure 7 shows that increase in delay occurs due to the offered
load, lower is the throughput for the particular scenario. Another
Simulation scenario focuses on such methods like increasing the
buffer size or reducing the network load by providing restricted
access to the users.
Case 3 Performance Analysis for varying Buffer Size
Buffer size (bits) specifies the maximum size of the
higher layer data buffer in bits. Once the buffer limit is reached,
the data packets arrived from higher layer will be discarded until
some packets are removed from the buffer, so that the buffer has
some free space to store these new packets.
Table 4 Performance analysis scenarios for varying Buffer size
Parameters Case 3.1 Case 3.2
Number of Workstations
12 12
RTS Threshold(bytes)
None None
Fragmentation Threshold(bytes)
None None
Data Rate (bps) 11 Mbps 11 Mbps
Physical Characteristics Direct Sequence Direct Sequence
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-5

Buffer Size (bits) 256000 1024000
The optimum size of buffer can stabilize the queue
size, the packet drop probability and hence the packet loss rate.
The benefits of stabilizing queues in a network are high resource
utilization. When the queue buffer appears to be congested the
packet discard probability increases. On the other hand, the
buffer overflow can be used to manage congestion.

Figure 8 Media Access Delay for varying buffer size
The buffer configuration defines the buffer size, the
maximum allocated bandwidth and minimum guaranteed
bandwidth. If an incoming flow suddenly becomes bursty, then
it is possible for the entire buffer space to be filled by this single
flow and other flows will not be serviced until the buffer is
emptied. If the buffer size is increased, (Figure 9) then the
number of retransmission attempts would be reduced. Also the
size of the queue will be decreased for larger buffer due to the
fact that the larger buffer will take less time to send the packets,
so the queue size will not build up continuously for larger
buffer. This shows the reduction in delay as shown in Figure 8.
The result of packet loss is the change of queue length. When
packet loss is relatively low, packets usually can be transmitted
without retransmission, so the queue length may be relatively
low. But when the packet loss is high, the MAC packet
retransmission (Figure 9) will prolong the delay of packet
(Figure 8). So, a small buffer size can increase packet drop rate
and hence change the queue length and thus impact the
throughput and delay. Hence, the buffer size has been increased,
to improve performance by reducing packet drop rate and thus
increasing throughput.

Figure 9 Retransmission Attempts for varying buffer size
The increase of packet discard rate can lead to the
decrease of throughput. This happens due to frequent
retransmissions of the MAC layer data packets when the packet
loss rate increases. But the packet loss may happen due to low
buffer size.

Figure 10 Throughput (bits/sec) for varying buffer size (Node level)
The analysis in figure 10 shows that if buffer size is
increased then the retransmission attempts would be reduced as
the size of the queue is decreased for a large buffer size. The
time to deliver the packets decreases, due to large buffer size.
The throughput always increases monotonically with the buffer
size, reaching a maximum above a threshold buffer size.
Case 4 Performance Analysis for varying Data Rates
This attribute specifies the data rate that will be used by
the MAC for the transmission of the data frames via physical
layer. The set of supported data rates depending on the deployed
physical layer technology are specified in IEEE's 802.11,
802.11a, 802.11b and 802.11g standards. The Wireless LAN
model in OPNET Modeler supports data transfer at 1, 2, 5.5, 11
Mbps. These data rates are modeled as the speed of the
transmitter and receiver connected to wireless network. A
station can transmit data only at the data rate specified by the
attribute. However, it can receive data at any rate. Four
scenarios are modeled as shown in Table 5. In these the data
rates have been varied while keeping the rest of parameters
same in all the scenarios
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-6

Table 5 Performance Analysis scenarios for varying Data Rates
Parameters Case 4.1 Case 4.2 Case 4.3 Case 4.4
Number of
Workstations
12 12 12 12
RTS
Threshold(bytes)
None None None None
Fragmentation
Threshold(bytes)
None None None None
Data Rate (bps) 1 Mbps 2 Mbps 5.5Mbps 11Mbps
Physical
Characteristics
ERP
(802.11g)
ERP
(802.11g)
ERP
(802.11g)
ERP
(802.11g)
Buffer Size (bits) 256000 256000 256000 256000
The Figures 11 shows the average media access delay
and figure 12 shows the average throughput for varying data
rates. The figure 11 shows that the delay is relatively more for
data rates ( 1Mbps), whereas it is less for scenarios with high
data rates (11 Mbps).

Figure 11 Media Access Delay for varying Data Rate
The high data rate provides more bandwidth, so the
data packet utilizes this bandwidth to be transmitted at high
speed. Thus the delay increases for low data rates
whereas it is less for high data rates (Figure 12).

Figure 12 Throughput (bits/sec) for varying Data Rate (Global)
From the above simulation scenarios it has been
observed that as we increase the data rate form 1 Mbps to 11
Mbps, there is increase in throughput. This is also predictable
from the theoretical viewpoint that as we increase the data rate,
the number of bits received increases. Also the media access
delay reduces with the increases in the data rate. It occurs with
increase in data rate the data will stay in the buffer for lesser
time and hence the number of retransmissions will decrease.
This indicates that that more number of packets can be delivered
safely with increase in data rate.
Case 5 Performance Analysis for varying RTS Threshold
RTS technique is used to reduce the probability of two
stations colliding with each other. The collisions occur because
they cannot hear each other due to the absence of medium
(wired link) between them. In this mechanism, a station wanting
to transmit a packet first transmits a short packet called RTS,
which includes the source, destination and duration between the
packet sent and the acknowledgement received. The destination
station responds (if the medium is free) with a response control
packet called CTS (clear to send), which includes the same
duration information. The mechanism reduces the probability of
a collision. The reduction in collision reduces the overhead of
collisions.
Table 6 Performance Analysis scenarios for varying RTS threshold
Parameters Case 5.1 Case 5.2 Case 5.3
Number of
Workstations
12 12 12
RTS
Threshold(bytes)
None 256 1024
Data Rate (bps) 1 Mbps 1 Mbps 1 Mbps
Physical
Characteristics
Frequency
Hopping
Frequency
Hopping
Frequency
Hopping
Buffer Size (bits)
256000 256000 256000
The RTS or the CTS overhead are short in size as
compared to the data packets to be transmitted. Moreover the
RTS and CTS frame size is controlled per station by a parameter
called RTS Threshold.

Figure 13 Media access delay for varying RTS
The increase in delay for high value of RTS threshold
shows the increase in delay in Figure 13. Though RTS helps to
reduce the collisions occuring due to hidden nodes but the delay
increases due to the additional overhead introduced by thehigh
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-7

value of RTS threshold . So, RTS threshold value must be
optimally chosen.

Figure 14 Retransmission Attempts for varying RTS
The number of retransmission occur with an increase in
number of collisions (Figure 14). The RTS standard reduces the
collisions, so, the retransmissions occuring due the RTS
threshold value 1024 bytes decreases as compared to the
threshold value of 256 bytes.

Figure 15 Throughput (bits/sec) for varying RTS (Global)
RTS/CTS are a four way handshaking technique for
packet transmissions. The above scenarios indicate that the
media access delay is increased when using the RTS/CTS
mechanism. It occurs because sending RTS frame and waiting to
receive CTS frame will take a certain period of time (while the
data waits in the transmission buffer). However, without using
RTS/CTS mechanism the data will be sent immediately once it
is ready to send. The probability of sending RTS increases in
case of 256 byte threshold rather than 1024 byte threshold and
correspondingly the throughput varies with the number of
reservations (Figure 15).
Case 6 Performance Analysis for varying Fragmentation
Threshold
Typical LAN protocols use packets several hundred
bytes long (the longest Ethernet packet could be up to 1518
bytes long). The mechanism of fragmentation divides the large
sized packets in to smaller packets called fragments. The need
of fragmentation i.e. smaller packets in a wireless LAN
environment occurs because of the high Bit Error Rate of a radio
link and the probability of a packet getting corrupted (it
increases with the packet size but if the data packet size is
smaller, the overhead to transmit the data packet is also less).
Table 7 Performance analysis scenarios for varying Fragmentation threshold
Parameters Case 6.1 Case 6.2 Case 6.3
Number of
Workstations
12 12 12
RTS
Threshold(bytes)
None None None
Fragmentation
Threshold(bytes)
None 256 1024
Data Rate (bps) 11Mbps 11 Mbps 11Mbps
Physical
Characteristics
Direct
Sequence
Direct
Sequence
Direct
Sequence
Buffer Size
(bits)
256000 256000 256000
So, the fragmentation/reassembly mechanism at the MAC Layer
has been used to deal with large sized packets. Table 7 shows
the various scenarios designed using varying fragmentation
thresholds. In this section average media access delay (sec)
shown in Figure 16 and average throughput shown in Figure 17
has been investigated .

Figure 16 Media Access Delay for varying fragmentation threshold
The investigations from Figure 16 reveal that the
average media access delay increases for scenario with FTS
value of 256 bytes. Also the value of average media access
delay is low for scenario with higher fragmentation threshold
(1024). This variation occurs because when the fragmentation
threshold is very small and the packet overhead is heavy. This
packet overhead deteriorates the wireless LAN performance,
causing increase in average media access delay.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-8


Figure 17 Throughput (bits/sec) for varying fragmentation threshold (Global)
Longer frames tend to have higher error rates therefore
the transmission of long frames increases both the number of
required retransmissions and the amount of data that is dropped.
To combat this inefficiency the 802.11 standards specify a
mechanism for fragmenting frames when their length is above a
user-defined threshold. Fragmentation is allowed in the IEEE
802.11b standard at the MAC layer; so that the fragment size
could vary from 100 to 2350 bytes. When a frame is fragmented
each segment is transmitted as if it were a separate frame, while
the fragments are identified at the destination using various
fields in the packet header. Packet fragmentation allows large
packets to be broken into smaller units when sent over the air,
which is useful in congested environments or when interference
is a factor, since larger packets have a better chance of being
corrupted. This technique reduces the need for retransmission in
many cases and thus improves overall wireless network
performance. The MAC layer is responsible for reassembling
fragments received, rendering the process transparent to higher
level protocols. This fragmentation is implemented using the
built-in SAR (segmentation and reassembly) package in OPNET
[2]. But fragmentation again increases the overall delay due to
the reassembling of multiple fragments at the destination. It also
increases the channel reservation time per data frame exchange.
Fragmentation threshold is thus an important parameter that
affects WLAN performance. It is used to improve the WLAN
performance when the media error rate is high. Packet
fragmentation for every data rate must be used to achieve
throughput optimization (Figure 17).
III. CONCLUSION
In this paper, performance analysis of wireless computer
networks has been done for improving the performance of
wireless LAN. The investigations reveal that the performance of
Wireless LAN can be improved by fine tuning and properly
choosing the WLAN parameters. For the tested simulation
scenarios the performance is observed to be better with infra-red
type physical characteristics, higher buffer size (1024Kb), and at
11 Mbps data rate. Increased throughput was also observed with
an RTS threshold of 256 bytes and fragmentation threshold of
1024 bytes in the simulated networks.
REFERENCES
[1] Muhammad Ahsan, “Access Methods (Contention & Token
Passing),”
http://www.suite101.com/article.cfm/laning_retired/44958
[2] Todd Lammle, “CCNATM Cisco Certified Network
Associate Study Guide, 5
th
Edition,2006: BPB
Publications”.
[3] Behrouz A Forouzan, “Data Communications and
Networking, 4
th
Edition, 2007: Tata McGraw Hill”.
[4] Sameh H. Ghwanmeh, “Wireless network performance
optimization using Opnet Modeller”, Information
Technology Journal 5(1): pp 18-24, 2006.
[5] ISO/IEC 8802-11: 1999, “Wireless LAN Medium Access
Control (MAC) and physical Layer (PHY) specifications,
Standard reaffirmed in June 2003", 1999.
[6] Bradley Mitchell,”Wireless Standards - 802.11b 802.11a
802.11g and 802.11n: The 802.11 family explained”,
available at
http://compnetworking.about.com/cs/wireless80211/a/aa80
211standard_2.htm
[7] Sarah Shaban, Dr. Hesham M.El Badawy, Prof. Dr. Attallah
Hashad, “ Performance Evaluation of the IEEE 802.11
Wireless LAN Standards”, in the proceedings of the World
Congress on Engineering-2008 , vol. I, July 2-4, 2008.
[8] A. Goldsmith, “Wireless Communications, 1
st
Edition,
August 2005: Cambridge University Press”.
[9] Manju Sharma and Manoj Kumar, “ HTTP and FTP
Statistics for wireless and wire-line network with and
without Redundancy based on OPNET” in the proceeding
of National Conference on Optical and Wireless
Communication (NCOW-2008)”, pp:70-75,September 27-
28,2008.
[10] Dr. Mohammad Hussain Ali and Manal Kadhim Odah,
“Simulation Study of802.11b DCF using OPNET
Simulator”, Eng. & Tech. Journal,
vol.27,No6,2009,pp:1108-1117,2009.
[11] Hafiz M. Asif, Md. Golam Kaosar, “Performance
Comparison of IP, ATM and MPLS Based Network Cores
Using OPNET,” in 1st IEEE International Conference on
Industrial & Information Systems (ICIIS 2006), 8-11
August, 2006, Sri Lanka.
[12] Manju Sharma, Manoj Kumar and Ajay K Sharma, "HTTP
and FTP Statistics for Wireless and Wire-line Network with
and without Load Balance Based on OPNET",Published in
International Journal of Information and Systems Sciences,
Volume 5,Number 1, pp. 112-125, 2009.
[13] Bradley Mitchell,” Wired vs. Wireless Networking:
Which is better - wired or wireless?” available at
http://compnetworking.about.com/cs/homenetworking/a/ho
mewiredless_3.htm
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COMO2O1-9

[14] Arti Sood, “Network Design by Using Opnet™ IT Guru
Academic Edition Software", Rivier Academic Journal,
2007, Vol. 3 (1).
[15] Dibyendu Shekhar, Hua Qin, Shivkumar Kalyanaraman,
Kalyan Kidambi, “Performance Optimization of TCP/IP
over Asymmetric Wired and Wireless Links” in the
proceeding of conference on Next Generation Wireless
Networks: Technologies, Protocols, Services and
Applications (EW-2002),February 25-28, 2002 – Florence,
Italy,2002.
[16] Guiomar Corral, Agustin Zaballos, Cesc Canet, “Proposal
of new challenge labs for the OPNET IT Guru Academic
Edition”, available at
www.salle.url.edu/~zaballos/opnet/OPNET2004a.pdf
[17] Born Schilling, “Qualitative Comparison of Network
Simulation Tools,” 2005, www.ipvs.uni-
stuttgart.de/abteilungen/vs/lehre/.../CUBUS.../schilling.pdf
[18] Xinjie Chang, “Network Simulations with OPNET,”
Proceedings of Winter Simulation Conference, 1999, pp.
307-314
[19] “Introduction to OPNET Simulator” available at
http://bolero.ics.uci.edu/~ypan/OPNET/Introduction to
OPNET simulator.pdf
[20] “OPNET IT Guru Academic Edition” available at
http://www.opnet.com/university_program/itguru_academic
_edition
[21] Ranjan Kaparti, “OPNET IT GURU: A Tool for
Networking Education”
http://www.opnet.com/university_program/teaching_with_o
pnet/textbooks_and_materials/materials/ITGAE_Tool_Ntw
rk_Ed.pdf.
[22] Dr. Reinhard Kuch, “Studienbrief 2: Simulation of
networks using OPNET ITGURU Academic Edition v
9.1”, version: 1.0, 26-04-2009.
[23] Marco Galeazzi, “Lezione 11: OPNET IT Guru Academic
Edition 9.1 A.A. 2007-2008 - OPNET ITGAE 9.1 (1/5)”
available at
http://www.tlc.deit.univpm.it/sistemi/AA2007_08/Lezione1
1.pdf
[24] “Opnet_Modeler_Manual”, available at
http://www.opnet.com
[25] T. Velmurugan, Himanshu Chandra and S. Balaji, “
Comparison of Queuing disciplines for Differentiated
Services using OPNET”, IEEE, ARTComm.2009.128, pp.
744-746,2009.Y



















CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO202-1

Comparative Analysis of Core Migration
Techniques in Wireless Ad Hoc Networks
Rahul Malhotra
1
, Reena Aggarwal
2
1, 2
Department of Electronics & Communication Engineering
Adesh Institute of Engineering & Technology, Faridkot (Pb.), India
{blessurahul@gmail.com, reena12aggarwal@gmail.com}

Abstract—A number of application level multicast protocols
have been proposed for core selection and core migration in
Mobile Ad Hoc Networks. Core migration is necessary to
minimize any disruptions on the transmission of data due to
the changes in tree structure and to achieve improvement in
the delivery of media streams in multicast group. In this
paper, core migration is performed on the varying size
network graph model. The nodes within this ad hoc
arrangement take on the values of the edge (W
e
) and node (W)
randomly. With the varying numbers of iterations done on the
core migration algorithm, every node and every edge picks up
different random values. The core migration in wireless ad
hoc network has been achieved by comparing the experimental
results based on the node weight and the edge weight.

Keywords—Core-based multicast, mobile ad hoc network, core
based tree, dynamic multicast membership, and core
migration
I. INTRODUCTION
In the past few years, a new wireless architecture has
been introduced that do not rely on any fixed infrastructure.
In this architecture, all nodes are mobile and no node plays
any special role. In fact, nodes reach other nodes they need
to communicate with using their neighbors as shown in
figure1. Nodes that are close to each other discover their
neighbors. When a node needs to communicate with
another node, it sends the traffic to its neighbors and these
neighbors pass it along towards their neighbors and so on.
This repeats until the destination of the traffic is reached
[6]. Such architecture requires that every node in the
network play the role of a router by being able to determine
the paths that packets need to take in order to reach their
destinations.

Fig. 1. Ad hoc network (client to client)

Mobile ad hoc network (MANET) is based on rapid
deployment of independent mobile users. The routers are
free to move randomly and organize themselves arbitrarily
thus; the network‟s wireless topology may change rapidly
and unpredictably [1]. Such a network may operate in a
stand-alone fashion, or may be connected to the Internet.
Multihop, mobility, large network size combined with
device heterogeneity, bandwidth, and battery power
constraints [19] make the design of adequate routing
protocols a major challenge. Multicast promises efficient
use of network bandwidth for multiparty communication by
allowing point-to-multipoint communication.
Recent [5] multicast protocols such as, Protocol
Independent Multicast (PIM) and Core Based Trees (CBT)
are based on the notion of group-shared trees. Multicasting
routing protocols generally build trees to deliver messages
to a multicast group. Delivering the information only
through edges belonging to the tree generates an efficient
form of multicast communication minimizing the amount of
the network resources as required with the unicast routing.
In PIM, a group-shared tree is rooted at a rendezvous point
(RP) similar to the CBT, which is rooted at the core node.
In a multicast tree, the core node maintains the flow of
traffic in the network. The [14] selection of this core node
influences the shape of the multicast tree influencing the
performance of the multicast tree with respect to the amount
of the delay experienced by leaves of the tree. Due to the
dynamically varying nature of the network topology, nodes
need to be migrated to the one that satisfies the specified
QoS metrics. Migration of the core node takes place after
selecting a core node from the set of candidate core nodes
that may be either the neighboring nodes or the nodes
adjacent to the neighboring nodes of the previous core
node.
In this paper [16] investigate the problem of finding a
best node in terms of maximum edge weight (distance
between two communicating nodes) and weight of
individual node. This implies the movement of the core
node on a hop-by-hop basis to reach an optimal location.
The rest of this paper is organized as follows: Section 2
discusses the basics of wireless ad-hoc network. Section 3
covers the need and the techniques of the core migration
followed by the results and conclusion the next two
sections.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO202-2

II. WIRELESS AD HOC NETWORKS
A. Network Model
A mobile ad hoc network can be modeled as a network
consisting of n identical mobile hosts (nodes). These mobile
hosts employ a packet radio network to communicate with
each other. Multiple nodes falling in the radio coverage area
of another node are said to be neighbors of this node and
they can simultaneously receive a message transmitted from
that node. The topology of the mobile ad hoc network is
dynamic. Such a network can be represented by an
undirected graph G = (V, E, W, W
e
), where V is the set of
nodes represented V = N
x
and E is the set of logical links
between neighboring nodes, W is the weight of each
corresponding node in a set of V nodes, and W
e
represents
the weight of the edge connecting the two nodes. It is
assumed that nodes leave and join the multicast group
arbitrarily.
B. Core Based Tree
The core-based tree protocol is a network layer multicast
routing protocol that builds and maintains a shared delivery
tree for the multicast routing. Core-based tree is suited to
inter- and intra-domain multicast routing in the Internet. To
establish the path between the sender and the receiver, core-
based tree may use a separate multicast routing table, or it
may use unicast routing. The core migration protocol
followed by the core selection protocol is designed to
construct core- based tree. The purpose of the core
migration algorithm is to locate that network node or a
router whose use as the core of the multicast group results
in the best-multicast tree with respect to the desired QoS
specified by multicast application. The core-based tree is
the shortest path tree rooted at some core node. The core
node is also referred to as center node or a Rendezvous
point [16], [22]. The core-based tree protocol was
developed for Internet [18]. It uses a shared multicast tree
rooted at core node, Rendezvous point or center node. It
uses a receiver-based tree construction. A node interested in
joining the multicast tree associated with a certain multicast
group sends a JOIN-REQUEST towards the core node of
that group. The JOIN-REQUEST is routed towards the core
node using the underlying unicast routing protocol. An
intermediate node receiving a JOIN-REQUEST simply
forwards it to the next node in the route towards the core
node unless it happens to be an established node in the
multicast tree in which case the node responds by sending
the JOIN-ACK to the node, which initiated the JOIN-
REQUEST. A node traversed by JOIN-ACK record the
parent (upstream) node and the child (downstream) node. In
effect, the JOIN-ACK grafts a tree branch from the node,
which responded with the JOIN-ACK, and the node, which
initiated the corresponding JOIN-REQUEST. A node
wishing to leave the multicast group simply sends a QUIT-
REQUEST towards the core node along its tree branch. A
node receiving a QUIT-REAUEST deletes the node along
which it received the QUIT-REQUEST from the children
list. Furthermore, it forwards the QUIT-REQUEST to its
parent node if its children list becomes empty and it itself is
not a multicast group node.
The CBT protocol provides mechanisms to deal with
link/node failures and the core node failure. Link failures
are detected by the exchange of periodic keep-alive
messages between the neighboring nodes. In case a node
does not receive a certain fixed number of consecutive
keep-alive messages from its parent node, it assumes that
the link to its parent node has failed. In case a non-core
node detects that its parent node or link to its parent node
has failed, it has two options for failure recovery: it can
attempt to rejoin the multicast tree by sending a JOIN-
REQUEST towards the core node and it can send a
FLUSH-TREE message downstream, thus allowing each
node in the sub-tree rooted at it to independently attempt to
reattach to the multicast tree [63]. The first option results in
the formation of loops in the multicast tree and, hence, the
CBT protocol provides a loop detection mechanism.
Core-based multicast routing provides a good mechanism
for scalable multicasting. However, migration of a core to
an “optimal” location is an essential task. In static networks,
core nodes may be pre-assigned and manually configured.
However, in mobile ad hoc networks this may not be
optimal since the topology often changes. Dynamic
selection and hence the migration of center nodes is
important for good performance. A good core node may be
one, which is at the center of the portion of a network which
spans all the multicast of the multicast group, which is the
member of the multicast group, and which is fairly stable
[18]. This core migration protocol is based on the edge
weight of current multicast tree. Core-based multicast tree,
as discussed about, have been effectively used in solving
the multicast routing problem in mobile ad hoc networks. In
this work, this technique has been used to find a core node
in which qualitative factor are of major significance with
the objective to extract and exploit the goodness of routing
and multicast strategies to reach at the best possible route in
mobile ad hoc networks. Performance of the core migration
method is generally measured using delay as a metric that is
the distance of the child (downstream) node from its parent
(upstream) node in the entire graph.
C. Dynamic Multicast Membership
Multicast group membership is dynamic, nodes are able
to join and leave the group at any time. One of the multicast
group members randomly assigns the responsibility of a
group leader viz. a core to the first node of a multicast
group which then starts broadcasting RREQs. Other group
members on receiving RREQs unicasts RREPs along the
shortest path to the core. Therefore the nodes are joined to
the tree when their corresponding RREPs are accepted
either by the existing members or by the core node.
III. FUNDAMENTALS OF CORE MIGRATION
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO202-3

A. Core Selection
Core selection is the problem of selecting the placement
of a core or cores in the domain for the purpose of
improving the performance of the tree(s) constructed
around these core(s). The core selection protocol is
designed to construct core-based trees using one of the
several different QoS metrics. The core selection algorithm
tries to find a router whose use as the core of the multicast
group results in an optimal multicast tree with respect to a
designed performance metric. The selection of the core is
based on the explicit as well as implicit core selection
algorithms where all network nodes of the multicast group
compute their weight functions and exchange weights
among themselves so as to select the node with minimum
weight as the core for the latter case, and for the former
case, the core monitors the delay by calculating the time
difference relative to the itself between the transmission of
a packet and their corresponding acknowledgements [28].
Core-based multicast routing scheme allows the migration
of the core node by selecting the candidate cores (the nodes
adjacent to the core node). The selection of the candidate
node is based on the heuristic of the comparison between
the current core and the set of the adjacent nodes. The node
with the significant better performance (in terms of the
distance between the adjacent nodes ad their individual
weight) than the current core is elected as the new core and
finally migrated. The migration of the core takes place for
the node having largest path delay and minimum node
weight, resulting into the optimal core selection and
migration. In a typical ad hoc environment, network hosts
work in a group to carry out a given task. Multicasting is
used for group-oriented communications.
These core selection techniques given in [23] used Core-
Based tree having a single node, or router, which acts as the
core of the tree, which enhances the scalability of the
multicast algorithm, also the tree creation was receiver
based. The core selection protocol was designed to
construct core-based tree so that the core located for the
multicast group resulted in the best-multicast tree with
respect to the desired QoS specified by the multicast
application.
Performance of the core selection method was measured
using Bandwidth i.e. the sum of all weights from the
B. Core Migration
Core migration occurs after selecting a new core for a
multicast group. During the lifetime of a multicast
application, the members of a group may change, and the
resources in the network may fluctuate. The motive of the
ore migration is to identify a new core node for the group
whose corresponding multicast tree, determined by the
current set of group members and present network status,
will likely result in significantly better multicast
performance than that of the current core [6]. For a fixed
network topology, the core migrates to an optimal position
on a hop-by-hop basis as opposed to the dynamic network
topology where the migration of the core to an optimal
position equalizes the edge weight (delay) coming from all
the subtree branches [13]. The core migration is based on
the core-based multicast routing. This means that, if the
performance metrics such as edge weight and node weight
of an adjacent core node is superior to the previously
considered node, then that adjacent core will be addressed
as the migrated core and the members of the multicast
group will then use this migrated core node as their new
core node [22]. Typical core-based routing selects one
router as a core for a multicast group and builds a single
multicast tree rooted at the core to deliver data to the entire
multicast group [24].
In this core migration approach, the core determines the
edge weight over the several edge weights to find a
neighbor experiencing the highest branch weight. The
neighbor with the highest branch delay or weight is selected
by the core as the new core if and only if the final migration
to this node will not increase the delay on other branches
[14].
1) TEMP_CORE Selection Method: The selection of the
intermediate node is termed the TEMP_CORE, which has
been selected for temporary basis until you find another
optimal node based on the edge and the node weight. This
node gets selected every time the algorithm is run for
finding an optimal position. For the node weight algorithm,
the lowercase node with largest path delay is selected and
for the edge weight algorithm, that particular node is
considered as the intermediate node having the minimum
weight value as mentioned by the table I. The on-route node
with the largest path delay accounts for the largest amount
of time required for the transmission of the queries sent and
their corresponding replies received. Then it is required to
migrate to this on-route node after comparing the weight
parameters of the current core with all its children nodes.
The migration to the on-route node with largest path delay
and minimum node weight is desirable.
C. Reasons for Migrating the Core Node
Due to unconstrained movement of nodes in mobile ad
hoc networks, the topology of the network keeps changing.
Core Migration can be invoked when the quality of the tree
degrades due to membership dynamics or w tree followed
by addition of the node to the tree. Further, new nodes may
join the multicast group and hence the tree [18]. Hence, a
node which is a good core node for the multicast tree at a
given time may not remain good due to these changes.
Hence, a new core node must be periodically chosen.
However, changing the core node involves informing all the
other nodes of the new core node as well as modifying the
multicast tree so that the new core becomes the root of the
multicast tree. The core migration in the multicast tree of
mobile ad-hoc network is invoked due to [1]:
1) Recovery from Core Failure: In core-based
multicasting, core is a single point of failure. If the core
fails, there is a large amount of packet loss, as many of the
receivers cannot receive the data sent by the senders.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO202-4

Therefore, recovery from both core failure and node/link
failure are important issues.
2) Core Degeneration & Migration: Due to the dynamic
nature of multicast group, core „degenerates‟ with time.
That is, the quality of the core and hence the multicast tree
will deteriorate with time. This means that the core
migration has to be invoked when the quality of the core
degenerates.
Core migration involves the following: (i)
selecting a new core that can offer better performance than
the current core, (ii) constructing a multicast tree based on
the new core, and (iii) migrating the group members from
the current multicast tree to the new multicast tree (tree
migration). Thus, core migration plays a vital role in tree
maintenance and core failure recovery.
D. Parameterized Techniques
When the membership migrates throughout the network,
current routing technique involves selection of a “core”
router through which all multicast communication is routed
[11]. Then the routing scheme is adapted efficiently to
support randomly varying groups by allowing the core node
to migrate. In order to evaluate the performance of a core
migration method, and therefore that of the resultant
multicast trees, the following techniques are generally
considered.
1) Node Weight: It indicates the unique value assigned to
every node in the network. The core node compares its
weight with all the neighboring nodes. If the weight of any
of the neighboring or the adjacent node is lower than the
weight of the core node, then it will stop looking for
another node and the node with the lower weight will be
considered as the new Core. The core migration is based on
the core-based multicast routing. This means that, if the
techniques such as path delay and weight of an adjacent
core node are superior to the previously considered node,
then that adjacent core will be addressed as the migrated
core and the members of the multicast group will then use
this migrated core node as their new core node. Simulations
can be done using high level programming using C++ or
any another platform.
2) Edge Weight: It is the distance between the core and
the child nodes adjacent to it. The distance of the core from
its group members should be less for efficient routing. The
core monitors the delay of all its adjacent nodes on a hop-
by-hop basis. Delay represents the weight of the edge
(connecting the two nodes). Therefore, a node with the
largest delay is taken for the new core node.
E. Core Migration Algorithm
1) Node Weight: An algorithm based on the individual
weight of the node is as follows:
1. Core multicast the (Edge_Query) message to all its
On Route Nodes (viz. to children nodes, here core
is the parent node for there ORNs)
2. Then it waits for (Edge_Reply) message from all
its ORNs.
3. On Route Node List is the union of all the received
edge weights (here the delays are entered
randomly)
ORNL = U (received delays from all ORNs)
4. The node with the largest edge weight is
considered as the Temp_Core.
5. It then finds the edge with the maximum weight by
selecting a set of candidate nodes.
6. If the selected candidate core has the maximum
edge value then it will be considered as the new
core.
7. Else pick the adjacent node with larger delay to be
the next Temp_Core, until the node with larger
delay is found.
2) Edge Weight: An algorithm based on the path delay
between the two nodes is as follows:
1. Core multicast the (Node_Query) message to all its
On Route Nodes (viz. to children nodes, here core
is the parent node for there ORNs)
2. Then it waits for (Node_Reply) message from all
its ORNs.
3. On Route Node List is the union of all the received
weights (here the node values are entered
randomly)
ORNL = U (received weights from all ORNs)
4. The node with the minimum node value is
considered as the Temp_CORE.
5. It then finds the node with the minimum weight by
selecting a set of candidate nodes
6. If its own weight is less than lowest adjacent node
weight viz. Temp_Core weight < weight of the
lowest adjacent node
7. Then migrate the core (viz. now the Temp_Core
will be considered as the new core)
8. Else pick the adjacent node with smaller node
value to be the next Temp_Core, until the node
with less weight is found.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO202-5


Fig. 2. Core Migration
TABLE I
MINIMUM NODE VALUE FOR CORE MIGRATION
Nodes
(N
x
)
Node Weight
(W)
Nodes (N
x
) Node Weight
(W)
1 17 8 16
2 15 9 16
3 5 10 11
4 12 11 7
5 11 12 11
6 9 13 9
7 14 14 7

The algorithm for the migration of the core node in
wireless ad hoc networks given in this section illustrates
the node migration based on the two migration
techniques discussed above in a D subsection. From
figure 2, it is stated that the core migrates on a hop-by-
hop basis by continuously comparing the delay and node
values with their child (downstream) nodes, until a node
with a minimum node weight and largest delay value
appears. Here, the core migration takes place at the N
x
=
10, as it is having the largest edge value of W
e
= 40 when
the edge weight algorithm is run and and N
x =
3 for
minimum node value of 5 for the case when the node
weight algorithm is run. The minimum node value of W
= 5 compared to TEMP_CORE nodes. But if the node
value of node 3 is changed to a value of 20, then the node
4 becomes the finally migrated core.
IV. RESULTS
The qualitative analysis of the core migration protocol
for wireless ad hoc network has been carried out by
comparing the two different algorithms for core migration
such as: edge weight algorithm and the node weight
algorithm.
Since the edge weight, the delay between the two
adjacent nodes are entered randomly, every edge
(connecting link between two nodes) takes on a different
values for a number of iterations done. Similarly the node
weight algorithm takes on the random values for each
iteration. For every network model designed, total of
5iterations are done to achieve the core migration.
A. Case Study I (using 7 nodes)
In the first case, the network graph was modeled using 7
nodes forming a multicast wireless ad hoc network. All the
nodes communicate through mesh structure. When the
above-mentioned issues related to the wireless ad hoc
network appears then core migration algorithm repeats for a
number of times, until a best core has been selected and
then migrated.



Fig. 3. Core Migration using 7 nodes

Table II shows the core migration based on the random
value of the link (edge) connecting the two nodes, where
the number of on-route nodes are participating in achieving
the core migration for the larger value of the link between
the parent node and the child node. Similarly, table III
shows the core migration on a hop-by-hop basis using the
minimum node value.

TABLE II
CORE MIGRATION USING 7 NODES FOR EDGE WEIGHT
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO202-6


TABLE III
CORE MIGRATION USING 7 NODES FOR NODE WEIGHT
B. Case Study II (using 10 nodes)
Finally, the network graph was modeled using 10 nodes
forming a multicast wireless ad hoc network.


Fig. 4. Core Migration using 10 nodes

TABLE IV
CORE MIGRATION USING 10 NODES FOR EDGE WEIGHT

TABLE V
CORE MIGRATION USING 10 NODES FOR NODE WEIGHT
Table IV and V shows the core migration using the edge
weight and the node weight respectively.
V. CONCLUSIONS
Research in wireless networks is progressing very fast
and numerous researchers from various fields focusing to
develop some workable scheme to find the best core node
or migration of the core if necessary. The motive of this
paper was to migrate the core to an optimal position by
No.
of
Itera
tion
T
C
1
Del
ay
of
TC
1
T
C
2
Del
ay
of
TC
2
T
C
3
Del
ay
of
TC
3
Migra
ted
Core
Delay
of
Migra
ted
Core
1 3 2 4 4 6 31
2 2 10 4 11 5 20 6 31
3 3 5 4 11 5 33
4 4 11 5 31 6 34
5 4 11 5 35 6 38
No.
of
Itera
tion
T
C
1
Val
ue
of
TC
1
T
C
2
Val
ue
of
TC
2
T
C
3
Val
ue
of
TC
3
Migra
ted
Core
Value
of
Migra
ted
Core
1 5 14 7 10
2 3 11 6 9 4 5
3 2 11 4 6 6 5
4 7 24 5 16 3 11 1 9
5 5 11 3 9 1 6 6 5
No.
of
Itera
tion
T
C
1
Del
ay
of
TC
1
T
C
2
Del
ay
of
TC
2
T
C
3
Del
ay
of
TC
3
Migra
ted
Core
Delay
of
Migra
ted
Core
1 2 23 4 29 5 31 8 32
2 2 23 4 31 5 33
3 5 12 9 13 10 35
4 5 5 6 18 8 30 10 40
5 5 18 8 30 7 33 10 40
No.
of
Itera
tion
T
C
1
Val
ue
of
TC
1
T
C
2
Val
ue
of
TC
2
T
C
3
Val
ue
of
TC
3
Migra
ted
Core
Value
of
Migra
ted
Core
1 5 17 7 13 9 8
2 4 20 8 16 5 11
3 1 11 4 8 10 6
4 2 11 6 10 7 4
5 3 16 9 12 7 11 10 6
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO202-7

considering the two different techniques (edge weight and
node weight) related to the core migration. These two
parameterized techniques take on the random values every
time the core migration algorithm is run. This paper
suggests the need and possible technique of core migration
wherever applicable. The results for the core migration in
wireless ad hoc networks obtained with edge weight were
verified with that obtained with the node weight, and it was
concluded that the core migration achieved with the edge
weight technique results into the optimal location of the
core than that obtained from the network graph modeled
using node weight parameter. This was due the fact that the
edge weight considered the distance between the parent and
the child node and hence maintained a path delay between
the nodes which was not achieved with the node weight
algorithm. But on the whole, core migration achieved with
node value parameter was faster compared to the edge
weight parameter.
VI. FUTURE WORK
Further some other parameters except delay and node
weight such as hop count, total path cost, throughput, and
bandwidth need to be considered in future when migrating
the core to a best optimal position.
REFERENCES
[1] A. Chakrabarti and G. Manimaran, “A case for Scalable
Multicast Tree Migration,” Global Telecommunications
Conference (GLOBECOM 2001), IEEE, Vol.3, pp. 2026 –
2030, August 2001.
[2] A. Chakrabarti and G. Manimaran, “A case for tree migration
and integrated tree maintenance in QoS multicasting,”
Computer Communications, Volume 26, Issue 9, No.2, pp.
1007-1017, June 2003.
[3] A.Dvir and M.Segal, “Placing and Maintaining a Core Node
in wireless Ad Hoc Networks,” in Wireless Communications
and Mobile Computing, Vol.00, pp.1-14, 2008.
[4] Ayse Karaman and Hossam Hassanein, “Core-Based
Approach in Multicast Routing Protocols,” 2003.
[5] C.V.Ravishankar et.al, “Distributed Center-
Location Algorithms,” IEEE Journal on Selected
Areas in Communications, Vol.15, No.3, pp.293-303,
April 1997.
[6] Eric Fluery, Philip K. McKinley, and Yih Huang, “LCM: A
Multicast Core Management Protocol for Link-State Routing
Networks,” in Proceedings of IEEE International Conference
on Communications, Vol.2, pp.1197-1201, June 1998.
[7] Hwa-Chun Lin, Zhe-Hong Lin, “Selection of Candidate
Cores for Core-Based Multicast Routing Architectures,” in
Proceedings of IEEE International Conference on
Communications, pp.2662-2666, 2002.
[8] Jean-Francois Macq and Michel X. Goemans, “Trade-offs on
the Location of the Core Node in a Network,” Wiley
InterScience Vol. 44, Issue: 3 (October 2004) pp. 179 – 186,
2004.
[9] Kenneth L. Calvert, Ellen Witte Zegura, and Michael J.
Donahoo, “Core selection methods for multicast routing,” in
Proceedings of the 4th International Conference on Computer
Communications and Networks (ICCCN ‟95), Las Vegas,
Nevada, pp. 638+, September 20–23, 1995
[10] M. Fleury, A. C. Downton, and A. F. Clark, “On the
performance and feasibility of multicast core selection
heuristics,” IEEE Transactions on Parallel and Distributed
Systems, vol. 11, no. 11, pp. 164+, 2000.
[11] M. J. Donahoo and Ellen W. Zegure, “Core Migration for
Dynamic Multicast Routing,” Proceedings of International
Conference, Computer Communication and Networks, 1996.
[12] Mario Gerla, Rajive Bagrodia, et.al, “A Performance
Comparison Study of Ad Hoc Wireless Multicast Protocols,”
in Proceedings of 19th Annual Joint Conference of the IEEE
Computer and Communications Societies (IEEE INFOCOM
2000).
[13] Manish Kochhal, Sandeep Gupta et.al, “An Efficient Core
Migration Protocol for QoS in Mobile Ad Hoc Networks,”
IEEE International Performance Computing, and
Communications Conference, pp.387-391, 2002.
[14] Manish Kochhal, Sandeep Gupta et.al, “QoS-Aware Core
Migration for Efficient Multicast in Mobile Ad Hoc
Networks,” IEEE International Performance Computing, and
Communications Conference, October 2002.
[15] Nen-Chung Wang and Si-Ming Wang, “An Efficient Location-
Aided Protocol for Mobile Ad Hoc Networks,” Proceedings
of 11
th
International Conference on Parallel and Distributed
systems, Vol.1, pp. 335-341, July 2005.
[16] Rahul Malhotra and Savina Bansal, “Investigation of Core
Selection techniques in multicasting,” M.Tech thesis, 2008.
[17] Roger Zimmermann, Rahul Hampole, and Beomjoo Seo,
“Distributed Core Migartion in Multicast Peer-to-Peer
Streaming,” in Proceedings of International Conference on
Parallel and Distributed processing techniques and
Applications (PDPTA 2004), June 21-24,2004.
[18] S.K.S.Gupta, P.K.Srimani et.al, “Adaptive Core Selection and
Migration method for Multicast Routing in Mobile Ad Hoc
Networks,” IEEE Transactions on parallel and Distributed
Systems, Vol.14, No.1, January 2003.
[19] Stephen Mueller, Rose P. Tsang, and Dipak Ghosal,
“Multipath Routing in Mobile Ad Hoc Networks: Issues and
Challenges,” In Performance Tools and Applications to
Networked Systems, volume 2965 of LNCS 2004.
[20] Sudarshan Vasudevan, Jim Kurose and Don Towsley, “Design
and Analysis of a Leader Election algorithm for Mobile Ad
Hoc Networks,” Proceedings of 12
th
IEEE International
conference on Network protocols (ICNP), pp. 350-360,
October 2004.
[21] Ting-Yuan Wang, Lih-Chyau Wuu, and Shing-Tsaan Huang,
“A Scalable Core Migration Protocol for Dynamic Multicast
Tree,” Journal of Information Science and Engineering,
Vol.19, pp.479-501, 2003.
[22] Deering, S., Estrin, D. L., Farinacci, D., Jacobson, V., Liu, C.
and Wei, L., (1996), “The PIM Architecture for Wide-Area
Multicast Routing,” IEEE/ACM Transactions on Networking,
4(2), pp.153-162.
[23] Gerla, M., Chiang, C. C., and Zhang, L., (1999), “Tree
Multicast strategies in mobile, multihop wireless networks,”
In Mobile Networks and Applications, 4 (3), pp. 193-207.
[24] Puthhividhya, W., Tavanapong, W., Tran, M., and Wong, J.,
(2004), “Distributed Core Selection with QoS Support,” In
IEEE International Conference on Commumnications, 4, pp.
2132-2137.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO203-1

Performance comparision of RI & TBI based system
in fading envoirement using MRC scheme

M. Shukla
Department of Electronics Engineering
Harcourt Butler Technological Institute
Kanpur, India
manojs@hbti.ac.in
Aasheesh Shukla, Rohit, Sankalp
Department of Electronics & Communication Engineering
GLA University
Mathura, India
aasheeshshukla@gmail.com


Abstract—In interleave division multiple access (IDMA),
interleavers are used for the user separation. This paper
presents the performance comparison between random and
tree based interleaver based IDMA system, in fading
environment. We also know that Maximal Ratio Combining
(MRC) diversity approach is popularly used to mitigate the
fading problem in any wireless system so here MRRC scheme
is used with both RI & TBI based system to mitigate fading.
The approach involves receiver diversity analysis as 1:2
antenna systems in fading environment; no. of antennas can be
increased to improve diversity order.

Keywords- Random Interleaver,Tree Based Interleaver,
MRC diversity, IDMA
I. INTRODUCTION
The goal for the next generation mobile communications
system is to seamlessly provide a wide variety of
communication services to anybody, anywhere, anytime
such as high voice quality, higher data rates etc. The
technology needed to tackle the challenges to make these
services available is popularly known as the Third
Generation (3G) Cellular Systems using multiuser detection
[1]. The fundamental phenomenon which makes reliable
communication difficult is time varying multipath fading,
which is major impairment in any wireless communication
system. The performance improvement is very difficult in
such situation
Theoretically, improvement in signal to noise ratio may be
achieved by providing higher transmit power or additional
bandwidth which are not feasible solution as they are
contrary to the requirements of next generation wireless
communication [2]. On the other hand, the problem of fading
may be handled with suitable diversity technique without
expanding communication resources easily.
In most wireless channels, antenna diversity is a
practical, effective and widely used technique for reducing
the effect of multipath fading. The maximal ratio combining
(MRC) diversity technique, is implemented with interleave-
division multiple-access (IDMA) scheme, as MRC is
performed well in comparison with selection or equal gain
combining.[3]-[5]. The IDMA scheme is known as advanced
version of CDMA, which inherits many advantages from
CDMA such as dynamic channel sharing, mitigation of
cross-cell interferences, asynchronous transmission, ease of
cell planning, and robustness against fading. It also allows a
low complexity multiple user detection (MUD) techniques
[7] (CBC detection) applicable to systems with large
numbers of users in multi-path channels.
The objective of this paper is to use MRC diversity in
IDMA communication system to reduce the effect of fading.
The study of transmit as well as receiver diversity is taken
separately, because both have their own application area [2].
The paper is organized as follows. Concept of IDMA is
introduced in section II. Section III deals with classical
MRRC diversity approach used with IDMA. In section IV
transmit diversity is discussed with IDMA. Performance
analysis is provided in section V. Finally conclusions are
presented in section VI.
II. IDMA SCHEME
A. IDMA Mechanism
The performance of conventional code-division multiple-
access (CDMA) systems [1] is mainly limited by multiple
access interference (MAI), as well as intersymbol
interference (ISI). Also, the complexity of CDMA multi-user
detection has always been a serious problem for researchers
all over the world. The problem can be visualized from the
angle of computational cost as well complexity of multi-user
detection algorithms in CDMA systems. The use of user-
specific signature sequences is a characteristic feature for a
conventional CDMA system. The possibility of employing
interleaving for user separation in CDMA systems is briefly
inducted in [1] but the receiver complexity is considered as a
main problem. In interleave-division multiple-access
(IDMA) scheme, users are distinguished by user specific
chip-level interleavers instead of signatures as in a
conventional CDMA system. The scheme considered is a
special case of CDMA in which bandwidth expansion is
entirely performed by low-rate coding. This scheme allows a
low complexity multiple user detection techniques applicable
to systems with large numbers of users in multipath channels
in addition to other advantages.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO203-2

In CDMA scheme, signature sequences are used for user
separation while in IDMA scheme, every user is separated
with user-specific interleavers, which are orthogonal in
nature. The block diagram of IDMA scheme is shown in
figure 1 for K users. The principle of iterative multi user
detection (MUD) which is a promising technique for
multiple access problems (MAI) is also illustrated in the
lower part of Fig. 1. The turbo processor involves elementary
signal estimator block (ESEB) and a bank of K decoders
(SDECs). The ESEB partially resolves MAI without
considering FEC coding. The outputs of the ESEB are then
passed to the SDECs for further refinement using the FEC
coding constraint through de-interleaving block. The SDECs
outputs are fed back to the ESEB to improve its estimates in
the next iteration with proper user specific interleaving. This
iterative procedure is repeated a preset number of times (or
terminated if a certain stopping criterion is fulfilled). After
the final iteration, the SDECs produce hard decisions on the
information bits [8]-[11].
The complexity involved (mainly for solving a size KxK
correlation matrix) is O(K
2
) per user by the well-known
iterative minimum mean square error (MMSE) technique in
CDMA, while in IDMA, it is independent of user. This can
be a major benefit when K is large [12].
B. Scheme Model
Here, we consider an IDMA system [1], shown in Figure
1, with K simultaneous users using a single path channel. At
the transmitter, a N-length input data sequence d
k
= [d
k
(1),
………, d
k
(i) , … d
k
(N) ]
T
of user k is encoded into chips
c
k
= [c
k
(1), ………, c
k
(j) , … c
k
(J) ]
T
based on low rate
code C, where J is the Chip length.
The chips c
k
is interleaved by a chip level interleaver
„Π
k
‟, producing a transmitted chip sequence x
k
= [x
k
(1),
……,x
k
(j) , … x
k
(J) ]
T
. After transmitting through the
channel, the bits are seen at the receiver side as r = [r
k
(1),
……,r
k
(j) , … r
k
(J) ]
T
. The Channel opted is additive white
Gaussian noise (AWGN) channel, for simulation purpose.
In receiver section, after chip matched filtering, the
received signal form the K users can be written as
1
( ) ( ) ( ), 1, 2,....... .
K
k k
k
r j h x j n j j J
(1)
where h
k
is the channel coefficient for
th
k user and
{ ( ) n j } are the samples of an additive white Gaussian
noise (AWGN) process with mean as zero and variance σ
2
=N
0
/ 2. An assumption is made that {h
k
} are known priori at
the receiver.
The receiver consists of a elementary signal estimator
block (ESEB) and a bank of K single user a posteriori
probability (APP) decoders (SDECs), operating in an
iterative manner. The modulation technique used for
simulation is binary phase shift keying (BPSK) signaling.
The outputs of the ESEB and SDECs are extrinsic log-
likelihood ratios (LLRs) about {x
k}
defined as

( / ( ) 1)
( ( )) log , , .
( / ( ) 1)
k
k
k
p y x j
e x j k j
p y x j
(2)

Figure 1. Transmitter and Receiver structures of IDMA scheme with K
simultaneous users.

These LLRs are further distinguished by the
subscripts i.e., ( ( ))
ESEB k
e x j and ( ( ))
SDEC k
e x j , depending
upon whether they are generated by ESEB or SDECs.
Due to the use random interleavers {Π
k
}, the ESEB
operation can be carried out in a chip-by-chip manner, with
only one sample r(j) used at a time. So, rewriting (2) as
( ) ( ) ( )
k k k
r j h x j j (3)
where
' '
'
( ) ( ) ( ) ( ) ( )
k k k
k k
k k
j r j h x j h x j n j
(4)
is the distortion in r( j) with respect to user-k. ( )
k
j is
the distortion (including interference-plus-noise) in received
signal with respect to user-k.
A brief description of CBC algorithm [1] used in IDMA,
has been presented in [3]. The operations of ESEB and APP
decoding are carried out user-by-user.
The outputs of the ESEB as extrinsic log-likelihood
ratios (LLRs) is given as,


The LLR output of SDEC is given as,


Now, these steps are repeated depending on no. of
iterations and users.
2
( ) ( ( )) ( ( ))
( ( )) 2 .
( ) ( ( ))
k k
ESEB k k
j k k
r j E r j h E x j
e x j h
Var r h Var x j
1
( ( ( ))) ( ( ( )))
S
SDEC k ESEB k
j
e x j e x j
1,..., j S
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO203-3

III. RECIEVER DIVERSITY ANALYSIS FOR IDMA
The block diagram of maximal ratio combining (MRC)
diversity with IDMA scheme is shown in figure 2. In this
method, the diversity branches are weighted for maximum
SNR. As shown in block diagram in figure 2,
k
d is data of k
th

user, after encoding and spreading the data is randomly
interleaved and termed as „chips‟. Now this chip Signal
k
x is
sent from the transmit antenna, which will propagate from
both the channel.
If we consider 1 transmit and 2 receive antenna, then
channel between transmit antenna and the first received
antenna is h
0
and between the transmit antenna and second
receive antenna one is denoted by h
1
. The channel can be
modeled having magnitude and phase response [16]. So,
h
0 = 0
0 i
e
h
1
=
1
1 i
e (5)
Noise can be added at both the receiver. The resulting
received signals are
R
0
=h
0
x
k
+n
0

R
1
=h
1
x
k
+n
1
(6)
where n
0
and n
1
represents the noise and interference at
both the receiver separately.
Now the Receiver combining scheme for two branches
MRRC can be written as

K
=h
0
R
0
+ h
1
R
1
(7)
Now this output of maximal ratio combiner can fed to the
detector for the proper estimation of transmitted signal x
k
.


Figure 2. IDMA with proposed two branches MRRC diversity scheme for
k
th
user


IV. PERFORMANCE ANALYSIS
For simplicity, IDMA system with BPSK signaling is
assumed with uniform repetitive coding and spread length
16, for all users along with 15 iterations.

TABLE I

COMPARISON OF COMPUTATIONAL COMPLEXITY WITH TREE BASED
INTERLEAVER AND RANDOM INTERLEAVER

User
Number
No. of times Interleaving
in Tree Based Interleaver
No. of times Interleaving
in Random Interleaver
2 1 2
6 2 6
14 3 14
30 4 30
62 5 62
126 6 126

The interleavers used in simulations are random
interleaver (RI) [3] and tree based interleaver (TBI)
[15].
0 10 20 30 40 50 60 70 80 90 100
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
x 10
6
User Number
M
e
m
o
r
y

R
e
q
u
i
r
e
m
e
n
t

o
f

I
n
t
e
r
l
e
a
v
e
r
With Random Interleaver
With Tree Based Interleaver

Figure 3. Memory requirement comparison of Random Interleaver and
Tree Based Interleaver.

Table 1 denotes the computational complexity analysis
for both interleavers and it is clear that for large no. of users,
computational complexity of TBI based system is lower.
Figure 3 demonstrates the memory requirement of RI and
TBI. The memory requirement of RI is dependent on user
count [3] [4] while that of TBI is constant [15] due to use of
only two master interleaving sequences for generating the
other user specific interleaving sequences. The memory
required by Tree Based Interleaver generation method is
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO203-4

extremely less than that required for random interleaver
generation method [3] and is independent of user count.



Figure 4 demonstrates the performance of IDMA scheme
with using random interleaver and tree based interleaver.
Here we refer the channel as slow fading Rayleigh channel.
The interleavers used in simulations are random interleaver
[3] and tree based interleaver [15]. The block length is 200
with spread length to be 16. The iterations selected for
simulation in receiver is 15. The simulations have been
performed for one transmitter and two receiver arrangement
in the case of receiver diversity.
The BER performance shows that although RI based IDMA
performs better but TBI based system also performs good for
larger value of E
b
/N
0





2 3 4 5 6 7 8 9 10
10
-6
10
-5
10
-4
10
-3
10
-2
10
-1
Eb/No
B
i
t

E
r
r
o
r

R
a
t
e
TBI with receive diversity.k=131072
TBI with receive diversity,k=65536
RI with receive diversity, k=65536
RI with receive diversity, k=131072

Figure 4. Performance of RI- IDMA & TBI-IDMA with t receive diversity

V. CONCLUSIONS

After doing performance comparison, we proposed that TBI
based IDMA system can be optimum selection in fading
environment. It has been explained that tree based interleaver
performs well, near to random interleaver [figure 4], but in
the reference of computational complexity and memory
requirement it takes edge on random interleaver. It can also
conclude that IDMA with suitable diversity technique can
generate fruitful results in the area of wireless
communication. study can also enhanced to Multiple input
and multiple output (MIMO) antenna system to improve the
diversity order and hence the performance of IDMA
communication system.
REFERENCES
[1] M. Moher and P. Guinand, “An iterative algorithm for asynchronous
coded multi-user detection,” IEEE Commun. Lett., vol. 2, pp. 229–
231, Aug. 1998.
[2] S.M.Alamouti. “A Simple transmitter diversity scheme for wireless
communications” IEEE J. Select. Areas Commun., vol.16 , pp. 1451-
1458, Oct.1998.
[3] Li ping, Lihai Liu, Keying Wu and W. K. Leung, “Interleave-
Division Multiple-Access,” IEEE Trans. Wireless Commun., vol. 5,
pp. 938-947, April 2006.
[4] H. Wu, L. Ping and A. Perotti, “User-specific chip-level interleaver
design for IDMA systems,” IEEE Elec. Lett., vol. 42, no. 4, Feb.
2006.
[5] Li Ping and L. Liu, “Analysis and design of IDMA systems based on
SNR evolution and power allocation,” in Proc. VTC‟2004-Fall, Los
Angles, CA, Sept. 2004
[6] S. Brück, U. Sorger, S. Gligorevic, and N. Stolte, “Interleaving for
outer convolutional codes in DS-CDMA Systems,” IEEE Trans.
Commun., vol. 48, pp. 1100–1107, July 2000.
[7] Tarable, G. Montorsi, and S. Benedetto, “Analysis and design of
interleavers for CDMA systems,” IEEE Commun. Lett., vol. 5, pp.
420–422, Oct. 2001.
[8] F. N. Brannstrom, T. M. Aulin, and L. K. Rasmussen, “Iterative
multi-user detection of trellis code multiple access using a posteriori
probabilities,” in Proc. ICC 2001, Finland, June 2001, pp. 11–15.
[9] S. Verdú and S. Shamai, “Spectral efficiency of CDMA with random
spreading,” IEEE Trans. Inform. Theory, vol. 45, pp. 622–640, Mar.
1999.
[10] L. Liu, W. K. Leung, and Li Ping, “Simple chip-by-chip multi-user
detection for CDMA systems,” in Proc. IEEE VTC-Spring, Korea,
Apr. 2003, pp. 2157-2161.
[11] Li Ping, L. Liu, K. Y. Wu, and W. K. Leung, “A unified approach to
multi-user detection and space time coding with low complexity and
nearly optimal performance,” in Proc. 40th Allerton Conference,
Allerton House, USA, Oct. 2002, pp. 170–179.
[12] L. Liu, W. K. Leung, and Li Ping, “Simple iterative chip-by-chip
multiuser detection for CDMA systems,” in Proc. IEEE VTC 2003,
Korea, Apr. 2003.
[13] M. Shukla, V.K. Srivastava, S. Tiwari, “Analysis and Design of
Optimum Interleaver for Iterative Receivers in IDMA Scheme”,
Wiley Journal of Wireless Communication and Mobile Computing,
Vol 9. Issue 10, pp. 1312-1317. Oct, 2009.
[14] M. Shukla, Aasheesh Shukla,V.K. Srivastava, S. Tiwari, “Performance
evaluation of MRC diversity scheme for iterative IDMA receiver”,in
proc. INDICON 2009, India, Dec. 2009.




CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO205-1

Security Constraints in Peer-to-Peer Network on Overlay Networks


Raj Kumar Gaur
M. Tech. (Information Security)
Computer Science & Engineering Department,
M.M.M. Engineering College Gorakhpur
rajkumargaur@gmail.com



Divya Gupta
M. C. A.
Computer Science & Engineering Department,
M.M.M. Engineering College Gorakhpur
guptadivya28@gmail.com


Abstract

Network Information System, is not only the
information sharing and information exchanging; recently
there has been a significant amount of work on security
and privacy. Network Information System is not only
include information system network security and data
security, but also include the security of network business
running on information system network which is the
confidentiality, integrity, continuity and real-time of
network. The proposal of the concept "network information
security" provides theoretical basis for security defense of
enterprise automatic production system; and enterprise
management information system. Overlay network
monitoring enables distributed Internet applications to
detect and recover from path outages and periods of
degraded performance. Our scheme assumes only the
knowledge of underlying peer to peer topology, and any
link may loss or return to normal through paper, introduce
security constraints for privacy preserving set union in the
malicious adversary model. The proposal of the concept
"network information security" provides theoretical basis
for security system and enterprise management information
system.

Keywords: Information Security, Network Security, P2P Security
Model.

1. Introduction

P2Pnetworking allows computers to
communicate directly with one another rather than
through a central server like a website. P2P networks
have been the centerpiece especially involving
illegally; File-sharing across the peer-to-peer network
can lead to inadvertently breaking copy-right laws.
An application the Peer-to-Peer networking concept
is used to share files, i.e. the exchange of files.
However, Peer-to-Peer is not only about file sharing
[3], it is also




about establishing multimedia communication
networks based on Peer-to-Peer concepts or resource
sharing. Peer-to-Peer networks just as a collection of
heterogeneous distributed resources which is
connected with a network [1]. A lot of problems still
have to be solved and one of them is certainly the
traffic engineering problem encountered especially in
completely [15]. A large number of peer processes
are distributed in a peer-to-peer (P2P) overlay
network. Various types and a huge number of peer
computers are interconnected and the membership is
dynamically changed in a peer-to-peer overlay
network. P2P Security is concerns one major concern
of using P2P of network security. Security concerns
stem from the architecture itself. Now a day find most
blocking and routing handles by a specific server
within network, but the P2P architecture has no single
fixed server responsible for routing and requests. In
securing network against attacks and viruses there are
two main strategies where focus is on controlling the
network access or the focus is put on controlling the
files. Open networks are often insecure and provide
an opportunity for viruses and DoS activities to
spread. To make such networks more resilient against
these kind of threats, the propose the use of a peer-
to-peer architecture whereby each peer is responsible
for: (a) detecting whether a virus or worm is
uncontrollably propagating through the network
resulting in an epidemic; (b) automatically
dispatching warnings and information to other peers
of a security-focused group and (c) taking specific
precautions for protecting their host by automatically
hardening their security measures during the
epidemic. Vulnerabilities may result from software
bugs, weak passwords, or system miss configurations.
CVE (Common Vulnerabilities and Exposures) , a
standard dictionary with information about security
vulnerabilities and exposures, includes a set of
common identifiers that enable data exchange of
security products and provide a baseline index point
to evaluate coverage of tools and services. As a
result, a similarity method for security vulnerability
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO205-2

can speed up the process of vulnerability mitigation
as similar vulnerabilities may require similar
solutions. In other words, in order to take any action,
a level of priority can be determined based on known
data from other vulnerabilities. For example, a low
priority level can be assigned to a newly found
vulnerability if it is proved that it is similar to a
known vulnerability that has a low impact, while the
basic idea is simple, a practical system design is not
easy. P2P reputation system are consists of three
functional components collecting information on peer
behavior, scoring and ranking peers, and responding
based on peers’ scores.
All these components are nontrivial, especially
given the following consideration [4]:
Scalability: A large P2P network may have
hundreds of thousands of peers. For example, Skype
has several million online users. A reputation system
should be highly scalable in terms of peer number.
Adaption to peer dynamics: Peers may join or
leave at any time. If reputation information is
maintained at peers, peer leaving may lead to
information loss. A robust reputation system should
take peer dynamics into account.
Security: Malicious peers may endeavor to break
down the reputation system so that they can conduct
malicious actions without being detected. For
example, peers may purposefully leave and rejoin the
system with a new identity in order to shed any bad
reputation.
There are two fundamental issues in P2P
reputation systems. Reputation estimation [10]: An
estimation method describes how to generate peer
reputation based on others’ feedback. These classify
existing estimation methods into three categories:
social network, probabilistic estimation, and game
theoretic model. As many estimation methods rely on
specific feedback collection mechanisms the feedback
collection mechanisms. The necessary networks are
not trivial. First, efficient data storage and retrieval is
always a challenging issue in P2P networks. Then
efficient retrieval becomes nontrivial. Peer dynamics
bring more difficulty. Second, reputation data are
highly security-sensitive. There are many other
important issues in P2P reputation systems; for
example, how to prevent targeted and adversarial
attacks? How to interpret reputation? Interested
readers may refer to for a comprehensive overview of
P2P reputation issues to the end, each peer can
connect to the overlay topology [6] at multiple points
(i.e., receive content through multiple parent peers).
Therefore, participating peers form an unstructured
overlay [2] that can constructing a peer-to-peer
overlay for streaming applications should not only
accommodate global design goals such as scalability
and resilience but also satisfy the local design goal of
maximizing delivered quality to individual peers. The
commonly have the two drawbacks as follows. First,
the systems have static overlay networks [5] consisted
of reliable nodes under the administrative control, or
assume that a spanning tree of entire node is known
beforehand. Clearly, this is not feasible when a
system involves an enormous number of nodes, which
join and leave the overlay network dynamically. In
extreme case applications, nodes systems may be
organized only by client nodes without any
specialized node. Second, a node keeps a large
amount of routing states and its control message
overhead is huge. This is because every node can be
an intermediate router on the paths of an event
dissemination tree. A distributed sender/receiver
needs a structured overlay network, but it must be
designed with great care since the underlying peer-to-
peer overlay substrate has a significant effect on the
performance of the system.
To solve the above challenges, a new peer-to-peer
overlay; one important metric in evaluating
distributed content based systems is the percentage of
false positives over the total number of messages. A
false positive is defined as a message received by the
node not interested in the message. This section
presents some preliminary evaluation of the security
protocol [9] using simulations.
So peer-to-peer file-sharing are so dangerous?
There are a number of reasons:
Illegal, Unethical, and immoral--unfortunately the
most common uses of P2P are illegal ones such as
copy-righted music, copyrighted software,
pornographic material, and even child pornography.
Viruses--when downloading a shared file, one big
danger is to download a computer virus or Trojan
horse onto your computer. These viruses can cause all
kinds of problems from erasing all files on the hard
drive to the automatic sending of pornographic e-
mails to all of your friends in your mail directory.
Overlay Network Centralization--Although in their
purest form peer-to-peer overlay networks are
supposed to be totally decentralized in practice this is
not always true, and systems with various degrees of
centralization are encountered.
This typically renders them inherently un-scalable
and vulnerable to censorship, technical failure or
malicious attack and peer to peer objective in figure 1
shows.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO205-3


Figure: 1 Peer to Peer Objective

2. Application of P2P Network and
Architecture

Peer-to-peer networks, which will hereafter
reference as p2p networks, are often considered to be
security threats for organizations, companies or plain
users, mainly due to the use of p2p-based applications
for illegal file sharing, and to the ability of worms to
be spread through such applications. Our scheme
indicates, however, that p2p networks can also be
positively utilized to significantly reinforce network
security, by offering substantial help in the protection
against malicious applications. The propose an
effective way to achieve this by collecting and
exchanging information that will allow us to obtain a
global overview of the network status, with reference
to ongoing security attacks. The goal of our
methodology is to select the most appropriate security
policy, based on the level of danger posed by rapid
malicious code circulating in the network. P2p
networks leverage the principle that a much better
utilization of resources (processing power,
bandwidth, storage etc.) is achieved if the
client/server model is replaced by a network of
equivalent peers networks.
The applications on p2p architectures or
infrastructures derives to a large extent from their
adaptability to variable operating environments, i.e.
their ability to function, scale and self-organize in the
presence of a highly transient population of nodes (or
computers/users) hardware failures and network
outages, without the need for a central administrative
server.
Peer-to-peer architectures have been employed for
a variety of different application categories, which
include the following.
Communication and Collaboration: This category
includes systems that provide the infrastructure for
facilitating direct, usually real-time, communication
and collaboration between peer computers.
Distributed Computation: This category includes
systems whose aim is to take advantage of the
available peer computer processing power (CPU
cycles). This is achieved by breaking down a
computer-intensive task into small work units and
distributing them to different peer computers.
Internet Service Support [10]: A number of
different applications based on peer-to-peer
infrastructures have emerged for supporting a variety
of internet services. Examples of such applications
include peer-to-peer multicast systems, internet
indirection infrastructures, and security applications
providing protection against denial of service or virus
attacks.
Database Systems: Considerable work has been
done on designing distributed database systems based
on peer-to-peer infrastructures. The purpose of Local
Relational Model (LRM), in which are the set of all
data stored in a peer-to-peer network.
Content Distribution: Most of the current peer-to-
peer systems fall within the category of content
distribution, which includes systems and
infrastructures designed for the sharing of digital
media and other data between users.
The applications on p2p architectures or
infrastructures derives to a large extent from their
adaptability to variable operating environments, i.e.
their ability to function, scale and self-organize in the
presence of a highly transient population of nodes (or
computers/users), hardware failures and network
outages, without the need for a central administrative
server.

3. Pee-to-Peer Network Architecture

Streaming is a process of generating and
delivering a stream flow of data packets over
networking medium, e.g., the Internet, from a source
to a destination. Streaming media usually denotes
digital audio and video data; however other data may
be streamed as well. Theoretically, P2P architecture
can be built over any networking medium and at
potentially different layers of the network. However,
most of the existing P2P implementations and their
associated research have focused on application-level
overlay networks [11].
A streaming process can be separated into three
stages that overlap in time: data acquisition, data
delivery and data presentation. Data acquisition is the
stage that determines how the streaming content is
acquired, packetized and distributed for streaming.
The data presentation stage represents the methods on
how to buffer, assemble and render the received data.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO205-4

Data delivery is the process of how the stream data is
transported from the source to the destination. The
source, the destination and all the intermediate nodes
in a streaming system participate in a topology that is
constructed based on the specific system’s protocol in
a P2P streaming system.

Figure 2: The Architecture of P2P Network
with security system

Data Acquisition and Presentation: At the streaming
source, the content is prepared for distribution. If the
data was prerecorded and is available as files, it
categorizes this as on-demand streaming. Data
Delivery: The transition of one or multiple copies of
the content from a source node to a destination node
is called a streaming session. A streaming session
starts when a streaming request is made and ends
when all associated destination nodes have received
the last byte of the content. Depending on the number
of source and destination nodes involved in a
streaming session. Peer-to-Peer System Operation:
From the perspective of a peer, the life-cycle of a P2P
streaming session can be decomposed into a series of
four major processes: finding the service, searching
for specific content, joining or leaving the service,
and failure recovery when there is an error.

3.1 Analysis of Architecture

In this description and analysis of architecture,
systems and infrastructures that are based on peer-to-
peer architectures and aim at either offering content
distribution solutions or supporting content
distribution related activities:
Our approach is based on: identifying the feature
space of non-functional properties and characteristics
of content distribution systems,
Determining the way in which the non-functional
properties depend on and can be affected by various
design features, and
Providing an account, analysis and evaluation of
the design features and solutions that have been
adopted by current peer-to-peer systems, as well as
their shortcomings, The various design features and
solutions that examine, as well as their relationship to
the relevant non-functional characteristics, were
assembled through a detailed analysis and current
peer-to-peer content distribution systems that are
either deployed
The most important are attributes of peer to peer
contents distribution systems namely [14, 15].
Security: Further analyzed in terms of:
Integrity and Authenticity: Safeguarding the
accuracy and completeness of data and processing
methods. Unauthorized entities cannot change data
adversaries cannot substitute a forged document for a
requested one.
Privacy and Confidentiality: Ensuring that data is
accessible only to those authorized to have access and
that there is control over what data is collected, how
it is used and how it is maintained.
Availability and Persistence: Ensuring that
authorized users have access to data and associated
assets when required. For a peer-to-peer content
distribution system this often means always. This
property entails stability [7] in the presence of failure
or changing node populations.
Scalability: Maintaining the system's performance
attributes independent of the number of nodes or
documents in its network. A dramatic increase in the
number of nodes or documents will have minimal
effect on performance and availability.
Performance: The time required for performing the
operations allowed by the system, typically
publication, searching and retrieval of documents.
Fairness: Ensuring that users offer and consume
resources in a fair and balanced manner. May replies
on accountability, reputation and resource trading
mechanisms.
Resource Management Capabilities: In their most
basic form peer-to-peer content distribution systems
allow the publishing, searching and retrieval of
documents. More sophisticated systems may afford
more advanced resource management capabilities
such as editing or removal of documents,
management of storage space and operations on
metadata.
Semantic Grouping of Information: An area of
research that has attracted considerable attention
recently is the semantic grouping and organization of
content and information in peer-to-peer networks.
Various grouping schemes are encountered, such as
semantic grouping based on the content itself,
grouping based on locality or network distance,
grouping based on organization ties as well as others.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO205-5


4. Security in Peer to Peer Network

Peer-to-peer content distribution architectures
present a particular challenge for providing the levels
of availability, privacy, confidentiality, integrity, and
authenticity often required, due to their open and
autonomous nature. The network nodes must be
considered un-trusted parties [13], and no
assumptions can be made regarding their behavior.
This Section discusses various approaches that
address a number of security issues characteristic of
peer-to-peer content distribution systems. The
particularly focus on secure storage, secure routing,
access control, authentication, and identity
management.
Attackers who can choose nodeIds [12] can
compromise the integrity of a structured p2p overlay,
without needing to control a particularly large
fraction of the nodes. For example, an attacker may
partition a Pastry or Chord overlay if she controls two
complete and disjoint neighbor sets. Such attackers
may also target particular victim nodes by carefully
choosing nodeIds.
One solution to securing the assignment of
nodeIds is to delegate the problem to a central,
trusted authority. A set of trusted certification
authorities (CAs) [12,13,14] to assign nodeIds to
principals and to sign nodeId certificates, which bind
a random nodeId to the public key that speaks for its
principal and an IP address[8]. The CAs ensures that
nodeIds are chosen randomly from the id space, and
prevent nodes from forging nodeIds.
Sybil attacks: While nodeId assignment by a CA
ensures that nodeIds are chosen randomly, it is also
important to prevent an attacker from easily obtaining
a large number of nodeId certificates. One solution is
to require an attacker to pay money for certificates,
via credit card or any other suitable mechanism. With
this solution, the cost of an attack grows naturally
with the size of the network. The identity-based CA is
the preferred solution in “virtual private” overlays run
by an organization that already maintains employment
or membership records with strong identity checks.
For small overlay networks, must require that all
members of the network are trusted not to cheat.
Attacks: All structured p2p overlays provide a
primitive to send a message to a key. In the absence
of faults, the message is delivered to the root node for
the key after an average of routing hops. Solution:
detect faults, use diverse routes that is important to
devise mechanisms to route securely. It is secure
routing primitive that takes a message and a
destination key and ensures that with very high
probability at least one copy of the message reaches
each correct replica root for the key.

5. Results and Evaluation

The experimental is result security architecture that
provides to support multimedia communications in
heterogeneous and homogeneous network
environments. The implementation is shown in
Figure-2; message flows in the implementation are
nodeId to nodeId. In this implementation [15], the
architecture provides assumption results but aims at
providing dynamic multicast/uncast conversion
services to support an interactive multimedia
application. Our assumption is implemented some
major components:
Secure Application Layer: that provides request
and key management of the multicast/uncast
conversion services.
Detection System: All incoming and out going
queries the available and checked the authentication
service providers.
Secure Adaption Layer: provide the Service
confidentiality and integrity selected by the user using
as the peer-to-peer networking protocol.

6. Conclusion and Future Work

In this paper to improve the security and efficiency
of a peer-to-peer system by data streaming with
secure transformation of data nodes to nodes with
similar content in peer to peer Overlay Networks.
This can an efficiently process queries while
preserving a high degree of node autonomy. In this
introduced security, an approach that improves query
performance even more at a cost of a slight reduction
in the maximum achievable recall level. So conclude
is security and efficiency offer significant
improvements versus random overlay networks, while
keeping more secure system, that believe security and
efficiency in particular nodes can help improve the
search performance of current and future P2P systems
where data is naturally clustered with secure
transmission. The general architecture is for creating
robust and efficient use of the heterogeneity of the
hierarchical P2P systems.

The Network Security topology is self-maintained
and self healing, with low maintenance cost and
resilience to failures. The Network Security topology
allows the users to take advantage of the different
peers’ characteristics and the ability to rapidly adapt
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO205-6

to different situations (load balancing, failures,
network traffic, etc.). Network Security is well suited
to heterogeneous P2P networks. Our assumptions
show that Network Security efficiently handles
failures. It is highly stable and the resource oriented-
hierarchy provides an optimal use of the network. The
technique used to define and maintain the routing
tables allows the use of routing algorithms having a
logarithmic complexity.

7. References

[1] Wu Kehe, Zhang Tong, Li Wei, Ma Gang
“Security Model Based on Network Business
Security” IEEE 2009.

[2] Diego Doval and Donal O’Mahony Trinity
College Dublin “Overlay Networks a Scalable
Alternative for P2P” IEEE 2003.

[3] William Couch “Peer-to-Peer File-Sharing
Networks: Security Risks” 2002

[4] Yan Chen, David Bindel, Randy H. Katz
“Tomography-based Overlay Network
Monitoring” acm 2003.

[5] Kenichi Watanabe and Yoshiona kajima
“Ranking Factors in Peer-to-Peer Overlay
Networks” acm 2007.

[6] Arturo Crespo and Hector Garcia-Molina”
Semantic Overlay Networks for P2P Systems”
2004.
[7] Junghee Han, David Watson, and Farnam
Jahanian “Topology Aware Overlay Networks”
2004.
[8] Yongjin Choi and Daeyeon Park “Mirinae: A
Peer-to-Peer Overlay Network for Large-Scale
Content-Based Publish/Subscribe Systems” acm
2005.

[9] Laurent Massouli ´Erwan Le Merrer, Anne-
Marie Kermarrec and Ayalvadi Ganesh
Microsoft Research, Cambridge, U.K. IRISA and
FTR&D, Lannion, France INRIA/IRISA,
Rennes, France “Peer Counting and Sampling in
Overlay Networks: Random Walk Methods”
2006.

[10] S. Zhou, M. Hogan, S. Ardon, M. Portman, T.
Hu, K.Wongrujira, and A. Seneviratne “ALASA:
When Service Overlay Networks Meet Peer-to-
Peer Networks” IEEE 2005.

[11] Mei Li Wang-Chien Lee Anand
Sivasubramaniam “Semantic Small World: An
Overlay Network for Peer- to-Peer Search” IEEE
2004.

[12] Chunqiang Tang, Zhichen Xu, Sandhya
Dwarkadas “Peer-to-Peer Information Retrieval
Using Self-Organizing Semantic Overlay
Networks” acm 2003

[13] www http://napster.com.

[14] www http://vivisimo.com

[15] www.grokster.com





ENERGY CONSERVATION THEORY IN CONTEXT AWARE COMPUTING
Department of Computer Science, Banaras Hindu University
Varanasi-221005

Abstract – Awareness to the context is an energy
is plausible to design a handheld device which can perform
apparently to the higher level of energy conservation. Today
Ubiquitous Computing is an evolving area of computing
challenges. As per Ubiquitous Computing concepts it is
totally based on context-awareness. Context aware systems
consist with a large range of sensors which work on various
levels of energy ranges. Those energy ranges consumes
battery charge accordingly. So it is necessary to devise some
vital techniques which have control over the energy
consumption of battery and improving the on site
performance of hand-held devices.
Deliberating the performance of current hand-held devices,
the extracted result is found that it give more time in sensing
the context rather than sleeping. Hence some maintain the
unbeatable accuracy on the cost of extremely energy
consumption. Many times it is unnecessary to sense the
context because the context is not changing with the
corresponding time. Here the concept is to minimizing
energy consumption while the sensors are not sensing the
context.
Keywords – energy efficiency, context aware computing,
sensors, Bluetooth, microphone, back-off function, advance
function.
I. INTRODUCTION
Energy efficiency is a vital issue in mobile context
sensitive system design, and for this reason, it has been
researched in many recent works. The EEMSS system [6]
is one of the most impressive works. This platform used a
hierarchical sensor management strategy to recognize
user states as well as to detect state transitions. Kiran et.
al.[3] proposed dynamic adaptation methods. Kang et. al.
[2] proposed SeeMon, a context monitoring service for
mobile devices based on several sensors, and it achieves
energy efficiency by performing context recognition only
when the there is a change in the context. Musolesi et. al.
[4] showed that continuous sensing is a viable option for
mobile phones by adopting efficient data uploading (to a
remote server) strategies. Ra et. al. [5] addressed the
problem of energy-accuracy trade-offs in smart phone
applications. Abdesslem et. al. [7] devised SENSELESS
systems which uses accelerometer extensively to sense
the movements and reduces overall energy consumption.
Jatinder Pal Singh [8] provided a method for GPS
location sensing with improving energy efficiency. The
Cell2Noitify [9] protocol developed for the management
of energy in smart-phones using VOIP over wireless
network. E-Watch [1] sensor platform consists with
several sensors like Bluetooth, accelerometer, GPS etc
and perform good to tackle energy-accuracy trade-offs.
Finally, several energy saving schemes for mobile
devices are discussed and compared by Wang et. al. [6].
II. CONTEXT-AWARE COMPUTING
Context-aware systems consist of several types of
sensors like accelerometers, location-sensors,
temperature-sensor, voice-sensor, light-sensor etc.
Sensors used to sample the input provided by the
changing-context and process that data on the defined
ontology and inference rules to produce calculated
output. As sampling frequency is directly proportional to
the accuracy of sensor [1], and also to the energy
consumption therefore it is inferred that there is a trade-
off between energy and accuracy of a sensor. In this
paper we propose a mechanism whereby energy-accuracy
trade-off can be reduced up-to 22%. The samples here are
taken by Sony Ericsson K530i with battery specification
3.6 volt, 950 mAh. Block diagram (Fig-1) shows the
mechanism of a sensor equipped with back-off and
advance functions.

  SENSE  SENSE‐
WORTHY 
EVENTS 
SENSE ‐
UNWORTHY 
EVENT 



ADVANCE 
FUNCTION 
BACK‐OFF 
FUNCTION 
SLEEP 


Figure-1: Block diagram of mechanism of a sensor
Sensors can react to the two types of events [3] i.e.
sense-worthy and sense-unworthy. The events which are
recognizable to the sensors are known as sense-worthy
events. The sensor uses sampling algorithms to sample
these events and consumes energy from battery. In order
to distinguish among these events CPU of device is
programmed at design time which is powered from the
battery for its operation. For a microphone a sense-
worthy event is some available voice signals and for an
accelerometer a sense-worthy event is something that is
moving. On the other hand those events for which sensors
are not aware are called sense-unworthy events. A sensor
can neglect these types of events by not wasting its
sampling-cycle.
Umesh Chaudhary , Chandrika Prasad

umesh9794@gmail.com, chandrika.bhumca08@gmail.com
consuming phenomenon. In the modern era of computing it
III. MINIMIZING ENERGY CONSUMPTION
In order to insinuate the energy conservation,
adaptive sampling based on dynamic method [3] can be
used. A sensor works on the “sense and sleep” cycle
which is shown in figure 1. Here back-off functions are
used by the sensors to control their sampling interval
during sense-unworthy events and advance functions are
used to control sampling interval. During these events, it
is devised to choose back-off and advance functions
strategically so as to minimize energy consumption.
Back-off Function - This function controls the sampling
interval in context of sense-unworthy events so it was
taken as-
f(x)=(√|sin(x)|) where x is the sleep interval.
The function values ranges from 0 to 1. Hence the
sensor can’t delay more than 1 second while taking
samples and for a value other than boundary value the
sampling interval is greater than sleep interval.
Advance Function - This function controls the sampling
interval in context of sense-worthy events so it was taken
as-
f(x)=(|sin(x)|) where x is the sleep interval.
This function also ranges from 0 to 1 but the
largest boundary value is monolithic for the sense-worthy
events, it can be reduced by adjusting sleep interval
according to the context.
The back-off function and the advance function
were then tested for Bluetooth sensor and microphone
sensor of K530i.
IV. RESULTS AND DISCUSSION-
Relationship between minimum sampling interval
(in seconds) and accuracy of the Bluetooth sensor has
been studied (fig-2). The graph depicts an inverse
relationship between accuracy and sampling interval.
0
0.2
0.4
0.6
0.8
1
0 0.5 1 1.5
Figure-2: Accuracy (Y-Axis) vs sampling interval in seconds
(X-Axis) for Bluetooth.
Figure-3: Energy consumed in Joule per hour (Y-Axis) vs
sampling interval in seconds (X-Axis) for Bluetooth.
Another relationship between energy consumed
per hour (Joules) and minimum sampling interval (in
seconds) in the Bluetooth sensor has been studied (fig-3).
Energy consumption decreases gradually. Further, using
the relation between energy and current (E=Vit, where
V=Voltage (volt), i=current (Ampere), t=time interval (in
seconds)) we can infer the relationship between battery
life and the sampling interval of device (fig-4).
0
5
10
15
20
25
30
0 0.5 1 1.5
Figure-4: Battery life in hours (Y-Axis) vs sampling interval in
seconds (X-Axis) for Bluetooth.
Similarly, using the above defined trade-off and
advance functions we have taken samples of microphone
sensor. The plot between accuracy and sampling interval
(in seconds) is prepared (fig-5). The plot presents that
accuracy decreases with respect to increment of sampling
interval while sampling from microphone.
0
0.2
0.4
0.6
0.8
1
0 0.5 1 1.5
Figure-5: Accuracy (Y-Axis) vs sampling interval in seconds
(X-Axis) for microphone.
The graph (fig-6) represents the cumulative
behavior of energy consumption in Joules per hour and
sampling interval in seconds. It clearly depicts that the
energy consumption also decreases with the increment in
sampling interval
0
50
100
150
200
0 0.5 1 1.5
Figure-6: Energy consumption in Joules per hour (Y-Axis) vs
sampling interval in seconds (X-Axis) for microphone.
The graph (fig-7) represents the spatial
relationship between sampling interval in seconds and
battery life in hours. The graph shows that by increasing
the sampling interval, battery life increases gracefully.
We observed that all these curves stabilize at
certain values. Therefore, high values of these intervals
do not necessarily imply low accuracy and high savings
in energy. It should suffice to use the values after which
there are no considerable improvements in terms of
performance. Above described functions appear to be the
best in terms of accuracy as compared to the other
functions for most of the cases; however, it is not always
the worst in terms of energy consumption.
0
2
4
6
8
10
12
0 0.5 1 1.5
Figure-7; Battery life in hours (Y-Axis) vs sampling interval in
seconds (X-Axis) for microphone.
Figure-2 reveals that the sampling interval of 0.3
seconds provides dominant results with respect to other
sampling rates as while taking sampling interval 0.3
seconds, accuracy of Bluetooth sensor remains 80%. Also
as per figure-3 the energy consumption is 88 Joules per
hour and figure-4 states the battery life of the
experimented device is 11.98 hours which is a better
performance in comparison with other back-off and
advance functions experimented previously on Bluetooth
sensors. Similarly figure-5, figure-6 and figure-7 describe
the results for microphone sensor. In figure-5 sampling
interval of 0.3 seconds suffices 80% accuracy and from
figure-6 energy consumption found to be 155 Joules per
hour. Also from figure-7 it is noted that fully charged
battery survives 6.4 hours with this sampling interval.
The results achieved reflect an approximate gain of about
22% over the results of dynamic adaptation methods used
in microphone sampling.
V. FUTURE EXPECTATIONS
Ubiquitous computing is an emerging research
field of implementations in which many computing-
elements are present everywhere and abstracted from
human-beings. Therefore, future hand-held devices will
be equipped with numerous real or virtual sensors to
increase context-awareness. Hence it will be highly
desired that the accuracy of these sensors should be high
as well as the energy consumption should be minimized.
So energy management will be the most challenging issue
in those implementations. It will really influence nano-
computing as how to compensate the nano particles to
store charge which can fulfill the energy requirements of
peripheral of computing-elements. So we find energy
management will play a vital role in future computing
techniques.
VI. CONCLUSION
In this paper we proposed how energy can be
conserved in other hand-held devices and wearable
sensors. The experimental results are experienced with
Sony Ericsson K530i having battery configuration of 3.6
volt, 995 mAh. Results clearly reflect that using the
devised back-off and advanced function we can
experience the lower battery-power consumption by
Bluetooth as well as microphone sensors. The graphs
plotted above against accuracy vs. sampling interval,
energy consumed vs. sampling interval, and battery life
vs. sampling interval, each for Bluetooth and
microphone, highlights the energy conservation achieved.
Another significant finding is that with the taken back-off
and advance functions, we find that the sleeping time as
well as sampling interval is not more than 1 second. This
incredible nature of these functions allows the sensors to
maintain their accuracy. These functions can also be
implemented in real time systems which support delay
not more than 1 second. The increasing and decreasing
nature of the defined functions take place in a particular
interval so if we are talking about a primary sleep interval
of more than 1 second then all next sleep intervals and
corresponding sampling intervals will be within 1 second.
Hence the gain on battery life is found while taking
samples. These functions perform better than dynamic
adaptation method functions in context of sensors in
hand-held devices.
REFERENCES
[1] Andreas Krause, Matthias Ihmig, Edward Rankin,
Derek Leong, Smriti Gupta, Daniel Siewiorek, Asim
Smailagic, Michael Deisher, Uttam Sengupta-“
Trading off Prediction Accuracy and Power
Consumption for Context-Aware Wearable
Computing” 2005 IEEE.
[2] S. Kang, J. Lee, H. Jang, H. Lee, Y. Lee, S. Park, T.
Park, and J. Song. SeeMon Scalable and Energy-
efficient Context Monitoring Framework for Sensor-
rich Mobile Environments. In Proceedings of
MobiSys '08, 2008.
[3] Kiran K. Rachuri, Mirco Musolesi, Cecilia Mascolo-
“Energy-Accuracy Trade-offs in Querying Sensor
Data for Continuous Sensing Mobile Systems” 2010.
[4] M. Musolesi, M. Piraccini, K. Fodor, A. Corradi, and
A. T. Campbell. Supporting Energy-Efficient
Uploading Strategies for Continuous Sensing
Applications on Mobile Phones. Proceedings of
Pervasive '10, 2010.
[5] M.-R. Ra, J. Paek, A. B. Sharma,R. Govindan, M. H.
Krieger, and M. J. Neely. Energy-Delay Tradeoffs in
Smartphone Applications. In Proceedings of
MobiSys '10, 2010.
[6] Y.Wang, J.Lin, M. Annavaram, Q. A. Jacobson, J.
Hong, B. Krishnamachari, and N. Sadeh. A
Framework of Energy Efficient Mobile Sensing for
Automatic User State Recognition. In Proceedings of
MobiSys '09, 2009.
[7] Fehmi Ben Abdesslem, Andrew Phillips, Tristan
Henderson-“Less is More: Energy-Efficient Mobile
Sensing with SenseLess”. In Proceedings of
MobiHeld’09,2009.
[8] Zhenyun Zhuang, Kyu-Han Kim, Jatinder Pal Singh-
“Improving Energy Efficiency of Location Sensing
on Smartphones” In Proceedings of MobiSys '10,
2010.
[9] Yuvraj Agarwal, Ranveer Chandra, Alec Wolman,
Paramvir Bahl, Kevin Chin, Rajesh Gupta-“Wireless
Wakeups Revisited: Energy Management for VoIP
over Wi-Fi Smart-phones”. In Proceedings of
MobiSys '07, 2007.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-1

1

Abstract—Mobile ad hoc networks (MANETs) can be
defined as a collection of large number of mobile nodes
that form temporary network without aid of any existing
network infrastructure or central access point. Each node
participating in the network acts both as host and a router
and must therefore is willing to forward to packets for
other nodes. The characteristics of MANETs such as:
dynamic topology, node mobility, provides large number of
degree of freedom and self-organizing capability of that
make it completely different from other network. Due to
the nature of MANETs, to design and development of
secure routing is challenging task for researcher in an
open and distributed communication environments. The
main work of this paper is to address the security issue,
because MANETs are generally more vulnerable and we
proposed a secure routing protocol for MANETs, are
named ASRP (Authenticate Secure Routing Protocol)
based on DSDV. This protocol is work on various modes;
each mode corresponds to specific state of the node. This
protocol is design to protect the network from malicious
and selfish nodes. We are implementing Extended Public
key Cryptography mechanism in ASRP in order to achieve
security goals.
Keywords: - MANETs, Security, Cryptography.

I. INTRODUCTION
HE Wireless cellular system has been in use since
1980s. Wireless system operates with the aid of
a centralized supporting structure such as an access
point. These access points assist the wireless users
to keep connected with the wireless system, when
they roam from one place to other. In wireless
system the device communicate via radio channel to
share resource and information between devices.
Due to presence of a fixed supporting structure,
limits the adaptability of wireless system, so this
generation of wireless system is required easy and
quick deployment of wireless network. Recent
advancement of wireless technologies like
Bluetooth [3], IEEE 802.11 [4] introduced a new
type of wireless system known as Mobile ad-hoc
network (MANETs) [1, 2, 5, 6], which operate in
the absence of central access point. It provides high
mobility and device portability’s that enable to node


connect network and communicate to each other. It
allows the devices to maintain connections to the
network as well as easily adding and removing
devices in the network. User has great flexibility to
design such a network at cheapest cost and
minimum time. Mobile ad hoc network consist large
number of node, it form temporary network with
dynamic topology. In this network each node
communicates with each other through radio
channel without any central authority. In MANETs
each node operates in a distributed peer-to-peer
modes [2], serves as an independent router to
forward message sent by other nodes. MANETs has
shows distinct characteristics, such as:
Weaker in Security
Device size limitation
Battery life
Dynamic topology
Bandwidth and slower data transfer rate
Apart from these limitation MANETs has many
extensive application like: Military application,
Natural disaster, Medical service.

Figure 1.1
In ad hoc network there can be node that will try to
disrupt the proper functioning network. These nodes
can be malicious or selfish [8]. They try to disrupt
network function by modifying packets, injecting
packets or creating routing loops. So, security is an
important task, because MANETs has
characteristics such as; dynamic topology,
AUTHENTICATE SECURE ROUTING PROTOCOL
ABHAY SHUKLA (M.Tech., persuing P.hd.)
T
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-2

2
infrastructure less. There are large numbers of
secure routing protocols proposed by many
researchers they fulfill different security
requirements and prevent specific attacks. They are
divided into three categories: Reactive routing
protocol [5, 6], Proactive routing protocol [5] and
hybrid routing protocol [6]. In reactive routing
protocol the route is discovered when it required, in
proactive each node maintain network information
regarding to network connectivity and route
information to all others node within the network
and proactive is one which is neither reactive nor
proactive.
Now, the Most of the solution uses cryptography
mechanism to detect selfish, malicious behavior of
nodes and securing information from other types of
attacks. The mechanisms which are used by
different secure routing protocol to detect malicious
and selfish node have address separately in different
protocol. No secure mechanism has been proposed
till date that can address to detecting malicious and
selfish node collectively. We proposed a
mechanism, Extended Public key Cryptography
(EPKCH) [9, 10] that able to detect the malicious
nodes and selfish nodes collectively in order to
achieving security goals such as;
Authentication, Integrity, Confidentiality and Non-
Repudiation. Also, we proposed a routing protocol
named Authenticate and Secure Routing protocol
for mobile Ad hoc Network (ASRP). We
implemented EPKCH mechanism in monitor mode
of ASRP to securing MANETs. To design of this
protocol follows the table-driven approach, in which
each node maintain the information, regarding to
network structure and route from a particular source
to its all possible destination in its node info table.
ASRP is a proactive secure routing protocol.
2. PURPOSE
The secure protocol design and development has
become the most challenging task in securing
mobile ad hoc network. Most of the existing
protocol has been develop based on specific security
scenarios. So the main purpose of this research is to
understand and evaluate the existing secure protocol
and implement a secure protocol which is not
address the solution for particular security (attacks)
but it is prevent different kind of security scenarios.
3. DESIGN, DEVELOPMENT AND
SIMULATION OF ASRP
3.1 Introduction
The ASRP is a proactive secure routing protocol.
The design of ASRP follows the table driven
approach, each node maintain a node info table
regarding to network structure, route information
from a particular source to its all possible
destination and information about others node.
When a new node enters into network all the node
updates its own info table.
This means that every node have complete
knowledge about the network structure. The ASRP,
protocol works on the terminology of the modes of
node, in which each mode define the working
pattern of node. The different modes correspond to
the activity of node and each mode defines
particular state of the node in ASRP. When
MANETs is established then every node in the
Initialization Mode (IM: it is the network
initialization mode) which enables every node setup
the network infrastructure and store initial
information in its own Node Info table and also the
information regarding to network structure. When
nodes are finish initialization they switch
themselves to Lazy mode (LM), where they are
waiting for forwarding the packet. Lazy mode is the
default mode of the each node when they do not do
anything. If any node wants to forward the
PackFoward packet they switch from lazy mode to
monitor mode (MM), then packet forwarding mode
(PFM). As soon as they finished the packet
forwarding they switch to lazy mode. In lazy mode
the lazy node forward the PackLazy packet to its
neighbor node.
So, there are four mode correspond to different
activity of node. The IM responsible for the
establishment of MANET, LM is the default mode
of the node where they do not do anything. MM
mode is responsible monitoring the network and
node, while node leaving the network and joining
the network. The IM also detect the malicious and
selfish node within the network. The PFM is also
detects the malicious and selfish node during packet
forward operation.
3.2 Assumptions
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-3

3
Our proposed secure protocols aim to protect the
network from attackers. Our proposed schemes
work under several assumptions as follows:
The network link is bidirectional. That is, if
node A is able to transmit to node B, then B is also
able to transmit to A.
The wireless interface supports promiscuous
mode operations. That is, each node can receive a
copy of the messages being transmitted by other
nodes within its receiving range.
A public key infrastructure exists in the
MANET under consideration. Each mobile node
stores the public key of all other nodes.
The trust relation could be instantiated. For
example: by knowing public key of other nodes.
There is a security association between source
node and destination node.
The existence of security association is justified
because, host chose to employ a secure
communication schemes and consequently, should
be able to authenticate each other.
3.3Packet types and their Structure
There are two types of packet being transmitted by
the various nodes; conditional packet and
unconditional packet.
3.3.1 Unconditional Packets
The unconditional packets are those packets that a
node sends to neighbor which have fixed size and
irrespective of the condition.
a) Pack Hello
This packet is send by node to its entire neighbor
when the node is in IM.
Table 1.1 PackHello Packet and Structure
Packet_Id
NodeNameA1 Signatur_of_A1
PublicKey KA1+
b) PackLazy
This packet is send by the node to all its neighbour
when the node is in LM.
Table 1.2 PackLazy Packet and structure
Packet_Id
NodeName ………

PacketType PackLazy
3.3.2 Conditional Packet
The conditional packets are those packets that a
node sends after detecting the particular condition
has been satisfied. Some of the conditional packets
are fixed size and some are variable size.
I. Conditional Fixed Size Packet
a) PackUpdate
This packet is send by various nodes to their
neighbor when they detect that there is any change
in their neighbor position of the nodes.
Table 1.3 PackUpdate Packet and Structure

Packet_Id


NodeName ………

PacketType Conditional

NodeNomeother Info_about_other_node

b) PackError
This packet is send by the node to the source node
to indicate that the destination does not exists
Table 1.4 PackError Packet and Structure
Packet_Id


NodeName ………

PacketType Conditional

NodeNomeother PackError

c) PackMalicious
This packet is send by the node to its neighbor when
it finds any malicious activity by some node.
Table 1.5 PackMalicious Packet and Structure
Packet_Id


NodeName ………

PacketType Conditional

NodeNomeother Info_about_other_node or
PackMalicious


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-4

4
d) PackSelfish
This packet is send by the node to its neighbor when
it finds any selfish activity by some other node
Table 1.6 PackSelfish Packet and Structure
Packet_Id


NodeName ………

PacketType Conditional

NodeNomeother Info_about_other_node
or PackSelfish

II. Conditional Variable size packet
The variable size packets are.
a) PackInitialized
This packet is send by node to neighbor when
neighbor node sends PackHello to him.
Table 1.7 PackInitialized Packet and Structure
Packet_Id


NodeNameA1 PublicKey(KA1+)

NodeNameA2 PublicKey(KA2+)


------------
-
---------

Nonce_of_A1 TimeStamp_of_At

Number_of_node No_of_Hopes

PacketType PackInitialized

b) PackForward
This is the packet which is prepared by the source
node first then it is modifying by the all node which
are on the route to destination, by adding its own
information. All the intermediate node encrypt the
NPart of packet by the public key of source node
and verify it and then sign on it by using its own
private key.
Table 1.8 PackForward Packet and Structure
Packet_Id


NodeName ………

PacketType PackForward

PrivacyPart(PPart) Non-repudiation
part(NPart)

3.3.3General Structure of Packet
I. Fixed Size Packets
The general structure of fixe size packets for ASRP
is
Table 1.9 Fixed Size Packets Structure
Packet_Id

MAC Sender

Cond_control_bits

MACother

Info_control_bits

Table 1.10 Cond_control_bits meaning
Bits Meaning

00 Unconditional PackHello
Packet

01 Unconditional PackLazy
Packet

11 Conditional Packet

The Info_control_bits are meaningful only if the
Cond_control_bits are 11. The 3 bit for the
Info_control_bits stand for whether the packet has
left neighbour of the sender node has just joined the
network or sender node detect the MACother to be the
malicious node or the selfish node. These 3 bit stand
for the conditional packets so for the
Cond_control_bits are 11.
II. Variable Size Packets
The variable size packet structure for the ASRP is
Table 1.11 Variable Size Packets Structure
Packet_Id

MAC Sender

Cond_control_bits

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-5

5
MACother

Info_variable_size

The Info_control_bits identifies whether the packet
is PackForward or PackInitialized Packet.
Table 1.12 Info_control_bits Meaning
Bits Meaning

00 PackForward Packet

01 PackInitialized Packet

3.4Activities of Nodes
There are various activities of node in mobile ad
hoc network:
I. The node is sending HelloPacket packet to it all
neighbor node to getting the certificate and public
key of all other nodes that participate in the
network.
II. The neighbor node sends the public key and
certificate signed by its own private key, to the node
by sending PackCert packet.
III. The node will forward the PackFoward packet
that will receive from its neighbor if it is not
destination node.
IV. The node will monitor any topology change in
its neighborhood that is if any node is leaving the
network, joining the network or changing its
position with respect to neighborhood.
All these activities correspond to the specific state
of node in mobile ad hoc network. And the group of
activities represents the mode of nodes. The first
and second activity grouped into IM (network
initialized mode), the third activity will come under
the PFM (packet forward mode) and the fourth
activities come under MM (monitor mode). We
discussed every mode of the node detailed in next
section.
3.5Initializing Mode (IM)
I. Description
This mode is responsible for the establishment of
MANETs. It can be established MANET in two
ways. One way can be that all the nodes get on at
the same time and second way can be that the node
get added to the network with the posses of the
time. For the first way all the node are in the IM
simultaneously and for latter situation the node
added to the network is in IM. After initialization
mode node switches to lazy mode. In IM mode,
each node got the public key of all other nodes
participating in the network before joining the
network. In this mode a trust relationship exists
between each node that participates in the network
by sharing public key and other information.
II. Procedure
In this mode during the initialization of the
network the node, say A, send the PackHello packet
to its neighbour to get the PackInitialized packet. If
other node send the PackHello packet to the node,
say B, it send the PackInitialized to the node.
When node receives the PackInitialized packet
it updates its NodeInfo table, and broadcast
PackInfo to all node that participate in the network
which contains its public key, timestamp t,
timestamp e, address and signature and then switch
to lazy mode.
When the others node receive the information
about new node updated its NodeInfo table and also
check there is any change in the table or not. If there
is no change in the table then it switches to LM.
If the node not receives the PackInitialized to
any its neighbour it again sends the PackHello
packet.
This mode is can also help to find the malicious
nodes as they may change the public key of node
intentionally. The node can compare the public key
send by particular node with the public key send by
others nodes and if the find any deviation it send the
PackMalicious packet to other neighbour nodes.
It the node find that a particular node is not
responding with the PackInitialized packet
corresponding to its PackHello it send the
PackSelfish to its neighbour nodes.
All the nodes switch to IM from the LM if all
the node going to change their keys. Then they
perform the process of Initializing with neighbors.
The node can change their mode according to their
activity, if a node finish the Initialization process it
switch to lazy mode and if they want to be forward a
PackForward to any node then it switch lazy mode
to packet forward mode through monitor mode.
Transition between modes depends upon the activity
of nodes.
Table 1.13 Packet exchanged and Transition in IM
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-6

6
In the Network Initialization procedure, if a node is
in IM and send the PackHello Packet.

1. Start
2. Send the PackHello packet to all node which
are at the distance i=0.
3. Stop.


A. In the Network Initialized procedure, if a node is
in IM and get the PackHellopacket.

1. Start
2. Update its NodeInfo table.
3. Send the PackInitialized packets which
contain the public key all other
node and other information regarding to network
structure which are
available in its own NodeInfo. It send to that
node which are at the
distance of i=0.
4. Stop.

B. In the Network Initialized procedure, if a node is
in IM and get the PackInitialized Packet.
1. Start
2. Set change = false
3. Updates its NodeInfo table
4. Check for any change in its NodeInfo table, if
there is any change
Then
Change = True
5. If change = true
Send the PackHello again
Go to step 2.
6. Stop


III. Structure of the NodeInfo table
The node stores the information regarding the
network structure and information about other
nodes (such as public key, address, timestamp t,
timestamp e, nonce and signature to all others node
that participate in the network)
Table 1.14 Structure of NodeInfo Table
IV. Main Goals
The main goals of Network Initialized mode are
listed below
All the nodes can get the public key of all other
nodes that participate in the network.
After IM procedure ever node updates its
NodeInfo table and gets complete knowledge about
the route of all other nodes that are reachable to
source node and also get the information regarding
to the network topology.
All the node have get approximate idea about
how much distance and time (millisecond) taken to
sending a data to the particular node.
This mode can help us to detect malicious and
selfish node within the network, by comparing the
same node public key send by the different nodes.
This mode is responsible for the establishment
of mobile ad hoc network, by adding node in the
network.
The transition from IM to either lazy mode or
MM or packet forwarding mode via monitor mode.
Lazy Mode
I. Description
The LM is the default mode of the node. The node
switches this mode either from IM, or from MM, or
from PFM. When MANET is established every
node are in the network initializing mode, as soon
as they finished the network initialization process it
switch form IM to LM. The node is in PFM, and if
there is no packet to be forwarded they switch LM.
In this mode, if all the nodes want to change their
public keys it switch from lazy mode to initializing
mode. If the node in LM and detect any
neighborhood activity it switch to MM or it has
receive any PackForward packet it switch itself to
from lazy mode to packet forwarding mode through
monitor mode, and after finishing packet forwarding
Mode Packet
exchanged
Transition to
Mode

IM PackHello,
PackInitialized
MM, LM

NodeName Publi
c key
Nonc
e
Timestamp NextNodeName NumberofHop
s

----- ----- ----- ---- ---- ----

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-7

7
procedure it again back to LM, and forward
PackLazy packet to all neighbor nodes.


Figure 1.2 Transitions between Modes in LM
II. Procedure
The Lazy mode is called the default mode of
the node.
As soon as the node finished the network
initialization they switch from IM to LM.
If there is no PackForward packet to be
forwarded they switch from PFM to LM.
When the node in LM detect that there is
activity of neighbor node it switch itself to MM
If the PackForward packet comes from the
neighbor node then it switch them to PFM through
MM and after finishing packet forwarding
procedure it comeback to LM.
If all the nodes want to change their public key
then they are also switch LM to IM.
In the nodes comes in the Lazy mode it send the
PackLazy packet to all the neighbor node.
Monitor Mode
This mode is basically for monitoring the network
topology when nodes leaving the network, joining
the network or changing its position with respect to
neighborhood. This mode also detect malicious and
selfish node. If the node in mode found any activity
of neighbour node or receives any PackMalicious
Packet or PackSelfish Packet or PackForward
Packet it switches itself to MM. This is the main
mode for ASRP, it provide high level of security.
This mode is also called the protector mode. These
above condition can divided into two part; general
condition and special condition.

Figure 1.3 Transitions between Modes in MM

A. General Condition
I. Procedure for MM when node are in Lazy mode
and detect any activity of neighbor nodes

II. Procedure for MM when nodes are in Lazy mode
and receive PackMalicious or PackSelfish or
PackForward to neighbour node.

B. Special Condition
Nodes joining the network, nodes leaving the
network and nodes are changing its position within
the network. These three special conditions are also
handling in monitor mode by sending or receiving
PackUpdate packet. These three conditions are
explained below. In the case of new node joining
the network, say node A want to join the network.
There are two cases possible, first is the network
has no any node initially and the second case is
node A join the network through nodes B1 to Bn. in
first case node A enter into network without any
condition. In second case if n=1, then it join through
B1 and if n=2, then it join through B1 and B2.
1. Start
2. Set Flag = 1 (i.e. If there is any activity of
nieghbour, then)
Switch to MM
Else flag = 0
End loop
3. It check the neighbour node, for malicious or
selfish , if yes then send
PackSelfish to all neighbour node
4. If not then return to LM
5. Stop.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-8

8
a. The Procedure for node joining the network is
given below
1. Start
2. Flag Join = False
3. While ( Join = False)
The node A checks that if at least one node out of
all the node through which it is
going to join the network. if there is one node, then
Join = True
Else Join = False
End Loop
4. The node A switch to IM and send the
PackHello packet to nodes out of B1 to Bn
which are in LM.
5. The node which are in LM, out of B1 to Bn
switch themselves to IM, when it receive
the PackHello packet from node A, then updates it
NodeInfo table and then it send the
PackInitialized Packet to A.
6. Node A match all the PackInitialized packet
send by B1 to Bn if there is no change,
then
Node A Update its NodeInfo table and send its
own information to all node within the
network
All node in MANET Updates its own NodeInfo
table
Else send PackMalicious to all neighbours.
7. After finishing IM procedure they switch to LM
8. The resident nodes again switch to LM.
9. Stop.
b. Procedure when the node, say D changes its
position or detects the change in its neighborhood or
node D leaving the network.
Packet Forward Mode
The ASRP is proactive secure routing protocol, so
when MANET is established every source node
know the route to the all its possible destination
node which are reachable from it. The source node
then prepare the packet and send it to first node
which are on the route of destination node, then first
node send to next in this way the data is reaches to
the destination node. Now the packet that is sent,
consist two parts
PPart NPart


The first part i.e. Privacy part contain the data
which is to be transmitted is encrypted by source
node by using public key of destination node. The
NPart i.e. non-repudiation part contains the address
of destination node, a nonce of destination node and
timestamp, and the address of all the node along
they transmitted. It also contains the address of
source and signature of source node which is
encrypted by private key of source node. Every
intermediate node along the decrypt the NPart of
Packet and verify its addresses, if itself destination
node or not then forwards the packet to the next
node in the path. This way packet reached to
destination node. The NPart contains only the
information related to route. This mode detect
malicious and selfish node by verifying NPart of
packet. It also detect if link is broken to destination
node or if destination not exists by generating
PackError packet. If any node behaves as a selfish it
detect by resending packet again. If packet send by
source node to particular destination and the
intermediate node not find the particular destination
it send PackError packet to source node.
The data forwarding process is better understood by
using an example. Let us suppose the source node
want to send the data in PackForward packet has to
destination node along the way to destination
through the source A B C D destination.
1. Start
2. Check any change in its neighbors, if node D
leave the network, then
3. The neighbor node E of D send the
PackUpdate to all its neighbor
4. If node D change its position, then D send
the PackUpdate to its neighbors
5. The neighbor node after receiving it switch
to MM, and updates their NodeInfo
table and send PackUpdate to all its own
neighbors
6. Stop.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-9

9
Source A {PPart, NPart}
PPart =EPA (EPB (EPC (EPDestination (Data, Source, t,
n, C), B,) A), Source)
NPart = EPRSource (A, B, C, (Destination, t, n)
KSource-
A B{PPart, NPart}
PPart = EPB (EPC (EPDestination (Data, Source, t,
n, C), B), A)
NPart = EPRA (B, C, Destination, t , n) KBB
C {PPart, NPart}
PPart= EPC (EPDestination (Data, Source, t, n, C), B)
NPart = EPRB (C, Destination, t, n) KCC
EPDestination {PPart, NPart}
PPart= EPDestination (Data, Source, t, n, C)
NPart = EPRC (Destination, t, n)


Additional Security Features
The others security features enhanced the security of
MANET, are given below.
I. The key of the nodes have to be change regularly
after some specific time interval, so as keep the
MANET secure all time.
II. The dummy packet forwarding operation can be
implemented by the intermediate node to detect the
node who misbehaves.
III. To detect malicious and selfish during the
network initializing process by verifying public key
and signature send by all neighbour node to new
node.
IV. In case of special scenario of military operation,
like some nodes are to join the network latter on,
only those nodes join which has the MAC address
that has been already verify
Inbuilt Defense
The various inbuilt defense are in ASRP, to
securing MANET, these are:
I. Cooperation Enforcement
The ASRP is the protocol that works by enforcing
the cooperation between the nodes. That makes the
nodes are more responsible for their action and the
probability of the non-intentional behavior of the
node also become less. The MANET does not pass
the any infrastructure so the nodes have to cooperate
fully in order to enforce the operation on the nodes
in the correct way.
II. Robust Modular Implementation
There are four modes of the ASRP each mode has
its own modular implementation and independent
working. That paves the way for the modification of
the working of the modes in the future as the need
arises without modifying the working of the other
modes. Moreover the modular approach is best for
the debugging also.
Simulation
We are using turbo C, for the simulation, has been
performed to simulate the transfer of data between
various nodes in initializing mode, lazy mode,
packet forward mode and monitor mode. The
simulation has been performing on eight nodes.
Figure 1.4 Simulated MANET
The simulation program after taking the input
performs the simulation as mention below.
I. Network input by adding the node in form of
adjacency list
II. Simulate the packet transfer in the Initializing
Mode
III. Simulate the packet transfer in the Lazy Mode
IV. Simulate the packet transfer in the Packet
Forwarding Mode
Since the ASRP is a proactive secure routing
protocol, so in every step it displays the status of
tables of all the nodes. The simulation step along
with screenshot has been given below. I. Firstly the
simulation program takes the input, in terms of
asking for the number, name and name of neighbour
of nodes.


5
3
6
0
1
4
7
2
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-10

10

Figure 1.5 Network Input
a) The simulation program displays the network
enters in terms of adjacency list.


Figure 1.6 Adjacency list of Nodes in the Network

b) The simulation program input the time (in
millisecond) to sending the data between
each nodes.

Figure 1.7 Input Times in Millisecond
II. Program to display the packet during the IM, i.e.
one node sends the PackHello packet to other and it
sends the PackInitialized to previous node.
Initially we assume that node 0 is only single node
in MANET, and then enter node 1, and then 2, and
so on.


Figure 1.8 Step II Network Initialization Mode
IV. Packet Transfer in Monitor Mode
In this mode if a node detect any activity of
neighbour node it goes to MM and send a
PackCheck to all its neighbour for detecting
Malicious node.
a) After step II every source node shows the
route information to its all destination, and
time in millisecond.




Figure 1.9 Information about the route of all
destinations from a source

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-11

11
b) Network after (establishing MANET) step II.


Figure 1.10 MANET after step II
III. Packet Transfer in Lazy Mode

Figure 1.11 Packet transfer during Lazy Mode
IV. Packet Forwarding Procedure

Figure 1.12 Shows the data are going to source to
destination via same route.
Conclusion and Future Scope
Wireless mobile ad hoc networks present difficult
challenges to routing protocol designers. Mobility,
constrained bandwidth, and limited power cause
frequent topology changes. The very basic nature of
the mode of communication is the main concern
because anything that moves over the open air
medium is susceptible to be picked up by
unauthorized access. For any mission critical or
organizationally sensitive information, ad hoc
networks add an element of insecurity. In the existing
secure routing protocols most of the security attacks are
possible with a compromised node. In this work we focus on
how to detect malicious and selfish node and to design and
implement a secure routing protocol.
In ASRP protocol we discuss various activity of node which
they are shown during the MANET operation and these
activities are grouped into modes along their working. We also
discussed the packets that are going to be exchanged in
different mode of nodes. The conclusion that comes are given
below.
I. The problems of malicious and selfish node are handling
simultaneously. We discussed Extended Public key
cryptography mechanism to handle the malicious and selfish
node during network operation. As the selfish node cannot
malicious at same time, but if nodes are not malicious then
they may be malicious.
II. The protocol is handling the some special situation like
nodes joining the network, node leaving the network and nodes
are changing its position within the network. The monitor
mode of ASRP handles all three situations.
III. In ASRP, there are four modes, the IM corresponds to
network initialization phase, the LM corresponds to default
phase, PFM responsible for forwarding the packet form source
to destination and MM is the protector mode of the network.
IV. The protocol has to develop in the way so that the future
modifications are possible without changing overall protocol.

Future Scope
The proposed secure routing protocol, ASRP, is a
proactive routing protocol based on table driven
approach. For the future work we can use the hybrid
approach or reactive approach in ASRP to
implement a new secure routing protocol. We can
also add another mode or existing one can be
extended to handle some exceptional conditions.
The public key cryptography algorithm can also be
extended to securing MANET.
References
1] C. Perkins, “Ad hoc Networks,” Addison-
Wesley, 2001.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO207-12

12
[2] M. Ilyas, “The Handbook of Ad Hoc Wireless
Networks,” CRC Press, 2003.
[3] Chatchik Bisdkian, “An overview of Bluetooth
Wireless technology”IEEE Communication
Magazine, Vol. 39, No. 12, pp 86-94 December
2001.
[4] Brian P. Crow, Indra Widjaja, Jeon Geun Kim
and Prescott T. Sakai, “ IEEE 802.11 Wireless
Local Area Network,” IEEE Communication
Magazine, Vol. 35, No. 9, pp 116-126, September
1997.
[5] C.K.Toh, “Ad Hoc Mobile Wireless Networks:
Protocols and Systems,” Prentice Hall Englewood
Cliff, NJ 07632, 2002
[6] C. Murthy and B.Manoj, “Ad hoc Wireless
Networks: Architectures and Protocols,” Prentic
Hall PTR,2005.
[7] IETF MANET Working Group. Mobile Ad Hoc
Networks (MANET). Working Group, Charter
available at
http://www.ietf.org/html.charters/manet-
charter.html.
[8] Sonja Buchegger and Jean-Yves Le Buddec,
“Increasing Routing Security in Mobile ad hoc
Network,” IBM Research Report: RR 3354, 2001
[9] B. Schneier, Applied Cryptography,Wiley, 1996.
[10] A. Salomaa, “Public-Key Cryptography,”
Springer-Verlag, 1996.






CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO209-1














Abstract— This paper describes a study on RF
attenuation path loss behavior in suburban coverage
areas in Haryana. Hata model is the most popular model
that extensively used in Europe and North America. The
model developed by Y. Okumura and M. Hata and is
based on measurements in urban and suburban areas at
Japan. There are practical problem faced in India if this
model is used for the prediction of the signal in suburban
areas of India. The main problem is that the path loss
calculated with Hata model for less than 1km shows
large practical difference. This difference can be due to
the empirical Okumura Hata model is not according to
the land profile of India, building and environmental
conditions are different from the basic inferences used in
formulation of this model in Japan and the Vegetation
and foliage losses etc. are different. An effort to improve
the Hata empirical model for sub urban area is done in
this research paper.

Keywords- Path Loss; Okumura Hata model; Optimization;
Suburban


I. INTRODUCTION

e are facing degradation of the mobile phone signal
due to various obstacles between base station and
mobile stations in suburban area. Thus, an accurate
path loss model is useful to predict the signal strength at
suburban area as well as to achieve a high quality of signal.
Path loss models are important for predicting coverage
area, interference analysis, frequency assignments and cell
parameters which are basic elements for network planning
process in mobile radio
systems. The propagation models can be divided into three
types of models, namely the empirical models, semi-
deterministic models and deterministic models. Empirical
models are based on measurement data, statistical properties
and few parameters. Examples of this model category will be
Okumura model and Hata model. Semi-deterministic models
are based on empirical models and deterministic aspects,
examples being the Cost-231 and Walfsh-Ikegami.
Deterministic models on the other hand are site-specific,
requires enormous number of geometry information about
the city, computational report and more accurate model.
The literature review shows that a lot of analytical


andexperimental works have been done for radio
propagation in mobile environment by using measurement as
in [1, 2]. In this paper, Hata-model is selected for
optimization using







numerical analysis and simulation in Matlab since this model
shows good performance compared to other models. The
field measurement data was collected using suitable
equipment for outdoor measurements and is then compared
with the simulation results.

II. THEORETICAL ANALYSIS
Hata model is the most popular model that extensively used
in Europe and North America. The model developed by Y.
Okumura and M. Hata [3] and is based on measurements in
urban and suburban areas at Japan in 1968. Validity range of
the model is frequency f
c
between 150 MHz and 1500 MHz,
TX height h
b
between 30 and 200 m, RX height h
m
between
1 and 10 m and TX-RX distance r between 1 and 10 km.
Hata's Equation is classified into three models [4]:

Rural: open space, no tall trees or building in path
Suburban area: Village Highway scattered with
trees and house with some obstacles near the mobile
but not very congested.
Urban area: Built up city or large town with large
building and houses.

Definition of parameters:

h
m
: mobile station antenna height above local terrain
height [m]
d
m
: distance between the mobile and the building
h
0
: typically height of a building above local terrain
height [m]
h
te
: base station antenna height above local terrain
height [m]
r: great circle distance between base station and
mobile [m]
R = r x10
-3
great circle distance between base station
and mobile [km]
f: carrier frequency [Hz]
f
c
= f x 10
-6
carrier frequency[MHz]
= free space wavelength [m]
Path loss for Hata-model is defined as below:

(i) Urban area path loss (dB)
L(dB) = 69.55+26.16 log f
c
-13.82 log h
te
+(44.9-
6.55 log h
te
) log R-E (1)


W
Improvement of Hata Propagation Prediction
Model for Suburban Area


Himani Kaur*, Er. Ravneet Singh **

*A.B.E.S. Institute of Technology, NH-24, Vijay Nagar, Ghaziabad (UP) 201009
** JTO, BSNL India



CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO209-2


(ii) Suburban Areas Path Loss,
L(dB) = L(urban) -2 4 . 5
28
log
2
c
f
(2)

(iii) Rural path loss,
L (dB) = L(urban) - 4.78 (log f
c
)
2
+ 18.33 log
f
c
+ 40.94 (3)

For urban area divided into:

For large cities:
E = 3.2 [log (11.7554h
m
)]
2
– 4.97 f
c
400 MHz
E = 8.29 [log (1.54h
m
)]
2
– 1.1 f
c
200 MHz

For small and medium-sized cities:

E= [1.1 log (f
c
) – 0.7] h
m
- [1.56 log (f
c
) – 0.8]


The area of Haryana is mostly is as sub-urban area. Thus,
the Hata Equation being used as below:
L (dB) = L (urban) - 2 4 . 5
28
log
2
c
f
(4)


As like any other empirical propagation models, Hata model
contains three basic elements [5, 6]:
Initial offset parameter,
Initial system design parameter,
Establish slope of the model curve.

L (dB) = 64.54 + 22.16 log (f
c
) – 13.82 log (h
te
) + [44.9 –
6.55 log (h
te
)] log (R) –E (5)

From Eq (5), the initial offset parameter is fixed at:
E
0
= 64.54
From Eq (5), the initial system design parameter may be
expressed as:
E
sys
= 22.16 log (f
c
) –13.82 log (h
te
)
The slope of model curve is a constant as expressed from
Eq. (5) as follow:
sys
= [44.9 –6.55 log((h
te
)]


III. METHODLOGY

The optimization process was obtained by using least square
method. The theoretical path loss is to be compared with the
measured data in the field. As the suburban area is to be
optimized the BTS should be selected in the different
suburban areas of Haryana. The optimization process can be
done with the help of data extracted from the various
measured log files recorded in the field with the TEMS Kit.
The accuracy of the result will be based on as many data of
different BTS is available for post processing. The
measurement data is available in the log file [7].The details
of planning data for the selected sites consists of Lat/long,
Height of Antenna, Azimuth, BCCH frequency, Cell ID etc
in TABLE I.
TABLE I
PLANNING DETAILS OF 2G SITES
Site name Azimuth Long Lat
Ant.
Height
ARFCN
KABLANA
15
76.747 28.621
37 75
165 37 71
260 37 87
BUPANIA
0
76.824 28.636
38 111
120 35 73
190 38 71
DABHODA
KHURD
350
76.825 28.662
35 76
80 35 71
180 35 87
GUBANA MAJRA
0
76.845 28.603
35 74
120 35 80
240 35 88
SAHU
0
75.832 29.493
38 112
120 38 87
230 38 76
DABWALI ROAD
340
75.024 29.552
37 113
120 37 76
200 37 78
SAWANTKHERA
140
74.768 29.907
37 78
235 37 76
320 37 112
OTTU
50
74.888 29.501
37 78
120 37 111
305 37 71

IV. OPTIMIZATION OF MODEL

Figure 1 shows the optimization process to optimize the Hata
propagation model. Optimization process was obtained by
using least square method. The optimized model will be
validated based on sub urban areas of Haryana. Statistical
analysis such as relative error, standard deviation and
variance were used to compare between the optimized model
and other known models.



Fig. 1. Optimization flow diagram

Algorithm for Optimization
For the optimization into the Hata model, initial system
design parameter E
sys
and slope parameter
sys
are
considered [8].
In this paper, a new appropriate tool is proposed in
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO209-3


optimizing to fit the measurement data with Hata model. The
condition of a best fit of the theoretical model curve with a
given set of experiment data would be met if the function of
sum of deviation squares is minimum :


P (a, b, c…) = min ,..... , , ,
2
1
n
i
i R i
c b a x E y

Where:
y
i
= measurement result at the distance x
i
;
E
R
(x
i
; a; b; c; ) = modeling result at the x
i
based on
optimization;
a, b, c = parameter of the model based on optimization;
n = Number of the experiment data set.

All differential of the P function should be equal to zeros:

. ;......... 0 ; 0 ; 0
c
P
b
P
a
P
(6)

Solution of Eq. (6) may simplify based on Parameters of the
P function are expressed simplified as:

a = E
0
+ E
sys
b =
sys

This would then mean that the expression of the Hata model
in Eq. (5) transformed into:
E
R
= a + b. log R (7)
Simplified logarithm base log R = x, the Eq. (7) becomes
E
R
= a + b. x
E
R
is the path loss (dB).

Both factors a and b are constant for a given set of
measurements. The solution of Eq. (6) may be expressed:


0 1 . ). , , , (
1
i i
n
i
R
i R i
bx a y
a
E
c b a x E y


0 . ). , , , (
1
i i i
n
i
R
i R i
x bx a y
b
E
c b a x E y


By re-positioning of elements in above equations would
provide the following expression:

i i
y x b a n.
) . (
2
i i i i
y x x b x a (8)

By substituting the variable a and b into equation eq. (8), this
would give the statistical estimates of parameter a and b:

2
2
2
~
) ( .
. .
i i
i i i i i
x x n
y x x y x
a ;

2 2
~
) (
. .
i i
i i i i
x x n
y x y x n
b ; (9)

Eq. (9) allows for calculation of parameter a and b of the
Hata model. The offset and slope parameters in the original
Hata model can be calculated from Eq. (5):
sys o
E a E
~ ~
;
te
h
b
log . 55 . 6 9 . 44
~
~



V. CALCULATIONS

The log file exported contains information Time, MS, Frame
Number, Direction, Message Type, Event, EventInfo,
Latitude, Longitude, Cell Name, Cell Id, ARFCN BCCH, Rx
Level (dBm). Data is collected with the help of software tool
that consists of test MS, Global Positioning System Receiver
Set (GPS system) and a laptop equipped with TEMS
software for various measurements in the field i.e Lat/long,
Rx power, Cell ID, ARFCN, various other events. The
measured data is to be collected in the dedicated mode i.e
Active Call from MS. The log files are to be exported in MS
Excel files with the help of TEMS Investigation 9.0 software
tool [9]. The parameters used for various calculations as
shown in TABLE II and the planning details of 2G BTS sites
selected the log files are to be processed. The measured data
is to be collected in the dedicated mode i.e Active Call from
MS. The base station information is required for each base
station, its location in longitude, latitude, antenna height,
transmit frequency, receive frequency, equivalent
isotropically radiated power (EIRP), bearing and base
station name, its power link budget for the calculation of
path loss[10].

TABLE II
PARAMETERS FOR CALCUATIONS

Mapinfo 9.0 tool is used for plotting of sites as well as Rx
power level with the help of TEMS export [11]. MS Excel
for processing of log files. MATLAB is to be used for the
calculation and plotting of the various results.


S no Parameters Values
1 Base Station Transmitter Power 43dBm
2 Mobile Transmitter Power 30 dBm
3 Base Station Antenna Height 27-58m
4 Mobile Antenna Height 1.5m
5 Transmitter Antenna Gain 17 dBi
6 Threshold Level for Mobile 102dBm
7 Threshold Level for Base Station 110dBm
8 Frequency Band 900MHz
9 Connector Loss 2dB
10 RF Cable Loss 1.5 dB
11 Duplexer loss 1.5 dB
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO209-4


The results from the different exported log files for each 2G
BTS site in processed in excel as discussed in previous
chapter. The MS Excel for each site is processed with
Matlab program as per eq. (5) for the calculation of values a
and b in eq. (9). The value for each site for a and b is
recorded in a TABLE III as shown in

TABLE III
VALUES OF A AND B
Site name Long Lat Value of a Value of b
KHERI JAT 76.7853 28.5971 127.44 15.7066
KABLANA 76.7474 28.6214 123.211 11.4782
BUPANIA 76.8248 28.6362 123.975 12.9398
DABHODA KHURD 76.8259 28.6628 124.572 7.5206
GUBANA MAJRA 76.8459 28.6033 123.227 8.2963
SAHU 75.8327 29.4936 133.073 12.1442
DABWALI-2 74.6984 29.9561 134.098 17.1094
DABWALI RD 75.024 29.5522 130.229 14.0702
DABWALI-3 74.7102 29.964 130.8 6.3378
OTTU 74.8883 29.5008 126.345 7.1927

The value of a and b is to be averaged and the final value is
calculated from the Table III

On taking average the value are as below:-
a = 130.706 b = 9.827

The value of E
o
and β is as below:-
(10)



V. COMPARISON OF RESULTS

The modified equation derived is to be validated on different
sites in the suburban area of Haryana and for each site the
error is calculated to validate the results and to be compared
with the results of original Empirical Okumura Hata model.
The results are verified on various sites in suburban area and
the modified results are better than the results of original
Empirical Okumura Hata model.
The plots are made with the help of Matlab code in
the previous chapter to see the results graphically and two
curve are plotted one curve is for the modified Hata model
and other is according to original empirical Okumura Hata
model as shown in the following plots. The relative error is
also calculated for each site to show the comparison of
modified model with original Empirical Hata model.

The Sahuwala BTS plot is shown in the figure 2 the
average relative error of pathloss is 0.0506 and with original
equation of Empirical Okumura Hata Model the error is
0.1562.





Fig. 2. Sahuwala BTS Pathloss Plot

The Sawantkhera BTS plot is shown in the figure 3. The
average relative error of path loss is 0.0282 and with original
equation of Empirical Hata Model the error is 0.1684.


Fig. 3. Sawantkhera BTS Pathloss Plot
The Ottu BTS plot is shown in the figure 4. The average
relative error of path loss is 0.0416 and with original
equation of Empirical Hata Model the error is 0.0919.

Fig. 4. Ottu BTS Pathloss Plot
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO209-5



The Dabwali Rd BTS plot is shown in the figure 5. The
average relative error of path loss is 0.0162 and with original
equation of Okumura Hata Model the error is 0.1972.



Fig. 5. Sawantkhera BTS Pathloss Plot

The Gubana Majra BTS plot is shown in figure 6. The
average relative error of path loss is 0.0827 and with original
equation of Empirical Okumura Hata Model the error is
0.0648.



Fig. 6. Gubana Majra BTS Pathloss Plot


V. Statistical Analysis of Modified Model

The modified model is verified for other sub urban sites
as shown in above section. Statistical analysis such as
relative error and standard deviation is calculated from
the measured log files both for Modified model as well
as Original Empirical Okumura Hata Model .The
average relative error for sites 0.04048 and standard
deviation is 0.00574 for the for Modified Hata Model
and the average relative error for sites 0.12778 and
standard deviation is 0.01222 for original empirical
Okumura Hata Model as shown in table IV.

TABLE IV
STISTICAL ANALYSIS

The actual path loss results are compare with the original
Empirical Hata Model as well as Modified Hata model with
new parameters value as Eq.(10). The results are compared
as shown in the Table IV.

VI. CONCLUSION
An effort to improve The Hata empirical model for sub
urban area is done in this research work. The results obtained
from the modified model are better than the theoretical
empirical Okumura Hata model as shown in the comparison
of results. The calculation of constant initial system design
parameters with value of a and calculation of related to slope
of model curve β
sys
. The average relative error for sites
0.04048 and standard deviation is 0.00574 for Modified
Hata Model and the average relative error for sites 0.12778
and standard deviation is 0.01222 for the for original
empirical Okumura Hata Model.


REFERENCES
[1] Andrea Goldsmith “Wireless Communications”, Cambridge
University Press 2005,pp 1-2.
[2] Micheal, D. Y., "Foundation of Mobile Radio Engineering",
CRC Press Inc., 2000.
[3] M. Hata, “Empirical formula for propagation loss in land mobile
radio services,” IEEE Trans. Veh. Technol., vol. VT-29,, Aug.1980,
pp. 317–325.
[4] John S. Seybold "Introduction to RF propagation", Willy
Puplisher,2005 pp 11-12.
[5] Micheal, D. Y., "Foundation of Mobile Radio Engineering", CRC
Press Inc., 2000.
[6] Jacques, L. and S. Michel, "Radio Wave Propagation Principles and
Techniques", John Wiley & Sons Ltd., 2000.
[7] K.Ayyappan, P. Dananjayan, "Propagation Model For Highway In
Mobile Communication System", UBICC journal,2009.
[8] M. B. Roslee and K. F. Kwan, "Optimization of hata propagation
prediction model in suburban area in malaysia," Progress In
Electromagnetics Research C, Vol. 13, 91-106, 2010.
[9] Ericsson Radio Systems AB, TEMSTM CellPlanner 3.4 User
Guide, 2001.
[10] http://www.wifinerd.com/wifi-calculators.html.
[11] http://reference.mapinfo.com/software/mapinfopro/english/9.5/ Map
Info Professional Supplement.pdf.
SNO NAME OF BTS
RELATIVE ERROR
Modified Hata
Model
Empirical Hata
Model
1 BUPANIA 0.0532 0.1045
2 DABWALI ROAD 0.0162 0.1972
3 GUBANA MAJRA 0.0648 0.0827
4 KABLANA 0.0586 0.0723
5 JANDWALA JATTAN 0.0356 0.1157
6 HABUANA 0.0506 0.1562
7 KALANWANLI 2 0.0053 0.1473
8 OTTU 0.0416 0.0919
9 SWANTKHERA 0.0282 0.1684
10 DOHBA KHURD 0.0507 0.1416
AVERAGE 0.04048 0.12778
VARIANCE 0.000033 0.000149
STANDARD DEVIATION 0.00574 0.01222

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO210-1

1

Abstract— Microwave systems have an enormous impact on
modern society. Applications are diverse, from entertainment via
satellite television, to civil and military radar system. In
particular, the recent trend of multi-frequency bands and multi-
function operations in wireless communication systems along with
the explosion in wireless portable devices are imposing more
stringent requirements such as size reduction tunability or
reconfigurability enhancement and multiband operations for
microwave circuits. This paper describes about the design of
stepped-impedance microwave low pass filter by using Microstrip
layout which works at 2.3 GHz for permittivity 4.4 value with a
substrate thickness 1.6 mm. The stepped-impedance low pass
filter have a pass band ripple 0.1 dB. The development of the
Microstrip filters are simulated by using IE3D simulator
software.


Index Terms— Low pass Filter, Dielectric Constant,
Microstrip filter, Millimeter wave filter.

I. INTRODUCTION
n a microstrip filters that we are trying to achieve are to have
an exact center frequency, good bandwidth and low return
loss level. There are no active devices to add uncertainty to
the fabricated results. The filters are one of the primary and
necessary components of a microwave system.


Manuscript received March 4,2011. This work was supported in part by
the Department of Electronics & Communication Engineering, Krishna
Institute of Engineering & Technology, Ghaziabad.
Navita Singh is Sr. lecturer with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone: 9999227633; e-mail: singh.navita@ rediffmail.com.
Saurabh Dhiman is Student with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone:9953480681;e-mail:dhiman.saurabh28@gmail.com.
Prerna Jain is Student with Department of Electronics & Communication,
Krishna Institute of Engineering & Technology, Ghaziabad, INDIA.
Phone:08860898015 ;e-mail: jainprerna20@gmail.com.
Shrestha is Student with Department of Electronics & Communication,
Krishna Institute of Engineering & Technology, Ghaziabad, INDIA.
Phone:9811374345 ;e-mail: shresthachita89@yahoo.com.
Tanmay Bhardwaj is Student with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone: 9899249500 ; e-mail: tanmaybhardwaj2008@yahoo.com.

Microstrip line is a good candidate for filter design due to its
advantages of low cost, compact size, light weight, planar
structure and easy integration with other components on a
single circuit board. Conventional filter structures like equal
ripple and butterworth low pass filters are requirement of
special fabrication methods. Conventional low frequency
techniques for fabrication does not fit at these frequencies due
to the very high losses associated. Although microstrip is not
the highest performance filter technology, still it is the
preferred choice in many thin-film on ceramic and printed
circuit board applications. RF pre-selector filters, image
rejection filters, local oscillator (LO) filters and intermediate
frequency (IF) filters can all be realized in microstrip. The
recent advance of novel materials and fabrication technologies,
including monolithic microwave integrated circuit (MMIC),
microelectromechanic system (MEMS), micromachining,
high-temperature Superconductor (HTS), and low-temperature
co fired ceramics (LTCC), has simulated the rapid
development of new microstrip and other filters.
In this paper, filter is optimized for high performance and an
efficient. Microstrip technology is used for simplicity and ease
of fabrication. The design and simulation are performed using
3D full wave electromagnetic simulator IE3D. This filter is
widely used today in radar, satellite and terrestrial
communication applications.

II. MICROSTRIP FILTER DESIGN
The design of low pass filter involves two main steps. The first
one is to select an appropriate low pass prototype. The choice
of the type of response, including Pass band ripple and the
number of reactive elements will depend on the required
specifications. The element values of the low pass prototype
filters, which are usually normalized to make a source
impedance go = 1 and a normalized frequency Ωc = 1.0, are
then transformed to the L-C elements for the desired cutoff
frequency and the desired source impedance, which is
normally 50 ohms for microstrip filters. The next main step in
the design of microstrip low pass filters [3] is to find an
appropriate micro strip realization that approximates the
lumped element filter. The element values for the low pass

Design of Stepped-Impedance Microstrip Line
Low Pass Filter for Wireless Communication
Navita Singh, Saurabh Dhiman, Prerna Jain, Shrestha, Tanmay Bhardwaj

I

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO210-2

2
prototype with Chebyshev response at pass band ripple factor
L
AR
= 0.1 dB, characteristic impedance source/load Zo = 50
ohms, are taken from normalized values g
i
i.e. g
1
, g
2
, g
3
, g
4......
,
g
n.
The filter is assumed to be fabricated on a substrate of
dielectric constant
r
and of thickness h mm. for Angular
(normalized) cutoff frequency Ω
c
, using the element
transformation[1].
The filter design steps are as follows:
1.Determine the number of sections from the specification
characteristics for microstrip parameters.
Filter Specifications:

Relative Dielectric Constant, є
r
= 4.4
Cut-off frequency, fc = 2.3 GHz
Height of substrate, h = 1.6 mm
The substrate used –
The loss tangent tanδ = 0.02
Z
o
= 50

Ω

c
= 1

2. Determine the values of the prototype elements to realize
the specifications. Also we have taken the element value for
low pass from table 3.2[1] for n=3.
L
i
= (Z
o
/g
o
) (Ω
c
/2πf
c
) g
i
......................... (1)
C
i
= (g
o
/Z
o
) (Ω
c
/2πf
c
) g
i
...................... (2)


l
L
= λ
gl
/2π Sin
-1

c
L
i
/ Z
OL
)…………… (3)
l
C
= λ
gc
/2π Sin
-1

c
C
i
Zoc).................... (4)
3.To calculate the width of capacitor and inductor we use the
following formula
W/h = 8 exp (A)/exp(2A)-2 …………… (5)
Where A= Zc / 60 {εr+1}^0.5+ εr+1/ εr-1{0.23+0.11/ εr}
................... (6)
Where Zc = η / 2 π √ εre [ln (8h/w+ 0.25 w/h)]
................... (7)

Where η = 120 π ohms is the wave impedance in free space.

4. The effective dielectric constant can be found by the
following formula

ε
re
= (ε
r
+1)/2+(ε
r
-1)/2 [(1+12h/W)
-0.5
]...... (8)

5. Effective wavelength is also found as

λ
ge
=300/(6√ε
ree
) ................................. (9)




TABLE I
DIMENSIONS FOR A STEPPED-IMPEDANCE LOW PASS FILTERS (FOR N=3)

S.No. Dimension Value

1.

Microstrip line width
in mm

WC=7.027,WO=3.059,
WL= 1.707


2. Characteristic impedance
in ohm

ZOC = 20,ZO = 50, ZOL = 60



3. Effective dielectric
constant

(εre)C = 3.5799, (εre)O = 3. 381,
(εre)L = 3.1858




Fig.1. Layout of a 3-pole, stepped- impedance Microstrip low
pass filter on a substrate with εr= 4.4and h = 1.6 mm at 2.3
GHz frequency.

III.SIMULATION RESULTS
In order to verify the validity of the above expressions in
millimeter wave regime, a simulation study was performed
using IE3D. To get the exact response for our purpose, an
optimization was performed using software. The design of the
filter is completed, the layout of the filter is given in figure 1
with all the determined dimensions. Figure 2 shows the EM
simulated performance of the filter.


Fig.2. Full-wave EM simulated performance of the stepped-
impedance low pass filter for n = 3 at 2.3 GHz.


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO210-3

3


IV. RESULTS & ANALYSIS
The Simulated filter as shown in Figure 1 and 2 shows the
geometry & response of low pass filters for n=3.The graph is
plotted by taking gain (dB) on the Y-axis and frequency in
GHz on the X-axis. From the graph it is clear that the cut-off
frequency is found to be 2.3 GHz for stepped-impedance low
pass filter. Hence the stepped-impedance low pass filter is
capable of passing the frequency less than 2.3 GHz & rejects
the frequency after 2.3GHz for the thickness of the substrate
1.6mm and also have the return loss performance up to -38 dB.


V. CONCLUSION
The filters are one of the primary and necessary components of
a microwave system and they are the very essential part of the
microwave system, not only in microwave are they very
important in communication field. Any communication system
cannot be design without filters. Low pass filter must be
included at the transmitting end and the receiving end of the
system to get desired spectrum. In conclusion, the authors
believe that the design can be archived compact filter design.
The measured characteristics of the filter agree with the
theoretical simulations. The conventional geometry of the
stepped-impedance low pass filter, shown in figure 1 and the
simulated response of the conventional geometry shown in
figure 2.The symmetrical approach tends to produce a more
compact filter with less coupling effect in its realization. Its
compact nature minimizes required space for realization and is
suitable for integration within Wireless system. The most
efficient way in order to obtain a filter with maximum size
reduction is by using the microstrip technique in which each
filter’s lumped component is realized as microstrip
transmission line.
ACKNOWLEDGMENT
The authors would like to thank authorities of Krishna
Institute of Engineering & Technology for all the support
provided.

VI. REFERENCES
[1] Jia-Shen G. Hong & M.J. Lancaster, “ Microstrip Filters
for RF/ Microwave Applications” John Wiley & Sons
Inc., 2001.
[2] D.M.Pozar, “Microwave Engineering, “John Wiley,2000.
[3] Jia-Sheng Hong, Lancaster M .J, “ Recent progress in
Planar microwave filters,” IEEE Trans. Antenna
Propagat, Vol. 2, pp. 1134 – 1137, August 1998.
[4] G. Mathaei, L.Young & E.M.T. Jones, “Microwave Filter
Impedance matching networks and coupling structures,”
Artech House, Norwood, MA, 1980.
[5] IEEE Trans.vol.CT-5, January 1964.
[6] IE3D Software Release – 8, Developed by M/S Zeland
Software Inc.
[7] Tae - Yeoul Yun and Kai Chang, “Transaction on
Microwave Theory and Techniques,” IEEE Trans, Vol
49, No.3, March 2001.
[8] M. Makimoto and S.Yamoshita, “Microwave Resonators
and Filters, “IEEE Trans.Wireless communication Vol.2,
August 1986.
[9] H. Ozalki and J. Ishii, “Synthesis of transmission -line
networks and the Design of UHF filter,” IRE Trans. On
circuit theory, vol. CT-2, pp.325- 336; December 1955.
[10] Jen-Tsai Kuo,,“ Parallel Coupled Microstrip Filters With
Suppression of Harmonic Response” In IEEE, Wei-Hsiu
Hsu, and Wei-Ting Huang Vol. 12, No. 10, Oct 2002
383.
[11] Z.M. Hejazi and A. Omar, “Modeling and simulation of
novel ultra narrow band miniature microstrip filters for
mobile and wireless critical applications,” Microw.
Opt. Technol. Lett., vol. 45, no. 1, pp. 35–39, 2005.
[12] Ozaki, H.,and J.Ishii, “Synthesis of a Class of Stripline
Filters,” IRE Trans., Vol. CT-5, 1958, pp. 104 – 109.


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO211-1

1

Abstract—Filters are significant RF and Microwave components.
Lumped element filters are impractical for compact designs of
wireless communications equipment, especially hand-held devices.
Distributed element filter design offers a much smaller area and
profile. With the advent of advanced substrate materials offering
high dielectric constants with low loss, the size reduction with
preserved efficiency is greatly enhanced.
In this paper we discuss end coupled band pass filters at
millimeter wave frequency for same permittivity condition. Both
end coupled band pass filters are designed at 10.5 GHz and 6
GHz. The cost and the fabrication complexity are much less since
it uses the thick-film process. The design and simulation are
performed using 3D full wave electromagnetic simulator IE3D .

Index Terms—Band pass filter, coupled lines, Micro strip filter,
Millimeter wave filters.
I.INTRODUCTION

HE term microwave may be used be describe
electromagnetic waves with frequencies ranging from
300MHz to 300GHz, which correspond to wavelengths (in
free space) from 1meter to 1mm. The EM waves with
frequencies above 30GHz and up to 300GHz are also called
millimeter waves because their wavelengths are in the milli----
meter range (1-10 mm). Above the millimeter wave spectrum
is the infrared, which comprises electromagnetic waves with
wavelengths between 1 μm and 1 mm beyond the infrared
spectrum is the visible optical spectrum, the ultraviolet
spectrum, and x-rays. Below the microwave frequency
spectrum is the radio frequency (RF) spectrum.

Manuscript received March, 01 2011. This work was supported in part by
the Department of Electronics & Communication Engineering, Krishna
Institute of Engineering & Technology, Ghaziabad.

Navita Singh is Sr. lecturer with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone: 9412582436; (e-mail: singh.navita@ rediffmail.com).
Avnish Kumar is student of Electronics & Communication, with Krishna
Institute of Engineering & Technology, Ghaziabad, INDIA,
Phone: 9997866775; (e-mail: avi.k.prajapati@gmail.com).
Manohar Kumar is student of Electronics & Communication, with Krishna
Institute of Engineering & Technology, Ghaziabad, INDIA,
Phone: 9634477323; (e-mail: manoharkumarald@gmail.com).
Ashish Kr. Verma is student of Electronics & Communication, with Krishna
Institute of Engineering & Technology, Ghaziabad, INDIA,


Phone: 9557593123; (e-mail: ashishvermadng@gmail.com).

The frequency boundary between RF and microwaves is
somewhat arbitrary, depending on the particular technologies
developed for the exploitation of that specific frequency range.
Therefore, by extension, the RF/microwave applications can
be referred to as communications, and other that explore the
usage of frequency spectrums, some of these frequency
spectrums are further divided into many frequency bands[1].A
microwave filter is a two-port network used to control the
frequency response at a certain point in a microwave system
by providing transmission at frequencies within the pass band
of the filter and attenuation in the stop band of the filter.The
electromagnetic spectrum is limited and has to be shared for
many RF/microwave applications; here filters play important
roles. The filters are used to separate or combine different
frequencies or to select or confine the RF/microwave signals
within assigned spectral limits [1]. Emerging applications such
as wireless communications continue to challenge
RF/microwave filters with ever more stringent requirements
i.e. higher performance, smaller size, lighter weight, and lower
cost. Depending on the requirements and specifications,
RF/microwave filters may be designed as lumped element or
distributed element circuits; they may be realized in various
transmission line structures, such as waveguide, coaxial line,
and microstrip lines. The recent advance of novel materials
and fabrication technologies, including monolithic microwave
integrated circuit (MMIC), microelectromechanic system
(MEMS), micromachining, high-temperature Superconductor
(HTS), and low-temperature co fired ceramics (LTCC), has
simulated the rapid development of new microstrip and other
filters. In the meantime, advances in computer-aided design
(CAD) tools such as full-wave electromagnetic simulators
have revolutionized the filter design.
Microwave filters specifically band pass filters have found
large number of application in varying fields such as satellite
communication, GSM networks, Bluetooth, remote sensing,
navigation etc. Due to their short wavelength, highly directive
and smaller antennas which are therefore more practical and
enable the use of multiple channels. Filters play important
roles in many RF/microwave applications. They are used to
separate or combine different frequencies. The
electromagnetic spectrum is limited and has to be shared ,
filters are used to select or confine the RF/microwave signals
within assigned spectral limits..

End Coupled Band Pass Filter for Wireless
Communication

Avnish Kumar ,Manohar Kumar, Ashish Kumar Verma, Navita Singh


T
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO211-2

2


In this paper, filters are optimized for high performance and an
efficient and cost effective fabrication methods is proposed.
Microstrip technology is used for simplicity and ease of
fabrication. The design and simulation are performed using 3D
full wave electromagnetic simulator IE3D. This filter is widely
used today in radar, satellite and terrestrial communication
applications.



II.MICROSTRIP FILTER DESIGN
An nth order filter will use N resonant sections of transmission
line With N+ 1 capacitive gap between them. The
discontinuities may be approximated as Series capacitors. The
resonators are approximately λ/2 long at the resonant
frequency ω0.

FIGURE 1. General Configuration of end coupled microstrip
band pass filter

Where each open end microstrip resonator is approximately a
half guide wavelength long at the mid band frequency
0
f of
the band pass filter [1]. The coupling from one resonator to the
other is through the gap between the two adjacent open ends,
and hence is capacitive. The gap can be represented by the
inverters [1]. These J-inverters tends to reflect high impedance
level to the end of each of half wave length resonator, and it
can be shown that this cause the resonator to exhibit a shunt
type resonance . Thus, the filter under consideration operates
like the shunt resonator type of filter whose general design
equations are:

J
01
/Y
0
= √(∏ FBW/2g
0
g
1
) (1)

J
j,j+1
/Y
0
= (∏ FBW/2)1/√ (g
j
g
j+1
) j = 1 to n-1 (2)

J
n,n+1
/Y
0
= √(∏ FBW/2g
n
g
n+1
) (3)

Where g
1
, g
2
, g
3
, g
4......
, g
n .
are the element of a ladder low pass
protype with a normalized cutoff Ωc = 1.0, and FBW is the
frational bandwidth of bandpass filter. The J
j,j+1
are the
characteristic of J-inverters and Y
0
is the characteristics
admittance of the microstrip line[3].
Assuming the capacitive gaps act as perfect, series-
capacitance discontinuities of susceptance B
j,j+1
as

B
J,J+1
/Y
0
= (J
j,j+1
/Y
0
)/(1-( J
j,j+1
/Y
0
)
2
) (4)

Ө
j
= ∏ -1/2[tan
-1
(2 B
J,J+1
/Y
0
) + tan
-1
(2 B
J-1,J
/Y
0
)] radians
(5)
Where the B
J,J+1
and Ө
j
are evaluated at fo. Note that the
second term on the right hand side of second equation
indicates the absorption of the negative electrical length of the
J-inverters associated with the jth half wavelength resonator.

As referring to the equivalent circuit of microstrip gap, the
coupling gaps s
j,j+1
of the microstrip end- coupled resonator
filter can be determined as

C
g

j,j+1
= B
J,J+1

o
(6)



Where ω
o
= 2∏

f
o
is the angular frequency at the midband.

The physical length of resonator are given by

l
j
= (λ
g0
/2∏ ) Ө
j
- ∆ l
j
e1
-

∆ l
j
e2
(7)

where ∆ l
j
e1,e2
are the effective lengths of the shunt
capacitances on the both ends of the resonator j because the
shunt capacitances C
p

j,j+1
are associated with the series
capacitances C
g

j,j+1
as defined in the equivalent circuit of
microstrip gap, they are also determined once C
g

j,j+1
in
equation are solved for the required coupling gaps. The
effective lengths can then be found by [1]

∆ l
j
e1
= (ω
o
C
p

j-1,j
/Y
0
)(λ
g0
/2∏ ) (8)

∆ l
j
e2
= (ω
o
C
p

j,j+1
/Y
0
)(λ
g0
/2∏ ) (9)

III.SIMULATION RESULTS
In order to verify the validity of the above expressions in
millimeter wave regime, a simulation study was performed
using IE3D. To get the exact response for our purpose,an
optimization was performed using software.The dimensions of
the filters are given in the table and the simulated filter
responses are depicted below.


TABLE I

DIMENSIONS OF THE FILTERS FOR n=3 (in mm)

S.No. Type Physical
Length
Wid
th
Gap
1. BPF at
10.5GHz
l
1
=l
3
=6.4274
l
2
=6.4794

W=
3.05

s
01
=s
34
=0.3
s
12
=s
23
=0.84
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO211-3

3
2. End Coupled
Filter at
6GHz
l
1
=l
3
=12.56
l
2
=13.32
W=
1.1
s
01
=s
34
=0.03
s
12
=s
23
=0.601


Fig.2. Layout of end coupled BPF on a substrate with a
relative dielectric constant of 4.4 and a thickness of 1.6mm

Fig.3.Full-wave EM simulated frequency response of the filter
for εr = 4.4 and h =1.6.

Fig.4. Layout of end coupled BPF on a substrate with a
relative dielectric constant of 4.4 and a thickness of 1.27mm

Fig.5.Full-wave EM simulated frequency response of the
filter for ε
r
= 4.4 and h =1.27.

IV. RESULTS & ANALYSIS
The conventional geometry of the chebyshev band pass filter,
shown in figure 1. the simulated response of the conventional
geometry in figure 2 & 4. The attenuation decreases at 10 GHz
frequency shows in figure 3 & 5 for the band edge frequency
of 9.66 GHz and 11.34 GHz(at 6 Ghz the band edge frequency
of 5.85 GHz and 6.15 GHz). The conventional geometry of
the Butterworth band pass filter, shown in figure 2 & 4. The
simulated response of the conventional geometry in figure 3 &
5, the attenuation decreases at 10 GHz frequency shows in
figure. for the band edge frequency of 9.66 GHz and 11.34
GHz, we getting a sharp cutoff at 10.5 GHz & maximally flat
response without any ripples in the pass band with non
attenuated frequency in the desired band. Considerable
improvements in the frequency response can be seen.
V. CONCLUSION
The filters are the very essential part of the microwave system,
not only in microwave but they are also very important in
communication field. Any communication system cannot be
designed without filters. The simulated end coupled band pass
filters achieved an insertion loss of less than 1.0dB and
10.5Ghz filters is designed on 1.6 mm thick substrate for ε
r
=
4.4 with loss tangent 0.0024. And 6Ghz filters is designed on
1.27 mm thick substrate for ε
r
= 4.4 with loss tangent 0.0024.
As frequency of operation increases return loss decreases.
Considerable improvements in the frequency response can be
seen.
ACKNOWLEDGEMENT
The authors would like to thank authorities of Krishna
Institute of Engineering & Technology for all the support
provided.

VI.REFERENCES
[1]. Jia-Shen G. Hong & M.J. Lancaster, “Microstrip Filters
for RF/ Microwave Applications” John Wiley &
SonsInc.,2001.
[2]. D.M.Pozar, “Microwave Engineering,”John Wiley,2000.
[3]. Jia-Sheng Hong; Lancaster M.J, “Recent progress in
planar microwave filters,” IEEE Trans. Antennas Propagat.,
Vol. 2, pp. 1134 – 1137, August 1998.
[4].G. Mathaei, L.Young & E.M.T. Jones, “Microwave Filter
impedance matching networks and coupling structures,”
Artech House, Norwood, MA, 1980.
[5]. IEEE Trans.vol.CT-5, January 1964.
[6].IE3D Software Release – 8, Developed by M/S Zeland
Software Inc
.
[7].Tae-Yeoul Yun; and Kai Chang, “Transaction on
Microwave Theory and Techniques,” IEEE Trans, Vol 49,
No.3, March 2001.

[8].M.Makimoto; and S.Yamoshita, “Microwave Resonators
and Filters, “IEEE Trans.Wireless communication Vol.2,
August 1986.

[9].H.Ozalki and J.Ishii, “Synthesis of transmission -line
networks and the Design of UHF filters,” IRE Trans. On
circuit theory, vol. CT-2, pp.325- 336; December 1955.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO211-4

4
[10]. E.N. Torogow, Senior member, IEEE, and G. E Collins,
“Band-Stop Filters for High-Power Applications”, IEEE
Trans. Vol 40, pp 185-196; September 1965.

[11]. Jen-Tsai Kuo,,“ Parallel Coupled Microstrip Filters With
Suppression of Harmonic Response” In IEEE, Wei-Hsiu Hsu,
and Wei-Ting Huang Vol. 12, No. 10, Oct 2002 383
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-1


SECURITY OF ELECTRONIC MONEY
(RFID CREDIT CARD)
Mr. Pankaj Singh Mr. Sanyam Agarwal Mr. Rohit Sharma
Asst. Professor Asst. Professor Lecturer
SRM University, Modinagar SBBGI Engg. College, Meerut DVSIET, Meerut

Abstract: Recently, payment card issuers in
the United States have begun mass deployment
of radio-frequency enabled payment cards. We
examine a few examples of this new class of
payment card and observe that while the card
issuers have implemented some new security
features, all of them admit practical attacks to a
greater or lesser degree. This paper serves as the
first of two installments describing the threats
and vulnerabilities of RFID-based credit cards
as demonstrated by our laboratory experiments
and also proposed a model to make card more
secure.

Keywords: RFID, credit cards, contactless,
vulnerabilities, proposed model

1. About RFID Credit Cards: An increasing
number of credit cards now contain a tiny
wireless computer known as an RFID chip
(Radio Frequency Identifier) or a contactless
smart card chip. There are reportedly over 20
million RFID credit cards already deployed in
the U.S., and this number is increasing rapidly.
According to Visa ``This has been the fastest
acceptance of new payment technology in the
history of the industry'. RFID credit cards are
growing in popularity because they permit
contactless payment transactions which are fast,
easy, can be more reliable than magstripe
transactions, and require only physical
proximity (rather than physical contact) between
the credit card and the reader. These same
features, however, are also the basis for our
concern about security and privacy
vulnerabilities. Traditional credit cards require
that an entity have visual access or direct
physical contact in order to obtain information
from the card such as the cardholder's name and
the credit card number. By contrast RFID credit
cards make these and other sensitive pieces of
data available using a small radio transponder
that is energized and interrogated by a reader. In
the remainder of this technical report we will
examine what data are conveyed, what kind of
equipment is required to receive such a
transmission, and whether such transmissions
can be initiated by adversarial (not authorized
by credit card companies) readers.

2. Summary of Findings: We examined
representative RFID enabled credit cards from
the major payment associations. We engaged
each of the cards in various transactions with
several different kinds of RFID readers,
including specialized point-of-sale equipment
deployed by major retailers. We determined that
cheap off-the-shelf hardware and software are
sufficient for an adversary with only modest
technical skills to obtain critically sensitive data
from the RF interface of the cards. While it is
possible that some existing RFID credit cards
have mechanisms for protection of sensitive
information, cards from most issuers reveal all
of the following information totally unprotected
by any cryptographic security mechanism:
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-2


1. Cardholder Name
2. Complete credit card number
3. Credit card expiration date
4. Credit card type
5. Information about software version and
supported communications protocols

Cards from one issuer revealed this same
information with the exception that the credit
card number used by the RF interface is a
different number from that encoded onto the
magnetic stripe. Off-the-shelf hardware requires
reasonably close proximity (on the order of
10cm) in order to read these data from an RFID
credit card. This is sufficient to read a credit
card through clothing or a wallet, but the reader
must still be close to the targeted card. The
maximum potential range available seems to be
hotly debated in press accounts of the situation.
Experiments conducted by Royal Dutch Shell of
Canada and reported in indicate a read range of
26 inches, while retailer 7-Eleven insists that the
range is only 2 inches. This longer range is
supported by the academic literature: detailed
instructions on how to build and operate such a
specialized reader are already available to the
public on the World Wide Web.
Additionally the range at which it is possible to
detect the presence of a tag or eavesdrop on a
transaction may be much longer than the range
required for an active skimming attac. This fact
is apparently behind the report from the
National Institute of Standards and Technology
that the protocol used with these kinds of RFID
chips can be read from up to 30 feet away using
specialized reading hardware.

3. Threat Model: RFID-enabled credit cards
include several noteworthy features that open
vulnerabilities not found on traditional
magstripe cards. For instance, RFID cards can
receive external power, communicate
wirelessly, include persistent state that is
modified by readers at each transaction, and
perform computation.
Our security analysis considers a set of
adversaries of various capabilities and goals.
RFID threat models have already been well
explored [6]. Our investigation focuses on
informal scenarios specific to RFID credit cards.

3.1 Attacks and scenarios: The RFID
security and privacy community defines several
conventional adversaries by their methods and
goals. Our informal analysis considers four
basic attacks to privacy and security.
Clandestine bearer tracking violates a tag
bearer’s privacy by exceeding expected
interactions. Clandestine scanning allows an
adversary to energize and read tags from either
close proximity or a distance. The replay and
relay attacks allow a more powerful man-in-
the-middle adversary to impersonate a card
holder. Finally, a cross-contamination attack
allows an adversary to use information from RF
transmissions to attack non-RF media such as
magstripe. This study does consider more
advanced attacks such as side-channel analysis
and physical probing.

Clandestine bearer tracking: In this scenario,
a legitimate merchant exceeds the expected use
of their RFID credit card readers. For example,
a merchant A may want to know whether a
credit card C has been used with any other
vendor since the last transaction between A and
C. Avoine describes examples of how such
attacks on the cardholder’s privacy could be
implemented. Some of the cards we examined
admit this sort of attack by means of a
transaction counter that could be co-opted.

Clandestine scanning: In this attack an
unauthorized and potentially clandestine reader
reads tags from either close proximity or from a
distance. One such scanning attack we call the
“Johnny Carson” attack. In a famous series of
comedic skits, Johnny Carson’s character
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-3

Carnac the Magnificent used his mental powers
to read the contents of sealed envelopes. Since
containers that are visually opaque are not
necessarily RF-opaque, security measures
sufficient for traditional credit cards such as
“security envelopes” must be re-examined in the
context of RFID-enabled cards. The Johnny
Carson attack on RFID credit cards occurs when
an attacker uses access to the physical mail
stream to read RF data from credit cards in
transit to their owners. This attack is particularly
powerful because the adversary gains accessory
knowledge such as cardholder address, and
because physical access to postal mail is
commonly easy in many areas (such as areas
with side-of-the-road mailboxes). Many other
scanning attack scenarios exist, such as the
“bump-and-run” close range skimming of the
contents of a victim’s wallet while standing in a
crowded line, elevator, or subway. Long range
skimming attacks are even more diverse, but the
absolute maximum read ranges for ISO 14443-B
transponders have not yet been clearly
established. Even if the read ranges of RFID-
enabled credit cards are short, their new uses
and form factors will engender new
opportunities for attack. Cards that support
sufficient read range may tempt consumers to
hold their wallets up to readers, rather than to
remove their cards first. In this case, a corrupted
reader for, say, physical access to a parking
garage, could skim customers’ credit-card
information at the same time that they read the
parking pass. Fob-type RFID credit cards are
now available for attachment to key rings,
exposing them to attack when consumers leave
their keys unattended. This behavior is seen
most often in valet-parking situations or in
gymnasiums where it is common for users to
leave their keys together in an unsecured box by
the door. The fact that such cards may not bear
embossed numbers can create a false sense of
security in addition to the fact that consumers
are skilled at protecting their wallets, but as we
have seen, often leave their keys exposed.

Replay and relay: In a replay attack, an
adversary broadcasts an exact replay of the
transponder end of the radio signal recorded
from a past transaction between an RF device
and a reader. Where mechanisms exist to foil
simple replays (such as time stamps, one-time
passwords, and challenge-response
cryptography) a related but more sophisticated
attack can frequently still succeed. This attack,
commonly known as the relay attack, uses a
man in the middle adversary to relay an
ephemeral connection from a legitimate reader
through one or more adversarial devices to a
legitimate tag which may be at a considerable
distance. The distance at which the relay attack
can succeed is limited only by the latency which
will be tolerated by the attacked protocol.
However, many portions of ISO-14443-B are
extremely tolerant to latency.

Cross contamination: The cross contamination
attack occurs when private information such as
cardholder name, number, and expiration date
learned by an adversary in an RF context are
then used by the adversary in a different
context. For example, the adversary could use
this data to create a magstripe card, re-encode
the stripe on an existing card, or use these data
in a card-not-present transaction such as a
telephone or online mail-order purchase.

4. Methodology and Experiments: Our
experiments used two different kinds of
commercial readers that read RFID credit cards
and produce serial output compatible with
standard charge-processing networks according
to ISO 7813. This is the standard that is used for
the magstripe data on a standard credit card.
One of these readers was obtained used from a
company at which it is the standard POS (Point
Of Sale) interface at hundreds or thousands of
locations nationwide. The other reader was
obtained new, and is also of a common type
used at POS locations. We examined
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-4

representative samples of many different RFID
credit cards from each of the three largest
payment associations in the U.S. Our
conclusions are based on observations of this
sample set of approximately 20 credit cards
obtained by five separate researchers from
several different issuing banks and three major
payment associations. All of these cards were
issued within the past year. Given the size and
diversity of our sample set we believe that our
results reflect the current state of deployed
RFID credit cards, however we wish to
emphasize that card issuers continue to
innovate, and as time goes by we hope and
expect that cards will change and additional
security features will be added. Therefore we do
not claim that our findings are exhaustive, in
particular there may already exist cards that use
security mechanisms beyond any we have
observed.

4.1 Eavesdropping experiments: Our
eavesdropping hardware consisted simply of a
tuned 13.56MHz antenna connected to an
oscilloscope. Using this setup we obtained
oscilloscope traces of complete transactions
between various RFID credit cards and our
various commercial readers. The serial output
obtained from the readers during these
transactions was saved for later correlation. The
raw oscilloscope trace was sufficient for
determination of carrier frequency and RFID
protocol due to the characteristic 10% amplitude
modulation associated with the ISO 14443-B
RFID protocol. Simple signal analysis software
based on information in section 3 of the ISO
14443-B specification was used to process
scope traces. This analysis revealed that all of
the data transmitted by the credit card reader
consisted of well-formed ISO 14443-B layer 3
and layer 4 commands. Since this ISO protocol
specifies that a cyclic redundancy check be
transmitted with each command, we were able
to confirm the accuracy (no garbled bits) of our
analysis of each command. BPSK demodulation
of PICC transmissions revealed that all PICC
communications also conform to the ISO-
14443-B protocol. Our hardware and software
are able to capture and demodulate any data
transmitted between RFID credit card readers
and cards that are within a certain distance of
the eaves- dropping antenna. Since the focus of
this work is not on extending read ranges, we
did not try to achieve great range. But we did
experimentally demonstrate that eavesdropping
with our setup is effective through materials
such as cloth, lending credence to the threat of
clandestine eavesdropping, perhaps through
clothing. Examination of data obtained through
these means immediately demonstrated the
efficacy of the simple eavesdropping attack,
since the full cardholder name and card
expiration date were present in clear text in all
transactions. Other data such as credit card
number are discussed in Section 5.

4.2 Skimming experiments: In our first
skimming experiment we took a commercial
RFID credit card reader described above and
determined that when presented with an RFID
credit card it produced serial output in
conformance with the ISO 7813 Track 1 and
Track 2 format. Since this is the exact data that
is normally transmitted by a POS terminal
across a charge processing network,


Fig. 1. Our assembled PICCAL credit card
emulator
We note that this extremely simple skimming
attack is clearly sufficient for perpetration of
certain kinds of financial fraud. In order to
perform more flexible active attacks against
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-5

RFID credit cards, we built a device capable of
impersonating our commercial RFID credit card
readers. We required only the ability to send
arbitrary bytes according to ISO 14443 layers 2
and 3, and we discovered that the Texas
Instruments s4100 Multi-Frequency RFID
reader possessed all of the hardware capabilities
that we required. This hardware together with
applications of our own design allows us to
rapidly challenge cards at a rate far exceeding
that observed on any commercial hardware. In
addition, this same hardware combined with
custom amplifiers provides the basis for some
noteworthy read range extension experiments.
Using libraries of our own design, we wrote a
program which simply sends the exact bytes that
we captured from the commercial readers in our
eavesdropping experiments. Eavesdropping on
transactions between our credit card reader
emulator and real RFID credit cards
demonstrated that all of the RFID credit cards
we tested responded to our emulator exactly as
they respond to a commercial RFID credit card
reader. This strongly suggests that cards operate
in a “promiscuous mode” interacting with any
reader, with no cryptographic or other secure
mechanisms in use to authenticate an authorized
RFID reader to a credit card.

4.3 Replay experiments: An RFID credit card
belongs to a class of RFID devices known in the
ISO standard as a “proximity integrated circuit
card” (PICC). Since the primary difference
between our device and a traditional PICC is
that ours uses actively powered logic circuitry
as opposed to the passively powered (antenna
powered) RFID credit card, we have named our
device a PICCAL (Proximity Integrated Circuit
Card with Active Logic). Our PICCAL is a
microprocessor controlled device capable of
sending arbitrary bytes over the ISO 14443-B
transport layer (layer 2). For ease of prototyping
and flexibility of experimentation we chose the
gumstix single board computer as the controller
for our device. This computer is approximately
the size and shape of its namesake stick of
chewing gum, and incorporates an ARM
PXA255 microprocessor . Four General Purpose
Input/Output pins (GPIOs) of the
microprocessor are used to control simple radio
circuitry of our own design. The small size and
low power requirements of this single board
computer contribute the feasibility of
clandestine use of a PICCAL. Our analog front
end is essentially a simple AM radio consisting
of three integrated circuits, and a few capacitors
and resistors. The integrated circuits are: a
comparator used to demodulate AM commands
from the PCD, a counter/divider to divide the
input carrier into subcarrier and baud rate
clocks, and an XOR to allow the microprocessor
to accomplish phase shifting for the 14443-B
layers 2 and 3 specified binary phase shift
keying.



Table 1. A summary of the four types of cards
and susceptibility to various attacks.

We programmed our PICCAL to expect the
RFID credit card reader commands that we
captured using our eavesdropping setup
described in Section 4.1, and to transmit replies
captured from real RFID credit cards during a
skimming attack performed with the reader
emulator described in Section 4.2. The output
from our commercial RFID credit card readers
is identical in the case where the reader is
presented with a real versus PICCAL emulated
credit card. Since the data thus output is the
same data we would expect to be transmitted
over the charge processing network, we cannot
think of a scenario in which the charge
processing network could distinguish a real card
from a PICCAL unless additional elements are
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-6

present that we have not been able to observe in
the laboratory. But as noted above, many pieces
of data go into an overall transaction approval
decision including sophisticated risk-based
fraud detection mechanisms on the backend. For
this reason, a valuable future research direction
would include field tests in which PICCAL
initiated transactions are tested with complete
purchases from real merchants.

5 Analysis and Results: To protect the
identity of our cards, we label the cards A, B
and C based on semantic equivalence classes
determined by observing behavior between
cards and readers. Table 1 summarizes the
vulnerabilities of four classes of cards.

5.1 RFID credit card protocols: In a
traditional card-present transaction, data is read
from the magnetic stripe of a credit card by a
POS terminal. The format of this data is
specified by ISO 7813.
In this section we shall explore some of the
RFID credit card protocols that are in current
deployment. We shall examine some of the
conclusions that can be reached through
examination of the ISO 7813 data output by the
serial port of RFID credit card readers when
presented with different types of credit cards.
Where pertinent we shall consider in correlation
with this serial output the raw RF data from the
same transactions as captured by our
eavesdropping apparatus. In keeping with a
philosophy of ethical attacks research, we have
redacted several pieces of information from the
following subsections in part due to a desire to
prevent criminal misuse of our findings.
Cardholder name and card number have been
concealed. Additionally we have obscured the
number of digits in the card number in order to
obscure which observations correlate with the
products of specific payment associations and
issuing banks.

Card A protocol: When presented (RF
transaction) with any sample of a card from
issuer A, our reader outputs serial data identical
to the data contained on the magstripe of the
same credit card. This finding was confirmed by
comparison with output obtained with
presentation of the magstripe rather than RF.
When presented with the same card, the output
is always the same: there is no evidence of a
counter, one-time-password, or any other
mechanism for prevention of replay attacks.
Figure 2 shows a sample of this serial output,
which includes all the usual components of an
ISO 7813 magstripe. The first line represents
Track 1. The start sentinel B is followed by the
primary account number. Following the field-
separator character, the cardholder name
appears, followed by another field-separator and
an “additional data” field. This field includes
not only the card expiration date (in this case
06/2009), but also a long string of digits. The
meaning of these additional digits is not clear,
but since this field is static for card type A, it
cannot be used to prevent a replay or cross-
contamination attack. The second line represents
standard Track 2 data, which is largely similar
to the Track 1 data. Track 2 does not contain the
cardholder name, and contains less room for
proprietary information.

{Bxxxxxx6531xxxxxx^DOE/JANE^09061010
00000000000000000000000000858000000}-T1
{xxxxxx6531xxxxxx=09061010000085800000}
-T2

Fig. 2. Serial output from a commercial reader
after an RF transaction with a card from issuer
A

Card B protocol: When presented with cards
from issuer B, our commercial readers output
data similar to that of the card A experiments,
with a few important differences. In the sample
card B output shown in Figure 3 we note the
presence of a counter, determined to be such
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-7

because of monotonic incrementation with
successive transactions. Additionally we
observe three digits which change with each
transaction in no pattern that we have identified.
Because of the relatively high entropy of these
three digits, we consider it likely that they are
the output of some cryptographic algorithm
which takes the transaction counter as an input.
If this is the case, then the algorithm must also
take a card-specific value like a cryptographic
key as an input since we observe that different
cards with the same counter value produce
different codes. We speculate that these data
may serve as a stand-in for the traditional CVC.

{Bxxxxxx1079xxxxxx^DOE/JANE^09011011
00000000000100000000000}—T1
{xxxxxx1079xxxxxx=09011011000001600221}
—T2
{Bxxxxxx1079xxxxxx^DOE/JANE^09011011
00000000000100000000000}—T1
{xxxxxx1079xxxxxx=09011011000007400231}
-T2

Fig. 3. Sample of reader serial output after RF
transaction with a card from issuer B. In this
sample we see a three digit code (shown in bold
italic font), and a four digit counter (shown
underlined).

Card C protocol: Card C’s protocol differs
from Card B’s in a few crucial details:

1. Its unique transaction codes are eight digits
instead of three
2. its transaction counter, now located in the
Cardholder Name field, displays only three
Digits instead of four
3. Rather than sending the embossed card
number over the air, it uses a fixed pseudonym
Shown in Figure 4 are transactions 017 and 018
from an issuer C RFID credit card. These
transactions correspond to codes 10691958 and
40146036, seen both at the end of Track 1 and
in different order at the end of Track 2. Note
that the counter value in Track 2 is followed by
a 0 instead of a 1, perhaps an indication to the
back end processing network of a different
algorithm.

{Bxxxxxx2892xxxxxx^DOE/JANE
017^1001101010691958}-T1
{xxxxxx2892xxxxxx=100110101069195801700
}-T2

{Bxxxxxx2892xxxxxx^DOE/JANE
018^1001101040146036}-T1
{xxxxxx2892xxxxxx=100110104014603601800
}-T2
Fig. 4. Sample output from an issuer C card
differs from output of an issuer B. Transaction
codes are shown in bold italic font, transaction
counter is shown underlined.
Proposed Model: - We have studied all three
protocols and we observe that card c protocol is
the most secure compare to other. But in some
condition card c protocol will also be fail. What
are this condition and what improvement we
have to do in this card to make it more secure?
Failure condition: Because here CODE is equal
to sum of key and counter value. That producing
the same code for all cards at same counters
value.

Example- If we have three cards and all cards
have a same transaction and have a same
counter value than the entire three cards will
getting the same code value. This is also a
sophisticated problem

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-8


(PREVIOUS MODEL)

In this model:

(CODE=KEY+COUNTER VALUE)

-----But in proposed model our code is the sum
of key and CVC value of card.

{Bxxxxxx2892xxxxxx^DOE/JANE
017^1001101010691958}—T1
{xxxxxx2892xxxxxx=10011010CODE0170}—
T2

CODE=KEY+X

X=CVC VALUE OF CARD + COUNTER
VALUE

This proposed model provides the different
codes to all those cards, which operate at the
same counter value. Because each cards have a
different CVC value.
(PROPOSED MODEL)

8 Comparative Analysis :

Factors Card A
Protocol
Card B
Protocol
Card C
Protocol
Proposed
Protocol
Security Less Moderate Moderate Most
Secure
Code
Reliability
Less Moderate Moderate Most
Reliable

9 Graphical Analysis :




10 Conclusion: Despite the millions of RFID-
enabled payment cards already in circulation,
and the large investment required for their
manufacture, personalization, and distribution,
all the cards we examined are susceptible to
privacy invasion and relay attacks. Some cards
may be skimmed once and replayed at will,
while others pose a modest additional
synchronization burden to the attacker. After
reverse engineering the secret protocols between
RFID-enabled credit cards and readers, we were
able to build a device to mount several advanced
replay attacks in laboratory conditions. While
absolute security and privacy in a contactless-
card form factor is difficult to achieve, we hope
that next-generation RFID-enabled payment
systems will protect against the vulnerabilities
that our study identifies.

References

1. Adida, B., Bond, M., Clulow, J., Lin, A.,
Murdoch, S., Anderson, R., Rivest, R.: Phish
and chips: Traditional and new recipes for
attacking EMV. Technical report, University of
Cambridge Computer Laboratory (2006)
http://www.cl.cam.ac.uk/˜mkb23/research/Phish
and-Chips.pdf.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO212-9

2. Anonymous: Chip and spin (2006)
http://www.chipandspin.co.uk/problems.html
Last Viewed October 11, 2006.
3. Associated Press: Wave the card for instant
credit (2003)Wired News
http://tinyurl.com/yc45ll Last Viewed October
16, 2006.
4. Averkamp, J.: ITS Michigan: Wireless
technology and telecommunications (2006)
http://www.itsmichigan.org/ppt/AM2005/Joe.pp
t Last Viewed October 11, 2006.
5. Avoine, G.: Privacy issues in RFID banknote
protection schemes. In: International
Conference on Smart Card Research and
Advanced Applications – CARDIS. (2004) 33–
48
6. Avoine, G.: Adversary model for radio
frequency identification. Technical Report
LASECREPORT- 2005-001, EPFL (2005)
7. Bono, S., Green, M., Stubblefield, A., Juels,
A., Rubin, A., Szydlo, M.: Security analysis of a
cryptographically-enabled RFID device. In:
14th USENIX Security Symposium. (2005)
8. Bray, H.: Credit cards with radio tags speed
purchases but track customers, too (2006)
Boston Globe, August 14, 2006
http://tinyurl.com/lmjt4.
9. CardTechnology: Paypass subway trial starts
in New York (2006) http://tinyurl.com/uya3k.
10. Carey, D.: NFC turns phone into a wallet.
EE Times (2006) http://tinyurl.com/yyxk28 Last
Viewed October 8, 2006.
11. DIFRWear: Faraday-Caged Apparel. (2006)
www.difrwear.com Last Viewed October 6,
2006.
12. Dougherty, G.: Real-time fraud detection
(2000) MIT Applied Security Reading Group
http://pdos.csail.mit.edu/asrg/02-28-2000.html
http://pdos.csail.mit.edu/asrg/02 28- 2000.doc.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO213-1

HIGH TEMPERATURE MEMS BASED TRANSMITTER
FOR WIRELESS SENSOR AND
COMMUNICATION NETWORK
Ruchi Gupta
1
(Asstt. Professor. JPIET (Meerut))

Neha Mittal
2
(Asstt. Professor. JPIET (Meerut))

(ruchi.siet@gmail.com)


ABSTRACT—The paper presents a thought
experiment as to the feasibility of using large scale
wireless sensor networks as a vehicle for high level
scientific investigation. A sensor payload is
proposed, which includes means of seismic,
chemical, temperature and visual exploration. The
power and communications systems are also
discussed, based on the needs of a mission profile
which provides no special ‘base station’ nodes on
the planet’s surface, requiring each sensor package
to be capable of information extraction, in-network
collaboration and communication with an orbiting
satellite. And different sensor technologies are also
discussed. A high-temperature, low-power silicon-
tunnel-diode-based oscillator transmitter with an
on-board optical power converter is proposed for
harsh environment MEMS sensing and wireless
data transmission applications. The prototype
sensing and transmitting module employs MEMS
silicon capacitive pressure sensor performing
pressure to frequency conversion and a coil loop
serving as the inductor of the LC tank resonator
and also as a transmitting antenna. A GaAs
photodiode converts an incoming laser beam to
electrical energy to power the prototype. The system
achieves a telemetry performance up to 250 oC over
a distance of 1.5 meters with a transmitter power
consumption of 60 μW.

Keywords:- Distributed computation, information
extraction, intelligent sensing, MEMS sensors,
wireless sensor networks.

I. INTRODUCTION
The research agenda in Wireless Intelligent
Sensor Networks (WSN) has often been
promoted in the specialist literature through
scenarios and case studies. High
temperature, low-power wireless sensor
communication network with on-board
power supply is critical for industrial,
automotive, and aerospace sensing and data
telemetry applications. Typical temperature
for these applications ranges from 200
o
C to
600
o
C. Conventional microelectronics BJT
and CMOS technologies suffer from severe
performance degradation and failure due to
excessive leakage currents for temperatures
above 150
o
C. Silicon on insulator (SOI) and
silicon carbide (SIC) device technologies are
promising for increased operating
temperatures of 250
o
C and 600
o
C
respectively. It is highly desirable to develop
a stand-alone high temperature sensing and
data telemetry system which, therefore, can
be powered by an on-board energy supply,
thus eliminating the need for feed-through
wires. A number of wireless sensing and
communication architectures have been
developed for room-temperature
applications such as biomedical implants. In
these applications, MEMS sensors such as
pressure sensors, strain gauge transducers,
etc. are interfaced with active electronics
that convert the sensing information to
frequency or to a voltage which is further
digitized for wireless transmission. Active
RF transmitters can achieve telemetry
distances adequate for high temperature
applications, but consume significant power
dissipation compared to the overall system
power dissipation, a critical bottleneck for
high temperature operations where power
source is highly limited. In this paper, we
present a stand-alone low-power wireless
sensor communication module with an on-
board optical-based power generator,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO213-2

achieving a telemetry distance of 1.5 meters
under operating temperatures up to 250
o
C.

II. SENSOR TECHNOLOGY
The initial design decisions of sensor
technology to use a single type of node, it is
required that every node carry a complete
sensor package. For a probe with the
envisaged dimensions, the overall sensor
package will need to be designed in an
integrated way, allowing for reuse of as
many subsystems as possible. As each
sensor type is discussed below:

A. Acceleration sensor:
Accelerometers are an important part of the
instrument package on the Daisy, reflecting
their ubiquity in the overall field of sensing.
High performance capacitive pick-off
accelerometers can be integrated „side-by-
side‟ with adequate ancillary electronics
using a CMOS process.

B. Atmospheric sensor
The chemical sensing technology for
atmosphere and soil composition detection
require a further design choice. Electronic
„noses‟ are an established MEMS
technology, but generally need to be tailored
for the detection of a specific chemical or
class of chemicals. One type of general
purpose sensor is a gas chromatograph.

C. Pressure sensing
Pressure sensors have been one of the
earliest MEMS applications, and well
established designs are available using either
capacitive or piezo-electronic pick off
techniques

D. Soil chemical sensor
The difficulty in chemical sensing soil in an
environment without water is conveying the
constituent chemicals to the analyzer. A
simple solution would be to use the
transmission laser to vaporize the sample.
The 7mW of power available should be
sufficient.

E. Magnetometer
Magnetometers typically require a
magnetically active element which moves
according to the prevalent magnetic field,
the movement being sensed using any of the
established MEMS pick-off techniques. The
magnetically active element can be provided
using either a permanent magnet using
electromagnetism, which consumes current.

F. Thermal sensing
Temperature sensors can use thermocouples
or semiconductor junctions, the latter having
the advantage of being readily available as
components within standard semiconductor
technologies. Thermal sensing has also been
used for wind speed and direction sensing.

G. Image sensor
This subsystem is straight forward, being a
silicon CCD or CMOS image sensor, as
used in modern digital cameras. The
resources that need to be integrated are an
estimated 500 pixel array for the pick-off
and image sensors for imaging the
surroundings. This requires an 8500 pixel
array, on a 5mm square chip, easily within
current technological capabilities.

III. HIGH TEMPERATUR SENSING
AND TELEMETRYSYSTEM

Fig1 presents the high temperature prototype
architecture. The system consists of a
silicon-tunnel-diode based LC-tuned
oscillator transmitter employing a MEMS
capacitive pressure sensor with an on-board
loop inductor also functioning as a
transmission antenna, and a GaAs
photodiode which converts an incoming
laser to a DC power at high temperatures.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO213-3

The negative resistance characteristic
exhibited by the diode under a proper bias
condition can compensate the tank loss, thus
developing an oscillation. The DC bias
voltage typically required to properly bias
the device ranges from about 100 mV to 200
mV. This low DC bias voltage can be
readily obtained from a GaAs photodiode.

Fig1. Wireless Transmitter Architecture

The low DC bias voltage also significantly
minimizes the power dissipation, which is a
key advantage over other conventional
electronic oscillator implementations
requiring a supply voltage of a few volts.
The simplicity of the tunnel diode structure
results in a reduced leakage current
compared to conventional electronic active
devices at elevated temperatures, thus
enabling a reliable system operation at high
temperatures. The oscillator output
frequency is determined by the LC tank
resonance. The MEMS capacitive pressure
sensor converts the environment pressure
information to a capacitance change
resulting in the oscillator output frequency
variation. This pressure to frequency
modulation scheme is attractive for
achieving a reliable data transmission
compared to other amplitude modulation
techniques. The GaAs photodiode converts
the power of an incoming laser beam into a
DC power supply to power the system, thus
eliminating any feed-through wire. The
optical powering method is capable of
achieving a much larger coupling distance
than conventional RF to DC power
conversion schemes. Fig2 shows the I-V
characteristics measured at various
temperatures for a silicon tunnel diode, used
for the prototype implementation.

Fig2. Tunnel Diode I-V Characteristics

The tunnel diode exhibits negative
resistance characteristics up to 250
o
C.
Voltage and current bias levels of approx.
120 mV and 500 μA correspond to a power
consumption of 60 μW.

Fig3. Photodiode I-V Characteristics

Fig4 shows the I-V characteristics of a 3 mm
x 4 mm GaAs photodiode measured over a
temperature range from 25
o
C to 290
o
C with
an 8 mW laser beam illuminating the
surface. It also shows a typical device
characteristic response between the
capacitance value and applied pressure,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO213-4

exhibiting a linear characteristic beyond the
touch point pressure.

Fig4. MEMS Pressure Sensor Characteristic
Fig5 shows a photo of a fabricated pressure
sensor. The device exhibits a touch point
pressure of 10 psi and capacitance values
ranging from 15 pF at 2 psi to 25 pF at 32
psi (absolute pressures). The device has an
estimated series resistance of 25 Ω, which
limits the oscillator operating frequency.

Fig5. MEMS Pressure Sensor Photo

A 23 MHz oscillation frequency is chosen to
ensure the LC tank resistance can be
compensated by the tunnel diode over the
temperature range. A 5-turn, 1 μH spiral
inductor with a peripheral dimension of 3.5
cm x 3.5 cm is employed in the prototype
system to achieve the desired frequency.
Low loss capacitive sensors and spiral
inductors are also critical for minimizing
bias current required for tunnel diodes,
crucial for low power applications. Fig7
shows a photo of the prototype wireless
MEMS sensing and data telemetry system.

Fig6. Prototype Board Photo

IV. EXPERIMENT RESULTS
The sensor telemetry system is positioned
inside a pressure testing chamber with
temperature elevated and controlled through
resistive heating tape. A spectrum analyzer
is used as an external receiver with a tuned
receiving loop antenna connected to the
input port through a buffer. The oscillator
operates around 23 MHz under 1 atm at 250
o
C and can be varied over 1.5 MHz through
pressure increase from 2 psi to 32 psi
limited by the tunnel diode parasitic
capacitance, as shown in Fig 7.


Fig7. Oscillator Frequency vs. Pressure

The oscillator exhibits an output frequency
shift of approximately 500 kHz over the
temperature range due to components
temperature dependent characteristics
variation and tunnel diode bias point shift.
Fig8 presents the received power versus
telemetry distance under 1 atm measured at
25
o
C,200
o
C,&250
o
C,respectively, indicating
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO213-5

that the spectrum analyzer can receive an
incoming signal with an SNR of at least 10
dB over telemetry distances of 1.5 m. Fig9
shows the corresponding received power
spectrum at 1.5-meter telemetry distance
from the prototype oscillator operating at
250
o
C. An extended communication range is
expected through using a more sensitive
receiver.
A high-temperature frequency variation over
time has been observed in the current
prototype. The prototype exhibits an initial
frequency decline of approximately 150 kHz
over 30 minutes, then a random frequency
variation of 20 kHz, thus limiting the system
resolution.

Fig8. Received Power vs. Distance

Fig9. Received Power Spectrum

V. CONCLUSION
Silicon-tunnel-diode-based-oscillator
transmitter with an on-board optical power
converter is attractive for stand alone, low
power high-temperature MEMS sensing and
data telemetry applications. The prototype
wireless sensing and communication module
achieves high-temperature operations up to
250
o
C over a telemetry distance of 1.5
meters with a transmitter power
consumption of 60 μW. Together with
MEMS actuation technologies, the sensing
technologies allow for new adaptive
systems, providing more efficient
aerodynamics and hydrodynamics. MEMS
sensors have already proven to be a potent
technology in terms of size and cost
reduction of everyday sensors. When
integrated with processing capability they
can handle a range of further applications, in
particular being more readily deployed and
adapted to a specific application problem.

REFERENCES

[1] A.J.Auberton-Herve, “SOI: Materials to
Systems,” Int. Electron Devices Meeting, 1996,
pp.3-10.

[2] D. M. Brown, et al., “High temperature silicon
carbide planar IC technology and first monolithic
SiC operation amplifier IC,” in Trans. 2nd Int.
High-Temp. Elec. Conf., 1994, pp. XI-17 to XI-22.

[3] M. Suster, et al., “Micro-Power Wireless
Transmitter for High-Temperature MEMS Sensing
and Communication Applications,” The 15th IEEE
Int.l Conference on MEMS, pp. 641-644, January,
2002.

[4] S. Chatzandroulis, et.al., “A miniature Pressure
System with a Capacitive Sensor and a Passive
Telemetry Link for Use in Implantable
Applications,” IEEE J. MEMS, V. 9, No. 1, pp. 18-
23, Mar., 2000.

[5] K. Stangel, et. al, “A Programmable
Intraocular CMOS Pressure Sensor System
Implant,” IEEE Journal of Solid-State Circuits,
Vol. 36, No. 7, pp. 1094-1100, July, 2001.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO213-6

[6] M. A. Fonseca, et. al., “Wireless
Micromachined Ceramic Pressure Sensor for
High-Temperature Applications,” IEEE Journal of
Micromechanical Systems, Vol. 11, No. 4, August
2002.

[7] W. H. Ko and Q. Wang, “Touch Mode
Capacitive Pressure Sensors,” Sensors and
Actuators 75 (1999), pp. 242-251.

[8] K. S. J. Pister, J. M. Kahn and B. E. Boser,
"Smart Dust: Wireless Networks of Millimeter-
Scale Sensor Nodes", Highlight Article in
Electronics Research Laboratory Research
Summary (1999).

[9] E. Gaura, Smart, Intelligent and Cogent MEMS
based sensors, in E. Gaura, R. Newman, Smart
MEMS and Sensor Systems, Imperial College
Press, 2006.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO214-1

An Overview on Load Balancing Routing
Protocols for Mobile Ad hoc Networks
First Author: Bhawana Bhardwaj, Second Author: Sanjay K. Dhurandher


Abstract- Mobile ad hoc networks (MANETs) are
collections of mobile nodes that can dynamically form
temporary networks on the fly without the need for
pre-existing network infrastructure or centralized
administration. Due to unorganized connectivity and
dynamic topology, routing in MANET becomes a
challenging task. Moreover, constraints like lower
capacity of wireless links, error-prone wireless
channels, limited battery capacity of each mobile node
etc., degrade the performance of MANETs routing
protocols. Heavily-loaded nodes may cause congestion
and large delays or even deplete their energy quickly.
In this article, we discuss the various load metrics and
summarizes the principles behind several existing load
balanced routing protocols.

I. INTRODUCTION
A mobile ad hoc network (MANET) is a kind of
wireless ad hoc network, and is a self-
configuring network of mobile routers (and
associated hosts) connected by wireless links –
the union of which form an arbitrary topology.
The routers are free to move randomly and
organize themselves arbitrarily; thus, the
network's wireless topology may change rapidly
and unpredictably. They have features of fast
deployment and able to communicate without
link breakage while on the move. They can
easily communicate anywhere at any time. Over
the years, numerous routing protocols have been
developed for ad hoc mobile networks. Routing
protocols in MANETs are classified into three
categories: table driven proactive, on-demand
reactive source initiated and hybrid routing [3].
Table Driven Proactive routing protocols
maintain consistent, up-to-date routing
information in each node by propagating route
updates throughout the whole network.
Although a route to every other node is always
available, such protocols incur signaling traffic
and power consumption overhead. On the other
hand, on-demand reactive routing protocols do
not maintain routing information at every node.
They create routes only when desired by the
source. It has been shown in [4] and stated in [5
that reactive routing protocols can perform better
than proactive protocols in MANETs.




There are various ad hoc routing protocols which
use route hop count as the routing metric. In
shortest path routing, nodes on shortest path get
more heavily loaded than others because they are
chosen frequently as routing path. Having a
heavy load, constraints like lower capacity of
wireless links, error-prone wireless channels,
limited battery capacity of each mobile node etc.,
degrade the performance of MANETs routing
protocols. Heavily-loaded nodes may cause
congestion and large delays or even deplete their
energy quickly. Therefore, routing protocols that
can evenly distribute the traffic among mobile
nodes and hence can improve the performance of
MANETs are needed.
This article is organized as follows. The
next section categorizes various types of load
metrics and load balanced ad hoc routing
protocols. Afterthat, we will compare various
load metrics. We then discuss load balancing
techniques used by existing load balanced
routing protocols. This is followed by a
comparison of these various protocols in the
following section. The final section concludes
the article.

II. CATEGORIZATION
Over the years, various load balanced ad hoc
routing protocols have been proposed. Most of
the approaches are on-demand-based protocols;
that is, they follow the load balancing strategies
with route discovery [2]. A route with the least
load among multiple possible routes from source
to destination is generally chosen. As shown in
Fig. 1, these routing protocols can generally be
categorized into three types (based on their load
balancing techniques):
1 Delay-based: Where load balancing is
achieved by attempting to avoid nodes
with high link delay. Examples of delay
based load balanced routing protocols is
Load-Aware on-Demand Routing
(LAOR) [6].
2 Traffic-based: Where load balancing is
achieved by evenly distributing traffic
load among network nodes. An example
protocol using this approach are
Associativity Based Routing (ABR) [7],
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO214-2

Load Balanced Ad Hoc Routing
(LBAR) [8], and Traffic-Size Aware
(TSA) scheme [4].
3 Hybrid-based: Where load balancing is
achieved by combining the features of
traffic- and delay-based techniques.
Examples are Content Sensitive Load
Aware Routing (CSLAR)and Zone
Routing Protocol (ZRP) [1].

This is the classification of protocols based on
their load balancing techniques, we also have to
consider the load metrics used by all these
protocols.

The term load metric basically reflects how busy
a node is engaged in receiving and forwarding
packets over the wireless media. It also refers to
processing, memory, bandwidth, and power load
on the node.



As shown in Fig. 2, different load balanced ad
hoc routing protocols use different load metrics:
• Active path: It refers to the number of active
routing paths supported by a node. Generally, the
higher the number of active routing paths, the
busier the node since it is responsible for
forwarding data packets from an upstream node
to a downstream node.
• Traffic size: This refers to the traffic load
present at a node and its associated neighbors
(measured in bytes).
• Packets in interface queue: This refers to the
total number of packets buffered at both the
incoming and outgoing wireless interfaces.
• Channel access probability: This refers to the
likelihood of successful access to the wireless
media. It is also related to the degree of channel
contention with neighboring nodes.
• Node delay: This refers to the delays incurred
for packet queuing, processing, and successful
transmission.


III. LOAD BALANCE BASED ADHOC
ROUTING PROTOCOLS

In this section we discuss load balanced based ad
hoc routing protocols. Specifically, we focus
how these load balancing routing protocols help
to balances the heavily loaded nodes in the
network. As shown in Fig. 1, most of these
protocols are traffic-based (ABR, LBAR, TSA),
while LAOR is delay-based. CSLAR and LARA
are considered hybrid-based, exhibiting features
of both traffic- and delay-based protocols.

A. DELAY BASED PROTOCOLS
LAOR is an extended form of ad hoc on-demand
distance vector (AODV) routing. It has two
phases: route discovery and route maintenance.
In LAOR, load balancing is achieved by
minimizing the estimated total route delay and
route hop count. A node initiates route discovery
by sending a route request (RREQ) when it does
not have a valid route to the destination.
Intermediate nodes that receive the RREQ
message update the total delay to RREQ and the
routing table. The destination node then sends a
reply message (RREP) when it receives the first
RREQ. If duplicate RREQs received at the
destination node have smaller total delay and hop
count than earlier ones, it sends a RREP message
to the source node to change the chosen route
immediately. When the source node receives the
RREP, it initiates data packet transmission. This
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO214-3

approach reduces delay due to the efforts made
in selecting the best route.
LAOR implements congestion monitoring
during the route discovery process. A node is
considered congested if the total delay of a node
is greater than the upstream node or the queue
length is more than 80 percent of its buffer size.
When a node becomes congested, it discards the
RREQ message. This approach reduces the
routing overhead and distributes the load evenly
in the network.

B. TRAFFIC BASED PROTOCOLS
ABR uses the property of “associativity” to
decide on which route to choose. In this
algorithm, routes are discovered by broadcasting
a broadcast query request packet; with the
assistance of these packets, the destination
becomes aware of all possible routes between
itself and the source. Based on these available
routes, a path is selected using the associativity
property of these routes.
The ABR algorithm maintains a “degree of
associativity” by using a mechanism called
associativity ticks. According to this, each node
in the network maintains a tick value for each of
its neighbors. Every periodic link layer HELLO
message increases the tick value by one each
time it is received from a neighbor. Once the tick
value reaches a specified threshold value, it
means that the route is stable. If the neighbor
goes out of the range, then the tick value is reset
to zero. Hence a tick level above the threshold
value is an indicator of a rather stable association
between these two nodes. Once a destination has
received the broadcast query packets, it has to
decide which path to select by checking the tick-
associativity of the nodes. The route with the
highest degree of associativity is selected since it
is considered the most stable of the available
routes.
LBAR’s load metric is similar to ABR, which is
based on active path activity. LBAR has two
phases: route discovery and route maintenance.
The route discovery process is initiated
whenever a source node needs to communicate
with another node for which it does not have a
route. It broadcasts the setup message to its
neighbors. A node receives a setup message then
forwards it after updating the nodal activity and
traffic interference information. Such
information is then used to calculate the route
load [2]. The destination node then selects a
route with the least load and sends an ACK
message back to the source node. During route
maintenance, when a route is broken, the
upstream node propagates an error message to
the destination. The destination node then selects
an alternative least loaded route stored in its
cache to send data. Using this technique, traffic
over the network tends to be distributed in the
long term.

B. HYBRID BASED PROTOCOLS
CSLAR is an extension of Dynamic Source
Routing (DSR), which also uses route discovery
and route maintenance.
CSLAR utilizes the contention information
obtained from the 802.11 distributed
coordination function (DCF) to take into account
channel contention from neighboring nodes.
Load balancing in CSLAR is also performed
during route discovery. When a node has packets
to send, a route request packet is flooded
throughout the network. Route request is also
forwarded by intermediate nodes after they have
appended their load values to the request
message. However, intermediate nodes are not
allowed to send route replies back to the source
even if they have routes to the destination in their
route caches. The destination node then decides
to reply to route requests based on the
comprehensive route load value. The destination
node replies to the first request or to the request
that has a smaller route load than previous
requests. Once the source node receives the
destination node’s reply, it then utilizes the
selected route to transmit data. If one or more
links in the selected route is broken, the source
node is notified by a route error packet, resulting
in the source updating its route cache and
reinitiating a new route discovery to discover a
new least loaded route.
ZRP is a protocol used under hybrid category for
ad hoc mobile routing protocols. It incorporates
the merits of on-demand and proactive routing
protocols. ZRP is similar to a cluster with the
exception that every node acts as a cluster head
and a member of other clusters. The routing zone
comprises a few mobile ad hoc nodes within one,
two or more hops away where the central node is
formed. Since ZRP uses both reactive and
proactive schemes, it exhibits better
performance. However, since hierarchical
routing is used, the path to a destination may be
suboptimal. Since each node has higher level
topological information, memory requirement is
greater.
LARA is another hybrid load balanced routing
protocol. LARA requires each node to maintain a
record of the latest traffic queue estimation of its
neighbors. The traffic queue is defined as the
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO214-4

average value of the interface queue length
measured over a period of time. Traffic density,
on the other hand, refers to the sum of traffic
queues at a node plus the traffic queues of all the
node’s neighbors.. LARA selects the best route
based on checking the traffic type (i.e., TCP or
non-TCP traffic). For TCP traffic, route selection
is based first on route hop count followed by
traffic cost. For non-TCP traffic, the route with
the minimum traffic cost is selected.

III. COMPARISONS OF LOAD BALANCED
AD HOC ROUTING PROTOCOLS
There is a comparison of several load balanced
routing protocols. The key differences among the
routing protocols lie in the method of load
estimation and the decision on route selection.

A. LEAST LOADED ROUTE DECISION
Protocols such as ABR, LBAR, and LARA use
destination-based decision load balancing. Once
the destination node receives a route request, it
waits for some time to select the route with the
least loaded path. When the source node receives
the reply message, it proceeds to send data. TSA,
however, uses source-based decision load
balancing. In TSA multiple reply messages are
sent from destination node to source node. After
waiting for a predefined interval, the source node
selects the least loaded path.

B. ROUTE REQUEST FLOODING METHOD
Most of the load balanced routing protocols are
on-demand-based protocols in which a source
node discovers routes by first flooding a route
request message. In the duplicate route request at
destination node (DQD) approach, the
destination node is allowed to receive duplicate
route requests. The destination node then decides
to reply to all/ some/one route request. Currently,
only the LAOR protocol uses this approach for
route discovery. Meanwhile, protocols such as
CSLAR, TSA, and LAOR use the duplicate
route reply at source node (DPS) approach,
where the source node is allowed to receive
duplicate route reply messages from the
destination node.

C. ROUTE DISCOVERY DELAY
The delay incurred while waiting for multiple
route request or reply messages is an important
factor to be considered in the design of ad hoc
routing protocols. To avoid the issue of large or
least waiting time, both CSLAR and LAOR
protocols allow the destination node to send a
reply whenever it
receives the first route request. Comparing route
discovery delay, on-demand load balanced ad
hoc routing protocols will incur higher delay
since nodes need additional waiting time to
discover possible routes to destinations.
To find the least loaded path among multiple
possible routes, ABR, LBAR, and LARA
protocols allow the destination node to wait for
some time (Δt) before replying with the best
route to the source node. Meanwhile, TSA
allows the source node to wait for a time interval
( Δt) to select the best route to the destination
node. Therefore, the route discovery delay is: 2T
+ Δt, where T is delay incurred for the
destination node to receive the first route request
message. LAOR and CSLAR protocols,
however, incur a shorter delay since the
destination node will reply to the first route
request received (Table 1), so the delay incurred
is 2T.

D. ROUTING OVERHEAD
LBAR, LARA, and TSA use control messages to
exchange load information among neighbors,
which is then used to calculate the load metric.
On the contrary, LAOR and ABR allow nodes to
compute load information (by counting the
number of active paths) without incurring
additional overhead.

E. COMMUNICATION COMPLEXITY
Communication complexity refers to the number
of messages needed to complete a protocol
operation. During the initiation phase, ABR,
LBAR, and LARA require O(N + y), where N is
the number of nodes in the network and y is the
total number of nodes forming the directed path
the REPLY packet transits. In TSA and CSLAR,
to obtain a route from source to destination, the
destination node must reply to all route requests.
The communication complexity of TSA and
CSLAR is similar: O(2N). The complexity of
LAOR is O(N2) due to the allowance eof
intermediate nodes to receive duplicate route
request messages.
When mobility causes route failures (i.e., post-
failure), the communication complexities of
ABR and LBAR are the same, O(x + y), where x
is number of nodes affected by a topological
change. Meanwhile, the communication
complexities of LARA, CSLAR, TSA, and
LOAR are O(N + y), O(2N), O(2N), and O(N2),
respectively, since these routing protocols do not
employ localized route discovery. When
mobility causes route failures (i.e., post
failure), the time complexity of ABR and
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO214-5

LBAR is O(l + z), where l is the diameter of
the affected network segment. Since these
routing protocols use localized route
discovery, time complexity therefore
depends on the diameter of the affected
network segment. Meanwhile, CSLAR,
TSA, LARA, and LAOR have the same time
complexity as in route initiation since they
reinitiate a new route discovery when route
failures occur. Thus, the time complexities
of CSLAR, TSA, LARA, and LAOR are
O(2d), O(d + z), O(d + z), and O(2d),
respectively.

F. TIME COMPLEXITY
Time complexity refers to the number of steps
needed to perform a protocol operation. During
route initiation, ABR, LBAR, TSA, and LARA
incur O(d + z), where d is the diameter of the
network, and z is the diameter of the directed
path the REPLY packet transits. However, for
CSLAR and LAOR, the complexity is O(2d)

IV. CONCLUSION
In this paper we discuss the various types of
load metrics and the principles behind some
existing load balanced ad hoc routing protocols:
ABR, LBAR, LARA, CSLAR, TSA, and LAOR.
While offering better representation of actual
load, LARA and CSLAR incur higher
complexity in capturing load information.
Comparing the operations of routing protocols,
only LBAR and ABR perform load balancing
during route maintenance. Other protocols
balance load only during route discovery.





REFERENCES
[1] Shafinaz Buruhanudeen, Mohamed Othman,
Mazliza Othman, Borhanuddin Mohd Ali
“Existing MANET Routing Protocols and
Metrics used Towards the Efficiency and
Reliability- An Overview“.
[2] Chai Keong Toh, University of Hong Kong
Anh-Ngoc Le and You-Ze Cho, KyungPook
National University “Load Balanced Routing
Protocols for Ad Hoc Mobile Wireless Networks
“.
[3] E. M. Royer and C.-K. Toh, “A Review of
Current Routing Protocols for Ad-Hoc Mobile
Wireless Networks,”
[4] A. H. Altalhi and G. Richard III, “Load-
Balanced Routing through Virtual Paths: Highly
Adaptive and Efficient Routing Scheme for Ad
Hoc Wireless Networks”.
[5] P. Johansson et al., “Scenario-based
Performance Analysis of Routing Protocols for
Mobile Ad Hoc Networks,”
[6] J.-H. Song, V. Wong, and V. Leung, “Load-
Aware Ondemand Routing (LAOR) Protocol for
Mobile Ad Hoc Networks,”
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO214-6

[7] C-K. Toh, “Associativity-Based Routing for
Ad Hoc Mobile Networks,”
[8] H. Hassanein and A. Zhou, “Routing with
Load Balancing in Wireless Ad hoc Networks,”








CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26 27 2011

COMO215-1



Localization of Wireless Sensor Network
using Geographical routing
Saurabh Dixit
1
and Arun Kumar Singh
2

1.Member IEEE, IETE
Department of Electronics and Communication
Northern India Engineering College, Lucknow
saurabh2911@ieee.org
2. Member IEEE, IETE
Department of Electronics and Communication
Bansal Institute of Engineering andTechnology , Lucknow
aksingh_uptu@rediffmail.com

ABSTRACT:
Wireless Sensor Networks have
evolved as a key technology in
sensing and detecting physical
phenomenon such as acoustics,
humidity, temperature, imaging and
seismic activity. They are often
deployed in hostile environments to
collect information in cooperative
and collaborative fashion to achieve
greater objective. Wireless Sensor
Networks have innumerable
applications in all walks of human life
such as military, disaster monitoring,
home security, environmental
monitoring, commercial, medical and
nuclear installation, etc., However,
much of the information gathered by
sensor nodes are of utility by
ascertaining the position of the
nodes which are providing those
information. The location information
can be utilized for routing of the data
via the sensor nodes to sink.
Geographical routing implies location
aided routing in the sense that the
location of the nodes are
predetermined as they are assigned
virtual coordinates which provide
information about the location of
sensor nodes. Geographic routing
also addresses the issue of
scalability and dynamic change in
the network topology.
Because of the vagaries and the
size of the network, the localization
error can severely hinder the
performance of geographic routing, it
is imperative to collectively evaluate
the localization and geographic
routing to understand the tradeoffs
among different self localization
schemes for location based routing
in WSNs. Geographical routing is an
excellent means to enable self
localization.
In this paper, we propose a
method to incorporate Geographical
routing for the purpose of
Localization. The existing methods of
Greedy Perimeter Stateless Routing
(GPSR) and Greedy Other Adaptive
Face Routing(GOAFR) are assessed
and a scheme is proposed to blend
the Geographic routing with
localization of wireless sensor
nodes.
Keywords:
Localization, Geographic routing,
Greedy Perimeter Stateless Routing,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26 27 2011

COMO215-2

Greedy Other Adaptive Face
Routing, Wireless Sensor Networks.

I. Introduction:-
A Wireless sensor
network(WSN) consist of a large
number of sensors which are tiny,
low cost, and low-power devices that
collect environmental data which is
then communicated through radio or
optical means to infrastructure
processing nodes. WSNs may
consist of up to thousands of nodes,
which can be deployed in very high
density, in homes, highways,
buildings, cities, and infrastructure
for monitoring and controlling
purposes. Applications may range
from detecting and monitoring
occurrence of natural disasters and
homeland security, to military
surveillance. The application of
wireless sensor networks is for
instrumenting the world. They can
provide monitoring and warning of
natural disasters and the effects
there from including floods, winds,
and other natural phenomenon. They
can be used for conducting military
surveillance, for battle damage
assessment, for nuclear, biological
and chemical attack detection and
reconnaissance. Environmental
applications include forest fire
detection, bio-complexity mapping of
the environment, flood detection,
precision agriculture. Health
applications include telemonitoring of
human physiological data, tracking
and monitoring of doctors and
patients inside a hospital, and drug
administration inside hospitals.
Phone applications can include
home automation, smart
environment control, and home
security.
The wireless sensor networks
are driven by technology
convergence including
microelectromechanical sensors,
microelectronics, signal processing
and communications devices. They
are enabled by the algorithms that
run under software that execute the
network protocols, the software for
the applications conformance
specifically, and the power
management technology for
operational endurance. Thus
wireless sensor networks integrate
sensing, computing and
communication.
The key technical challenges in
wireless sensor networking include
efficient networking methods. It is
imperative to be able to implement
adhoc networking of any number of
devices. It is important to be able to
support mobile or fixed location
devices depending upon the nature
of the systems applications. The
wireless sensor networks should be
self configuring so as to adjust the
function of the network in case
additional sensors of different kinds
are to be added to the array. The
collaborative signal and information
processing within the network is
used to detect, classify and track
events and patterns of events
occurring in the geographic area that
they are deployed in.
The routing can be both proactive
as well as reactive. In proactive
routing a router is maintained at
every node in the network. In
reactive routing, routes are
maintained to only those nodes
which are needed for the current
activity, so only information of
interest above the threshold levels
are actually accessed. Reactive
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26 27 2011

COMO215-3

routing does have an expensive cost
since flooding is involved requiring
the activation of many nodes.
However this approach is good for
low and medium traffic networks and
sensor networks typically have very
low data rates. In Geographic
routing, location aided routing for
mobile adhoc networks is employed
for being able to meet the demands
of scalability and efficiency in
dynamic networks.

Localization of wireless sensor
nodes are of vital importance as they
convey the node location which is
conveying the information.
Geographical routing algorithms are
an attractive alternative to traditional
adhoc routing algorithms for wireless
networks because they scale better;
the routing state maintained per
node is dependent only on local
network density and not network
size. Since the satisfactory
performance of geographic routing is
dependent on reliable low cost self
localization schemes, hence a
evaluation of their combined
performance needs to be done. In
this paper we have tried to blend the
two isolated topics of geographical
routing and localization so as to aid
the important issue of localization.
The remainder of this paper is
organized as follows: Section II
provides a review of related work. in
the field of geographic routing and
localization. Section III presents
strategies to augment the
performance of geographical routing
along with the self localization
schemes. Section IV provides
conclusions and some directions for
future works.

2. Review of Related Works:

Localization approaches typically
rely in some form of communication
between reference points with known
positions and the receiver node that
needs to be localized. Another
approach to localize sensor nodes is
to use the Routing protocols to
generate virtual coordinates. Hence
instead of using the traditional adhoc
routing algorithms, which assume
that both the node and the sink are
stationary; Geographic routing
algorithms can be alternatively used
to generate virtual coordinates. The
geographic routing algorithms scale
better for wireless networks: the
routing state maintained per node is
dependent only on local network
density and not network size. Since
geographic location devices like
GPS, Bat and Cricket are not cost
effective for ubiquitous deployment
on large scale network; virtual
coordinates can be effectively
employed when location information
is not available
Geographic Routing is a kind of
reactive routing which employs
single hop routing so as to select a
neighboring node within the radio
range of a given node. Geographic
routing protocols use location
information of the neighboring nodes
for making routing decision for a
specific destination node. Based
upon the knowledge of source and
destination, it measures the possible
advance that could be made by
using individual neighbor nodes and
the node providing maximum
advance is chosen as the next hop
for routing. This approach known as
greedy forwarding approach can
however fail occasionally when there
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26 27 2011

COMO215-4

are no neighbor node that can
advance further than the current
node and hence face routing
(perimeter routing) is to be used to
route under these conditions. The
packet returns to Greedy forwarding
as soon as it finds that greedy
routing is available again. This is the
concept of Greedy Perimeter
Stateless Routing (GPSR), which is
the most popular geographic routing
protocols. GOAFR( greedy other
adaptive face routing) is an improved
geographic routing that combines
greedy forwarding with face routing
and achieves worst-case optimality
of the path length by using limited
elliptic regions for face routing and
recursively increasing the ellipse size
until reaching a close-to-optimal
path.

3. Geographical Routing for
Localization:
For generating virtual coordinates in
the case of geographical routing,
Leong et al. had earlier proposed the
GSpring[5] family of coordinate
assignment algorithms for adhoc
wireless networks. Starting from a
set of initial coordinates derived from
a set of elected perimeter nodes,
Greedy Embedded Spring
Coordinates(GSpring) detects
possible dead ends and uses a
modified spring relaxation algorithm
to incrementally adjust virtual
coordinates to increase the convexity
of voids in the virtual routing
topology. This reduces the possibility
that packets will end up in dead ends
during greedy forwarding. Rao et al.
had proposed the NoGeo family of
coordinate assignment algorithms for
adhoc wireless networks[13].In the
most general version of their
algorithm for systems where nodes
have no location information, they
designate two nodes as beacon
nodes. Next nodes determine if they
are perimeter nodes from a heuristic
based on their hop count from the
beacons. Once the perimeter nodes
are determined, O(p
2
) messages are
exchanged, where p is the number of
perimeter nodes, and the perimeter
nodes use an error-minimization
algorithm to compute their
coordinates. Finally, the perimeter
nodes are projected onto an
imaginary circle and nodes
determine their virtual coordinates
using a relaxation algorithm that
works by averaging the coordinates
of neighboring nodes.
However the NoGeo system
assumes the system is static once
the perimeter nodes are determined.
Such a topology is bad for
geographic routing since there is a
mismatch between the geometric
and routing distance.
The novel approach to
Localization of Wireless Sensor
nodes would be to make use of
Geographical routing where virtual
coordinates are identified for efficient
routing. Hence there is a need to
incorporate a technology which can
generate virtual coordinates that
produces usable coordinates quickly
and improves the routing
performance of existing geographic
routing algorithms. For the purpose
of generating virtual coordinates the
GSpring method is a good option.
The primary focus of the work will be
to develop an algorithm which can
assign greedy virtual coordinates to
the sensor nodes.


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26 27 2011

COMO215-5

4. Conclusion
This paper addressed the issue of
Localization of nodes using
Geographical routing. It can be
achieved as Geographical routing
assumes that nodes are aware of
their location. However the degree of
accuracy becomes an issue
especially for dynamic networks.
Hence an analysis is done to
compare the GPSR and GAOFR
routing while for the generation of
virtual coordinates in greedy
forwardind GSpring is
recommended.

References:
[1] Masoomeh Rudafshani and
Suprakash Dutta “Localization in
Wireless Sensor Networks” IPSN’07,
April 2007, Cambridge, MA, USA
[2] Zenon Chaczko, Ryszard
Klempous, Jan Nikodem, Michal
Nikodem, “Methods of Sensor
Localization in Wireless Sensor
networks” Proceeding of the 14
th

Annual IEEE International
Conference Workshop of the
Engineering of Computer Based
System, 2007
[3] Stefano Basagni, Michele Nati
and Chiara Petrioli, “Demonstrating
the resilience of Geographic routing
to localization errors” Mobile Adhoc
and Sensor Systems, 2007.
MASS2007. IEEE Internatonal
Conference, Oct. 2007
[4] B.Karp and H.TKung “GPSR:
Greedy Perimeter stateless routing
for Wireless networks” Proceedings
ACM MOBICOM, July 2000
[5] Ben Leong, Barbara Liskov and
Robert Morris “Greedy virtual
coordinates for geographic routing”
Network protocol, ICNP 2007, IEEE
International Conference, Oct 2007
[6] Nirupama Bulusu, John
Heidemann and Deborah Estrin
“GPS- less Low Cost Outdoor
Localization for very Small Devices”
IEEE personal Communication,
October, 2000
[7] Battelli, M.; Basagni, S.,
“Localization for Wireless Sensor
networks: protocols and
perspectives” Electrical and
Computer Engineering, 2007.
CCECE 2007. Canadian
Conference, April 2007
[8] Huey –Ing Liu ; Chi- Chuan Liu
“An Intelligent agent routing over
wireless sensor networks” Mobile
Adhoc and Sensor System (MA) ,
2006, IEEE International Conference
on Oct 2006
[9] Sungwon Yang; Jiyoung Yi;
Hojung Cha “ HCRL: A Hop-Count-
Ratio based Localization in Wireless
Sensor Networks” Sensor, Mesh and
Ad Hoc Communications and
Networks, 2007. SECON '07. 4th
Annual IEEE Communications
Society Conference, June 2007
[10] Pirzada, A.A.; McDonald, C.
“Trusted Greedy Perimeter Stateless
Routing”
Networks, 2007. ICON 2007. 15th
IEEE International Conference,Nov.
2007
[11] Rahman, M.Z.; Kleeman, L.;
“Self-Localization Schemes for
Geographic Routing in Wireless
Sensor Networks” Vehicular
Technology Conference, 2008. VTC
Spring 2008. IEEE, May 2008
[12] Witt, M.; Turau, V. “The Impact
of Location Errors on Geographic
Routing in Sensor Networks”
Wireless and Mobile
Communications, 2006. ICWMC '06.
International Conference, July 2006
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26 27 2011

COMO215-6

[13] A.Rao, C.H. Papadimitriou, S .
Shenker and I. Stoica “ Geographic
routing without location information”
Proceedings of MOBICOM 2003,
San Diego, September,2003
[14] Ulema, M. “ Wireless sensor
networks: architectures, protocols,
and management”
Network Operations and
Management Symposium, 2004.
NOMS 2004. IEEE/IFIP
Volume 1, April 2004
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO216-1


TIMESTAMP BASED LOAD
BALANCING IN AODV IN MOBILE
AD HOC NETWORKS

First Author:Bhawana Bhardwaj, Second Author:Sanjay K. Dhurandher



Abstract:
Mobile ad hoc networks (MANET) are
infrastructure-less networks, dynamically formed
by an independent system of mobile nodes that are
connected via wireless links. Multipath routing
allows the establishment of multiple paths between
a source and a destination. It distributes traffic
among multiple paths instead of routing all the
traffic along a single path. In this paper, we
propose a new routing protocol named TAODV i.e
Timestamp based load balancing Ad hoc On
Demand Distance Vector Routing protocol, that
uses all discovered paths simultaneously for
transmitting data. By this approach data packets
are balanced over discovered paths on the basis of
timestamp set in the network.

Keywords– Ad hoc, multipath, timestamp, Load
Balancing

I. INTRODUCTION
Mobile ad hoc networks (MANETs) are
collections of mobile nodes that can dynamically
form temporary networks on the fly without the
need for pre-existing network infrastructure or
centralized administration. Each node can
function both as a network host for transmitting
and receiving data and as a network router for
routing packets to the other nodes. Ad hoc
networks have numerous practical applications
such as Sensor networks, military applications,
Rescue operations, Conferences etc.
Due to unorganized connectivity and
topology, routing in MANET becomes a
challenging task. Moreover, constraints like
lower capacity of wireless links, error-prone
wireless channels, limited battery capacity of
each mobile node etc., degrade the performance
of MANET routing protocols in such networks,









nodes are typically distinguished by their limited
power, processing, and memory resources as
well as high degree of mobility. They have
desirable features such as fast deployment and
the ability to communicate while on the move.
MANETs can be categorized into Proactive and
Reactive routing [1]. Proactive routing protocols
maintain consistent, up-to-date routing
information in each node by propagating route
updates throughout the whole network. Although
a route to every other node is always available,
such protocols incur signaling traffic and power
consumption overhead. On the other hand,
Reactive routing protocols do not maintain
routing information at every node. They create
routes only when desired by the source. It has
been shown in [2] and stated in [3] that Reactive
routing protocols can perform better than Pro-
active protocols in MANETs. Most of the
existing Reactive (or on-demand) routing
protocols (for example, Dynamic Source Routing
(DSR) and Ad hoc On demand Distance Vector
(AODV) build and rely on single path for each
data session. So route recovery process is
needed after each route failure, which causes to
lose transmitted data packets, in such protocols.
Multipath routing allows the establishment of
multiple paths between a single source and
single destination node. It is typically proposed
in order to increase the reliability of data
transmission (i.e., fault tolerance) or to provide
load balancing [5, 6, 9 and 10].
Some nodes that may lie on multiple routes
spend most of their energy in forwarding of
packets and deplete their energy quickly.
Consequently they leave the network early.
It may lead to undesirable effects such as longer
delays, lower packet delivery fraction and high
routing overhead.
Our goal here is to develop an on-demand
multipath distance vector protocol as an
extension to a well-studied single path routing
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO216-2

protocol known as Ad hoc On-demand Distance
Vector (AODV). We refer to the new protocol as
Timestamp baes Load Balancing Ad hoc On-
demand Distance Vector (TAODV) protocol.
Primary design goal behind TAODV is to
provide multiple routes on the basis of setting
timestamp art the source node on which the
source sends data simultaneously over them.
In AODV source sends data through shortest
path, thus only a few nodes are involved for
transmitting data and their remaining energy
reduce dramatically, but in TAODV since
multiple routes are used, data packets are
distributed across all nodes on timestamp based
discovered routes.
The further proceeding of this paper is organized
as follows. Section II describes AODV Routing
Protocol. Sections III describe our new proposed
algorithm for searching multiple routes on
timestamp bases and load balancing data over
discovered routes. Section IV follows with the
simulation results and concluding remarks are
made in section V.

II. AD Hoc ON DEMAND DISTANCE
VECTOR ROUTING
The aim of AODV [4, 7 and 8] is to reduce the
number of broadcast messages sent throughout
the network by discovering routes on-demand
instead of keeping complete up-to-date route
information.
A source node seeking to send a data packet to a
destination node checks its route table to see if it
has a valid route to the destination node. If a
route exists, it simply forwards the packets to the
next hop along the way to the destination. On the
other hand, if there is no route in the table, the
source node begins a route discovery process. It
broadcasts a route request (RREQ) packet to its
immediate neighbors, and those nodes broadcast
further to their neighbors until the request
reaches either an intermediate node with a route
to the destination or the destination node itself.
This route request packet contains the IP address
of the source node, current sequence number, the
IP address of the destination node, and the
sequence number known last. Figure 1 denotes
the forward and reverse path formation in the
AODV protocol. An intermediate node can reply
to the route request packet only if they have a
destination sequence number that is greater than
or equal to the number contained in the route
request packet header. When the intermediate
Fig. 1 a) Reverse path Formation
b) Forward path Formation
nodes forward route request packets to their
neighbors, they record in their route tables the
address of the neighbor from which the first copy
of the packet has come from. This recorded
information is later used to construct the reverse
path for the route reply (RREP) packet. If the
same RREQ packets arrive later on, they are
discarded. When the route reply packet arrives
from the destination or the intermediate node, the
nodes forward it along the established reverse
path and store the forward route entry in their
route table by the use of symmetric links. Route
maintenance is required if either the source or
the intermediate node moves away. If a source
node becomes unreachable, it simply reinitiates
the route discovery process. If an intermediate
node moves, it sends a link failure notification
message to each of its upstream neighbors to
ensure the deletion of that particular part of the
route. Once the message reaches to source node,
it then reinitiates the route discovery process.

III. TIMESTAMP BASED LOAD
BALANCING AD HOC ON-DEMAND
DISTANCE VECTOR ROUTING (TAODV)
The aim of AODV is to reduce the number of
broadcast messages sent throughout the network
by discovering routes on-demand instead of
keeping complete up-to-date route information.
While in TAODV we proposed a new solution to
search a route in Ad Hoc network by minimizing
the searching time and improved packet delivery.
TAODV is an on-demand routing protocol that
consists of three main phases:
1) Path Discovery Process:
When a node has data to send and no route
information is known. Then it initiates path
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO216-3

discovery by sending route request packet
(RREQ) to its neighbors. RREQ packet identifies
the host, referred to as the target of the route
discovery having address of the original initiator
of the request and the target of the request. Each
route request packet contains a route record, in
which is accumulated a record of hops taken by
the route request packet. Each RREQ packet also
contains a unique request id, set by the initiator.
If a node receives a RREQ with a new request id.
Having less hop count of the RREQ then
appends its address to the Route Record of
RREQ and increases the hop count of RREQ and
rebroadcast it.
To prevent flooding network with too many
RREQs, nodes only rebroadcast it .If the hop
count of received RREQs is less (or equal to)
than hop count stored in Route Record.
Rebroadcasting of the RREQs is continued until
they reach to the destination. By using this
method for propagation the RREQs, many
RREQs from different routes will be received to
the destination. Destination node will send back
a RREP with total traveled time from which
RREQ was received. When node receives RREP
and if it is not a source node then maintain the
timestamp of neighbor nodes and forward the
RREP with traveling time on the reverse path.
Also all the information about the node is stored
in Routing Table. When source receives RREP
then it updates the routing table based on the
node sending RREP. Source also updates the
timestamp of its neighbors too.
For example, in fig 2, S sends RREQ and A, B
and C receive this RREQ with no. of Hop sets to
1. According to our proposed algorithm when
node B rebroadcasts RREQ and C receives it,
since the hop count of received RREQ is greater
than the hop count of first RREQ, this RREQ
will be dropped. By using this method 7 RREQs
from different paths will be received by the
destination.


Fig. 2 Propagation RREQs
When a destination receives RREQs, reverses the
route in the route record from the received
RREQs with total traveled time, and uses this
route to send back the route reply (RREP) packet
to the source. As the RREPs travel back to the
source each node along the path sets up a
forward pointer to the node from which the
RREP came, NextHop, and updates its timeout
information for route entries to the source. When
a node receives multiple RREPs from a node, it
increments the number of route reply, Count
Reply, that received from this node in its route
table field which means how many routes from
this next hop to the destination are exist. This
process is continued until the RREPs reach the
source. These two phases create multiple routes
from source to the destination. As shown in fig.
2, C has three next hops for D as a destination,
also S receives 7 RREPs from next hops and
since C sends four RREPs to S the count Reply
of C is sets to 4.

Dest. Addr

Next Hop CountReply
D C 4
D B 2
D A 1
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO216-4

Fig. 3 Route Table of S and D
2) Sending Data
When the source receives RREPs, it can transmit
data packets through the discovered routes. Our
protocol uses hop-by-hop method and less
execution time for forwarding data. Each node
that receives data packets sends them to the next
hops according to their Count Reply values and
execution time from source. Each next hop that
has greater Count Reply and less execution time
receive more data. This process causes that all of
the discovered routes is used and data packets
distributed across all of the paths simultaneously.

For example, out of 7 data packets, S sends 4 of
data packets to C that means these packets are
distributed to the four different routes later, also
2 of data packets transmitted to B and 1 of them
to A. Each intermediate node that receives data
packets sends them to the next hops according to
their Count Reply and less execution time in
their Route Table.

// S is the source node; D is the destination node
// RT = Routing Table
S wants to communicate with D if RT of S
contains a route to D
S establishes communication with D
else
S creates a RREQ packet and broadcasts it
to its neighbors
// RREQ contains the destination
Address(DestAddr),
// Sequence Number (Seq) and Broadcast
ID (BID)
for all nodes N receiving RREQ
if (RREQ was previously processed)
discard duplicate RREQ
end if
if (N is D)
Send the Time stamp back node from which
RREQ was received
broadcast RREQ
end if
end for
while (node N receives RREP) and (N != S)
Maintain a time stamp of neighbor nodes
forward RREP with Time of traveling T
on the reverse path



Fig. 4 Algorithm of TAODV
3) Route Maintenance
If a route is not used for some period of time, a
node cannot be sure whether the route is still
valid; consequently, the node removes the route
from its routing table. If data is flowing and a
link break is detected, a Route Error (RERR)
packet is sent to the source of the data in a hop-
by-hop fashion. As the RERR propagates
towards the source, each intermediate destination
is broken. When Count Reply of each next hop
in Route Table reaches to 0 this next hop is
deleted from route table. If no entry for a
destination exists in Route Table of source, it
invalidates the route and reinitiates route
discovery process if necessary. For example in
fig 3, if the link between I and J breaks I sends a
RERR to the C, when C receives this packet it
changes the Count Reply of next hop I in Route
Table to 1 and forwards this packet to S, also
when S receives this packet changes the Count
Reply of C to 3.


Dest. Addr NextHop CountReply
D I 2
D E 1
D F 1
store information about the node sending
RREP in the RT
end for
S receives RREP
S updates its RT based on the node sending the
RREP
S receives and Updates its neighbors Time stamp T
S establishes communication with D
end if
send back a RREP packet to the node sending the
RREQ

else
if (N has a route to D with SeqId >= RREQ.Seq)
send back a RREP packet with total travel time T
else
record the node from which RREQ was received
broadcast RREQ
end if
end for
while (node N receives RREP) and (N!= S)
Maintain a time stamp of neighbor nodes
forward RREP with Time of traveling T
on the reverse path
store information about the node
sending RREP in the RT
end for
S receives RREP
S updates its RT based on the node sending the
RREP

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO216-5

IV. PERFORMANCE EVALUATION
1) Simulation Environment:
The simulation code was implemented within the
Global Mobile Simulation (GloMoSim) library
[11]. In the simulation, we modeled a network of
30 mobile hosts placed randomly within a 1000
X 1000m2 area. Radio propagation range for
each node was 100 meters and channel capacity
was 2Mbit/sec. Each simulation runs for 100
seconds of simulation time. The IEEE 802.11
Distributed Coordination Function was used as
the medium access control protocol.

2) Performance Metrics
We evaluate the performance of TAODV, we
simulated and compared the schemes:
AODV
TAODV.
We evaluated these schemes as a function of
mobility speed of nodes and the traffic of
sessions. The number of sessions was set to 1
and mobility speed of nodes was varied from 0
to 100 m/s with pause-time that is equal to 5
seconds. To evaluate the performance of our
protocol, each source sent data as a rate of
l00Kbit/s to 600Kbit/s.

We evaluated the following metrics for each
session:
Packet Delivery Ratio: The number of
data packets received by destinations
over the number of data packets sent by
the source.
Routing Overhead. The number of all
packets (data and control packets)
transmitted divided by the number of
data packet delivered to the destination.
Average of End-To-End Delay: The
end-to-end delay of a packet is defined
as the time a packet takes to travel from
the source to the destination.

3) Performance Results
a) Packet Delivery Ratio
The packet delivery ratios as a function of
mobility speed is shown in figure 5. In figure 5
data rate is set to 300Kbit/s. We can observe that
as speed increases because of links break the
packet delivery ratios decrease in AODV and
TAODV protocols. Since in TAODV, all
discovered routes are used simultaneously for
transmitting data, large amount of data packets
will receive to the destination in compare of
AODV that use only one route for transmitting
data.











Fig. 5 Packet Delivery Ratio v/s Mobility

b) Number of total packets transmitted per data
packet delivered Figures 6 shows the routing
overhead as a function of mobility speed. In
figure 6 data rate is set to 300Kbit/s. As shown
in figure 6 as speed increases, because of link
break and route reconstruction the routing
overhead in AODV and TAODV are increased.
Since TAODV transmits many RREQs and
RREPs in compare to AODV thus its routing
overhead is higher than AODV.


Fig. 6 No. of packets transmitted per data
packet delivered v/s Mobility

c) Average of End-To-End Delay: Figures 7
shows the average end to end delay as a function
of mobility and data rate. In figure 7 data rate is
set to 300Kbit/s since TAODV uses all
discovered routes simultaneously its end-to-end
delay is less than AODV.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO216-6




Fig. 7 Average end to end delay v/s Mobility

V. CONCLUSION AND FUTURE WORK
We introduced a new multipath routing that uses
discovered paths simultaneously; this technique
applied to AODV and evaluated via several
simulations scenarios. Simulation results show
that our protocol have better packet delivery ratio
in compare of AODV. We are going to examine
our work over other routing protocols such as
DSR and evaluate its performance with different
scenarios. Also, we will try to make TAODV
perform well in large size of routing table.

REFERENCES
[1] E. M. Royer and C.-K. Toh, “A Review of
Current Routing Protocols for Ad-Hoc Mobile
Wireless Networks,” IEEE Pers. Commun., Apr.
1999, pp. 46–55.
[2] A. H. Altalhi and G. Richard III, “Load-
Balanced Routing through Virtual Paths: Highly
Adaptive and Efficient Routing Scheme for Ad
Hoc Wireless Networks,” 23
rd
IPCCC, 2004.
[3] P. Johansson et al., “Scenario-based
Performance Analysis of Routing Protocols for
Mobile Ad Hoc Networks,” Proc. ACM/IEEE
MobiCom, Aug. 1999.
[4] ] Park, S., Shin, J., Baek, S., Kim,
S.C.,"AODV-Based Routing Protocol Considering
Energy and Traffic in Ad Hoc Networks." In Proc. of
International Conference on Wireless Networks.
(2003) 356-36.
[5] Cho, H.K., Kim, E.S., Kang, D.W."A load-
balancing routing considering power
conservation in wireless ad hoc networks” Proc.
of 16th International Workshop on Database and
Expert Systems Applications. (2005) 128-132.


[6] S.-J. Lee and M. Gerla, "AODV-BR: Backup
Routing in Ad hoc Networks," Proceedings of
IEEE WCNC 2000, Chicago, IL, Sep.2000.
[7] C. E. Perkins, E. M. Belding-Royer, and S.
Das, "Ad hoc On-Demand Distance Vector
(AODV) Routing", RFC 3561, July 2003.
[8] C. E. Perkins and E. M. Royer, "The Ad hoc
On-Demand Distance Vector Protocol", In C. E.
Perkins, editor, Ad hoc Networking, pages
173.219. Addison-Wesley, 2000.
[9] A. Nasipuri and S.R. Das, "On-Demand
Multipath Routing for Mobile Ad Hoc
Networks," Proceedings of IEEE ICCCN'99,
Boston, MA, Oct. 1999, pp. 64-70.
[10] V.D. Park and M.S. Corson, "A Highly
Adaptive Distributed Routing Algorithm for
Mobile Wireless Networks," Proceedings of
IEEE INFOCOM' 97, Kobe, Japan, Apr. 1997,
pp. 1405-1413.
[11] Lokesh Bajaj, Mineo Takai, Rajat Ahuja,
Ken Tang, Rajive Bagrodia, Mario Gerla, UCLA
“Parallel Computing Laboratory, GloMoSim: A
scalable simulaton environment for wireless and
wired network system.” CA90095







CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO302-1

Performance of Cross-Shaped Junction Microstrip
Antenna using IE3D


Jyoti Gupta, Babita Singh, D. C. Dhubkarya
Department of Electronics & Communication Engineering,
B.I.E.T., Jhansi (U.P.) INDIA
gold.jyoti@gmail.com, babitasngh20@gmail.com,dcd3580@yahoo.com


Abstract— In this paper a novel Cross-
Shaped patch antenna is presented.the
operating frequency of antenna is 2.25
GHz.The dielectric constant is 4.2,
thickness of antenna is 1.6 mm and loss
tangent is 0.0013.Simulation result at
operating frequency 2.25 GHz of antenna
are done by the help of IE3D Zeland
software (Version 12.0).The antenna is
small and is suitable for wireless portable
device.The bandwidth achieved is 24%.
Index Terms ― microstrip antenna,
cross shaped antenna.


I. INTRODUCTION

In recent years, microstrip printed antennas
have been of great interest to researchers
and engineers and have a wide spectrum of
applications. The popularity is due to its key
attractive features such as low cost, small
weight and ease of fabrication. The most
notable problem of patch antennas is that its
size tends to be large at the low microwave
frequencies. Hence, wireless applications
require antenna size that is compliant with
the space restrictions associated with the
portable devices.
There are a few methods presented in the
literature to reduce the size of a patch
antenna [1 - 3]. These include using high
permittivity substrates, loading the patch
with shorting pins and meandering the
ground plane or the perimeter of the patch.
In the case of using a high permittivity
substrate, the dielectric constant and the
resonant frequency of the antenna are related
as shown in (1).


(1)

From the above equation, we can see that as
the dielectric constant (ε
r
) increases, the
resonant frequency (f
r
) and the patch
antenna dimension decreases. However, it
should be noted that using high permittivity
substrate results in the excitation of greater
surface waves, which reduces the
performance of the antenna.

In this paper the above-mentioned size
reduction techniques are investigated on
probe fed rectangular antennas. The results
are then analyzed to determine which size
reduction technique gives the highest
reduction in the overall antenna size
compared to the conventional probe fed
antenna. A new cross-shaped shorted
antenna for mobile communications is then
designed. The resonant frequency of the
antenna is designed to be around 2.25 GHz
as shown in figure 1.



CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO302-2



A

C


B
D

E F

Figure 1: Cross -shaped Patch Antenna
with Parameters A=B=36mm,
C=D=E=F=12mm, Frequency f = 2.25
GHz, ε
r
= 4.2, h=1.6mm.


II. ANTENNA DESIGN &
STRUCTURE

In this paper, the Cross-shape microstrip
patch antenna has been designed with over
all dimensions 36mm×36 mm. The
designing of Cross-shaped microstrip patch
antenna, the resonant frequency f
r
= 2.25
GHz and the dielectric substrate glass epoxy
is used for fabrication. The dielectric
constant of the substrate is ε
r
= 4.2 and
thickness of the substrate h = 1.6mm to
design the Cross-shaped microstrip patch
antenna. The width and length of
the microstrip antenna are determine as
follows-

Effective dielectric constant

= (2)

Length of the patch is given by

(3)

Width of rectangular patch is given by

(4)

The equation (2),(3),(4) are used to
determine the length and width of a patch
antenna.




Figure 2: Design of cross-shaped patch
antenna on IE3D with feed point (19,15).
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO302-3


Figure 3: Return loss graph by using
IE3D
Figure 4: simulation result for VSWR

III. RESULTS
Figure 2 represents the design of Cross-
shaped patch antenna with feed point
(19,15).Notices that the central resonant
frequency is near to 2.25 GHz and, in
particular, at -21 db as shown in figure 3.
Fig. 4 shows the simulated VSWR
characteristics of the proposed patch
antenna. The impedance bandwidth (VSWR
< 2) of the configuration is 24%. The
bandwidth is quite significant for such a
small antenna.


IV. CONCLUSION

The cross-shaped shorted antenna was
simulated using Zeland’s IE3D
Software.The Cross-shaped antennas are
suitable choices for radiating element
because of the relative wider bandwidth
compared to other resonant conventional
microstrip. The choice of 2.25 GHz has
been done because of the facility to design
IE3D simulation, for 50 Ώ environment,
and due to the available material.But other
bands can be used in a more specific project.
Moreover the Cross-shaped antenna has a
good portability and compact structure.


REFERENCES

[1] D.P. Chrissoulidis, P.C. Liakow, and
D.Y. Notis, “Dual polarized
microstrip patch antenna, reduced in size by
use of peripheral slits,”
Aristotle university of Thessalaniki, Greece.
[2] R.B. Waterhouse, “Small microstrip
patch antenna,” Electron. Lett., vol. 31, pp.
604-604, April 1995.
[3] H. K. Kan and D. Pavlickovski, “Small
dual C–slot printed antenna,”
Electron. Lett., vol. 39,(7), pp. 593-594,
April 2003.
[4]J.Lal & H.K.Kan, “cross shaped shorted
patch antenna”
[5] C.A. Balanis, Antenna Theory: Analysis
and Design, 2
nd
ed.,New York: Jhon Wiley,
1997
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO303-1

Broadband Slot Loaded Microstrip Patch Antenna

J. A. Ansaari , kamakshi, Ram Brij Ram, A.Singh, Anurag Mishra, N.P.Yadav
Department of Electronics and Communication
University of Allahabad

E-mail: kamakshi_iert_el@yahoo.co.in, mishraanurag31@gmail.com


Abstract-- A U-slot loaded microstrip patch
antenna is analyzed using circuit theory concept.
It is found that broadband behaviour of the
antenna is achieved due to the effect of U-slot
inserted into the patch. The optimized antenna
shows operational frequency band of 1.63 GHz
with impedance bandwidth of 34.72 %. A
comparison between theoretical and simulated
results is also presented.

Key-words: broad band, microstrip antenna, patch
antenna, U-slot

1. INTRODUCTION
Microstrip antennas have become attractive
candidates in a variety of commercial applications
such as mobile and satellite communications.
Traditionally, microstrip antennas suffer from low
bandwidth characteristic. But many of
communication systems require wide bandwidth
which is not provided by the conventional microstrip
antenna. Hence serious efforts started among the
scientific community to remove its inherent
drawback of narrow bandwidth. There are numerous
and well known methods to increase the band width
of the antennas including: the use of the substrate
thickness [1], the use of low dielectric substrate [1],
the use of various impedance matching and feeding
techniques [2], the use of multiple resonators [3-5]
and the use of slot antenna geometry [6-8]. Among
these, the use of multiple resonators in stacked or
coplanar antenna geometries is very promising
method to enhance impedance bandwidth. However,
the stacked geometry increases the thickness of the
antenna on one hand and the coplanar geometry
increases its lateral size on the other hand. Therefore,
these geometries are not preferred for most of the
modern wireless communication systems that require
single layer single patch broadband microstrip
antenna.
Huynh and Lee [1], in 1995, experimentally
reported that the impedance bandwidth of single-
layer single-patch microstrip antenna could exceed
30 % if U-slot is cut into the patch. But the main
disadvantage of this antenna is its electrical thickness
which causes additional inductive component
associated with the input impedance. In this
endeavor, therefore, U-slot loaded microstrip patch
antenna with reduced thickness and wider bandwidth
is presented. The theoretical investigation was carries
out using circuit theory concept based on modal
expansion cavity model. The results so obtained are
compared with IE3D simulated [**] and
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO303-2

experimental results; the details which are given in
the following sections:

II. ANTENNA GEOMETRY AND
THEORETICAL FORMULATIONS
The geometrical configuration of the U-slot
loaded microstrip patch antenna is shown in Fig. 1.
The proposed antenna consists of a rectangular patch
of length L =36 mm and widthW =26 mm. The
rectangular patch is printed on a dielectric substrate
(foam) of constant
r
c =1.03 and thickness h =5 mm.
A U-slot of side length
s
L =20 mm, base-length
s
W =12 mm and slot width T =2 mm is incorporated
in the middle of the patch so that the side-arms are
symmetrically positioned with respect to the feed
point. The feed point (
0
X ,
0
Y ) is determined with
respect to the center of the patch as (0,-1.8). A simple
microstrip patch antenna can be modeled as a simple
LC-resonant circuit [9]. When U-slot is incorporated
into the patch, the resonance feature is changed as
shown in Fig. 2. The part of the patch enclosed by the
U-slot has current distribution like normal patch and
it resonates at a certain frequency. However, for the
part of the patch beyond the U-slot, the current has to
flow around the side and base arms of U-slot which
causes lengthened current path. As a result, currents
along the edges of U-slot introduce an additional
resonance in conjunction of the main patch. This
additional resonance frequency is determined by
L
2
C
2
; here L
2
and C
2
are inductance and capacitance
of the patch with U-slot. Thus U-slot loaded patch
has two resonance frequencies. If these two
resonances appear in one operational frequency band,
antenna will show broadband characteristic.

(a)


(b)
Fig. 1 Antenna geometry (a) top view (b) side view

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO303-3


Fig. 2. Current distribution at frequency 4.51 GHz

According to modal expansion cavity model a
microstrip patch is considered as a parallel
combination of capacitance, inductance and
resistance. Therefore, the equivalent circuit for main
patch is as shown in Fig. 3 (a), in which
1
C ,
1
L and
1
R are capacitance, inductance and resistance of the
main patch. The circuit parameters are calculated as
[9]. The impedance of the main patch without U-slot
can be derived as
1
1
1
1 1
1
L j
C j
R
Z
P
e
e + +
= (1)
The U-slot cut in the lower patch can be considered
as a combination of three narrow slots joint together
in the form of U-shape. Using duality relationship of
dipole and slot, the equivalent circuit of U-slot can be
represented as shown in Fig. 3(b) [10]. Here,
UH
R ,
UH
X and
UV
R ,
UV
X are radiation resistance and
reactive component of the base-arm and side-arm of
the U-slot that can be calculated as [10]. Therefore,
from the equivalent circuit of U-slot shown in Fig.3
(b), the impedance of the U-slot can be derived as
UH UV
UH UV
U
Z Z
Z Z
Z
2
2
+
= (2)

(a)

(b)

(c)
Fig. 3 Equivalent circuits of (a) fed patch without U-
slot, (b) U-slot and (c) U-slot loaded patch

Further, the U-slot loaded microstrip patch
antenna can be considered as a parallel combination
of impedances of the patch and the impedance of U-
slot as shown in Fig. 3(c), therefore, the impedance
of the U-slot loaded microstrip patch antenna can be
derived as
U P
U P
p n
Z Z
Z Z
L j Z
+
+ = e
1
(3)
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO303-4

where
p
L is the inductance due to co-axial probe of
50 ohm.
The return loss of the antenna can be calculated as
I = log 20 RL (4)
where
0
0
Z Z
Z Z
in
in
+
÷
= I
where Z
0
is the input impedance of the coaxial feed
(50 Ω)
The radiation pattern of U-slot microstrip
patch antenna can be calculated as [11]
2
0
cos sin sin
2
cos
sin sin
2
sin sin
2
sin
) cos cos( ) (
0
0
0
0
0
t
o
o o u
u o
u o
u
t
o
s s
|
.
|

\
|
×

|
.
|

\
|
=
÷
where
L k
W k
W k
kh
r
WVe jk
E
e
r jk
(5)
2
0
sin cos sin sin
2
cos
sin sin
2
sin sin
2
sin
) cos cos( ) (
0
0
0
0
0
t
u
o u o u
u o
u o
u
t
u
s s
|
.
|

\
|
×

|
.
|

\
|
÷ =
÷
where
L k
W k
W k
kh
r
WVe jk
E
e
r jk
(6)
where V is radiating edge voltage, r is the distance of
an arbitrary point, and
r
k k c
0
=


III. CALCULATIONS AND DISCUSSION OF
RESULTS
The calculation for variation of return loss
with frequency were carried out using equations (3-
4), the resulting data are shown in Fig. 4. It is
observed that there exist two resonance frequencies at
4.108 GHz and 5.17 GHz. The two resonances are so
closely spaced that the antenna gives broadband
characteristics. The calculated operational frequency
band ranges from 3.88 GHz to 5.51 GHz (i.e. band of
163.6 MHz) and antenna shows bandwidth 34.72 %.
On the other hand, the simulated [12] return loss
curve shows frequency band of 154.8 MHz with
bandwidth of 33.04 %. The theoretical result is in
good agreement with the simulated result. Similar
results were reported experimentally by Lee et al [6]
and Huynh et al [8]. Very small errors appear in the
resonance frequencies and the bandwidth due to
approximation in calculations of the impedances of
the patch and the U-slot.
3.5 4 4.5 5 5.5 6
-16
-14
-12
-10
-8
-6
-4
-2
Frequency (GHz)
R
e
t
u
r
n

l
o
s
s

(
d
B
)
Theoretical
Simulated

Fig.4 Variation of return loss with frequency
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO303-5

-80 -60 -40 -20 0 20 40 60 80
-30
-25
-20
-15
-10
-5
0
Angle (degree)
R
e
l
a
t
i
v
e

R
a
d
i
a
t
i
v
e

P
o
w
e
r

(
d
B
)
Theoretical
Simulated

Fig. 5 Radiation pattern at frequency 4.517 GHz
The calculations for radiation pattern for the
U-slot microstrip patch antenna at frequency 4.517
GHz were done using equations (5-6), the data so
obtained are shown in Fig.5. It is noted that the
calculated radiation pattern is also almost similar to
the simulated radiation pattern. Antenna shows gain
of 5.37 dBi to 9.02 dBi in the operation frequency
band. As shown in Fig. 6.
3.5 4 4.5 5 5.5 6
2
3
4
5
6
7
8
9
10
Frequency (GHz)
G
a
i
n

(
d
B
i
)

Fig.6 Variation of gain with frequency
IV. CONCLUSIONS
It is, therefore, concluded that the existence of
two resonance frequencies provides broad band
characteristic to the U-slot microstrip patch antenna.
The optimized antenna can be made to operate with
bandwidth of 34.72 % in frequency band from 3.88
GHz to 5.17 GHz.

REFERENCES
[1] D. H. Schaubert, D. M. Pozar and A. Adrian,
“Effect of microstrip antenna substrate thickness
and permittivity: Comparison of theories and
experiments”, IEEE Trans. Antenna Propag.,
vol. 37, pp. 677-682, Jun. 1989.
[2] H. F. Pues and A. R. Van De Capelle, “An
impedance technique for increasing the
bandwidth of microstrip antennas”, IEEE Trans.
Antenna Propag., vol. 37, no. 11, pp. 1345-1354,
Nov. 1989.
[3] G. Kumar and K. C. Gupta, “Broad band
microstrip antennas using additional resonators
gap coupled to the radiating edges”, IEEE Trans.
Antenna Propag., vol. 32, no. 12, pp. 1375-1379,
Dec. 1984.
[4] J. A. Ansari and Ram Brij Ram, “ Analysis of
compact and broadband microstrip patch
antenna,” MIcrow. Opt. Tech. Lett., vol. 50, no.
8, pp. 2059-2063, 2008.
[5] Z. F. Liu, P. S. Kooi, L. W. Li, M. S. Leong and
T. S. Yeo, “A method for designing broad band
microstrip antennas in multilayered planar
structures”, IEEE Trans. Antenna Propag., vol.
47, no. 9, pp. 1416-1420, Sept. 1999.
[6] K. F. Lee, K. M. Luk, K. F. Tong, S. M. Shum,
T. Huynh and R. Q. Lee, “Experimental and
simulation studies of the co-axially fed U-slot
rectangular patch antenna”, IEE Proc. Microw.
Antenna Propag., vol. 144, no. 5, pp. 354-358,
1997.
[7] R. Bhalla and L. Shafai, “Resonance behavior of
single U-slot and dual U-slot antenna”, IEEE
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO303-6

Antenna and Propagat. Society Int. Symp., vol.
2, pp. 700-703, July 2001.
[8] T. Huynh and k. F. Lee, “Single layer single
patch wideband microstrip antenna”, Electronics
Letters, vol. 31, no. 16, pp. 1310-1312, Aug.
1995.
[9] I. J. Bahl and P. Bhartia, “ Microstrip antennas”,
Dedham, MA: Artech House, 1981
[10] J. A. Ansari and R. B. Ram, "Analysis of broad
band U-slot microstrip patch antenna,"
Microwa. Opt.Tech. Lett., Vol. 50, No. 4, pp.
1069-1073, 2008.
[11] K. R. Carver and J. W. Mink, “Micro strip
antenna technology”, IEEE trans. Antenna
Propag., vol. 29, pp. 2-23, 1981.
[12] IE3D Zeland Software, Inc.(USA), Version
14.05,





CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO304-1

1

Abstract— Modern microwave communication systems require,
especially in satellite and mobile communications, high-
performance wide-band bandpass filters having low insertion loss
and high selectivity together with linear phase or flat group delay
in the passband. Usually, these criteria are fulfilled using
waveguide cavity or dielectric resonator loaded cavity filters
because of their low loss. However, in order to reduce size, weight
and cost, there has been a growing interest in planar structures.
The disadvantage of high conductor loss of the planar filters using
conventional conducting thin films can be overcome by replacing
them with high-temperature superconducting (HTS) thin films.
These can have a very low conductor loss. An alternative is by
combining with active MMIC devices to compensate the loss.
Under those driving forces many novel types of microstrip filters
have recently been developed.
In this paper we discuss end coupled band pass filters at
millimeter wave frequency for same permittivity condition. Both
equal-ripple and butterworth end coupled band pass filters are
designed at 6 GHz for 3.02 permittivity value and having a FBW
of 0.083. The cost and the fabrication complexity are much less
since it uses the thick-film process. The design and simulation
are performed using 3D full wave electromagnetic simulator IE3D .

Key Words—Band pass filter, WiMAX, coupled lines, Micro strip
filter, Millimeter wave filters.
I.INTRODUCTION
he advances of telecommunication technology arising
hand in hand with the market demands and governmental
regulations push the invention and development of new
applications in wireless communication.

Manuscript received March, 5 2011. This work was supported in part by the
Department of Electronics & Communication Engineering, Krishna Institute
of Engineering & Technology, Ghaziabad.
Avnish Kumar is Student with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone: 9997866775; (e-mail: avi.k.prajapati@gmail.com)
Navita Singh is Sr. lecturer with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone: 9412582436; (e-mail:singh.navita@ rediffmail.com)
Manohar Kumar is Student with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone: 9634477323; (e-mail: manoharkumarald@gmail.com)
Ashutosh Kumar Varun is Student with Department of Electronics &
Communication, Krishna Institute of Engineering & Technology, Ghaziabad,
INDIA. Phone: 8755478459; (e-mail: ashu.banty111@gmail.com)




These new applications offer certain features in
telecommunication services, that in turn offer three important
items to the customers. The first is the coverage, meaning each
customer must be supported with a minimal signal level of
electromagnetic waves, the second is capacity that means the
customer must have sufficient data rate for uploading and
downloading of data, and the last is the quality of services
(QoS) which guarantee the quality of the transmission of data
from the transmitter to the receiver with no error. In order to
provide additional transmission capacity, a strategy would be
to open certain frequency regions for new applications or
systems. WiMAX (Worldwide interoperability Microwave
Access) which is believed as a key application for solving
many actual problems today.
In realization of such a system like WiMAX we need a
complete new transmitter and receiver. A bandpass filter is an
important component must be found in the transmitter or
receiver. Bandpass filter is a passive component which is able
to select signals inside a specific bandwidth at a certain center
frequency and reject signals in another frequency region,
especially in frequency regions, which have the potential to
interfere the information signals. In designing the bandpass
filter, we are faced the questions, what is the maximal loss
inside the pass region, and the minimal attenuation in the
reject/stop region.

In the process to fulfill these requirements there are several
strategies taken in realization of the filters, for example, the
choice of waveguide technology for the filter is preferred in
respect to the minimal transmission loss (insertion loss). This
strategy is still actual in satellite applications. The effort to
fabricate waveguide filters prevents its application in huge
amounts.

As alternative, microstrip filter based on printed circuit board
(PCB) offers the advantages easy and cheap in mass
production with the disadvantages higher insertion losses and
wider transition region. In this work we would like to give a
way to conceive, design and fabricate bandpass filter for the
WiMAX application at the frequency 6 GHz with end-coupled
microstrips, which designed filter for wireless local area
network 6.25 GHz, and which used the capacitive resonators
and stepped impedance resonators for filter realization.
Therefore, by extension, the RF/microwave applications can
be referred to as communications, and other that explore the
Microstrip Bandpass Filter at 6 GHz for
WiMAX

Avnish kumar, Navita Singh, Manohar Kumar, Ashutosh Kumar Varun


T
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO304-2

2
usage of frequency spectrums; some of these frequency
spectrums are further divided into many frequency bands.
In this paper, filters are optimized for high performance and an
efficient and cost effective fabrication methods is
proposed.Microstrip technology is used for simplicity and
ease of fabrication. The design and simulation are performed
using 3D full wave electromagnetic simulator IE3D. This filter
is widely used today in radar, satellite and terrestrial
communication applications.

II.MICROSTRIP FILTER DESIGN

The general configuration of an end coupled microstrip band
pass filter is illustrated in fig. where each open end microstrip
resonator is approximately a half guided wavelength long at
the midband frequency f
0
of the bandpass filter [1]. The
coupling from one resonator to the other is through the gap
between the adjacent open ends, and hence is capacitive. In
this case, the gap can be represented by the inverters[1]. These
J-inverters tend to reflect high impedance levels to the ends of
each of the half wavelength resonators, and it can be shown
that this causes the resonators to exhibit a shunt type
resonance. Thus, the filter under consideration operates like
the shunt resonator type of filter whose general design
equations are give as follows:

J
01
/Y
0
= √(∏ FBW/2g
0
g
1
) (1)

J
j,j+1
/Y
0
= (∏ FBW/2)1/√ (g
j
g
j+1
) j = 1 to n-1 (2)

J
n,n+1
/Y
0
= √(∏ FBW/2g
n
g
n+1
) (3)

Where g
1
, g
2
, g
3
, g
4......
, g
n .
are the element of a ladder low pass
protype with a normalized cutoff Ωc = 1.0, and FBW is the
frational bandwidth of bandpass filter. The J
j,j+1
are the
characteristic of J-inverters and Y
0
is the characteristics
admittance of the microstrip line[3].

Assuming the capacitive gaps act as perfect, series-
capacitance discontinuities of susceptance B
j,j+1
as

B
J,J+1
/Y
0
= (J
j,j+1
/Y
0
)/(1-( J
j,j+1
/Y
0
)
2
) (4)

Ө
j
= ∏ -1/2[tan
-1
(2 B
J,J+1
/Y
0
) + tan
-1
(2 B
J-1,J
/Y
0
)] radians
(5)
Where the B
J,J+1
and Ө
j
are evaluated at fo. Note that the
second term on the right hand side of second equation
indicates the absorption of the negative electrical length of the
J-invewrters associated with the jth half wavelength resonator.

As referring to the equivalent circuit of microstrip gap, the
coupling gaps s
j,j+1
of the microstrip end- coupled resonator
filter can be determined as

C
g

j,j+1
= B
J,J+1

o
(6)



Where ω
o
= 2∏

f
o
is the angular frequency at the midband.

The physical length of resonator are given by

l
j
= (λ
g0
/2∏ ) Ө
j
- ∆ l
j
e1
-

∆ l
j
e2
(7)
where ∆ l
j
e1,e2
are the effective lengths of the shunt
capacitances on the both ends of the resonator j. because the
shunt capacitances C
p

j,j+1
are associated with the series
capacitances C
g

j,j+1
as defined in the equivalent circuit of
microstrip gap, they are also determined once C
g

j,j+1
in
equation are solved for the required coupling gaps. The
effective lengths can then be found by[1]

∆ l
j
e1
= (ω
o
C
p

j-1,j
/Y
0
)(λ
g0
/2∏ ) (8)

∆ l
j
e2
= (ω
o
C
p

j,j+1
/Y
0
)(λ
g0
/2∏ ) (9)

III.SIMULATION RESULTS
In order to verify the validity of the above expressions in
millimeter wave regime, a simulation study was performed
using IE3D. To get the exact response for our purpose,an
optimization was performed using software.The dimensions of
the filters are given in the following table.the simulated filter
responses are depicted below.

TABLE I

DIMENSIONS OF THE FILTERS FOR n=3 (in mm)

Type Physical
Length
Width Gap
Equal-ripple
BPF


Butter-worth
BPF

l
1
=l
3
=9.47
l
2
=10


l
1
=l
3
=8.21
l
2
=8.47
W=1.6



W=1.6
s
01
=s
34
=0.038
s
12
=s
23
=0.81


s
01
=s
34
=0.075
s
12
=s
23
=0.810

Fig.1. Layout of equal-ripple BPF on a substrate with a
relative dielectric constant of 3.02 and a thickness of 0.672mm

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


COMO304-3

3

Fig.2.Full-wave EM simulated frequency response of the filter
for εr = 3.02 and h =0.672mm


Fig.3. Layout of end coupled BPF on a substrate with a
relative dielectric constant of 3.02 and a thickness of 0.672mm



Fig.4.Full-wave EM simulated frequency response of the filter
for ε
r
= 3.02 and h = 0.672mm.



IV. RESULTS & ANALYSIS
The conventional geometry of the chebyshev band pass filter,
shown in figure 1. the simulated response of the conventional
geometry in figure 2. The attenuation decreases at 6 GHz
frequency shows in figure 2 for the band edge frequency of
5.85 GHz and 6.15 GHz. The conventional geometry of the
Butterworth band pass filter, shown in figure 3. The simulated
response of the conventional geometry in figure 4, the
attenuation decreases at 6 GHz frequency shows in figure. for
the band edge frequency of 5.75 GHz and 6.75 GHz, we
getting a sharp cutoff at 6 GHz & maximally flat response
without any ripples in the pass band with non attenuated
frequency in the desired band. Considerable improvements in
the frequency response can be seen.

V. CONCLUSION
The filters are the very essential part of the microwave system,
not only in microwave but they are also very important in
communication field. Any communication system cannot be
designed without filters. The simulated end coupled band pass
filters achieved an insertion loss of less than 1.0dB and both
filters are designed on 0.672 mm thick substrate for ε
r
= 3.02
with loss tangent 0.0016. Considerable improvements in the
frequency response can be seen.

ACKNOWLEDGEMENT
The authors would like to thank authorities of Krishna
Institute of Engineering & Technology for all the support
provided.

VI.REFERENCES
1. Jia-Sheng Hong; Lancaster M.J, “Recent progress in
planar microwave filters,” IEEE Trans. Antennas
Propagat., Vol. 2, pp. 1134 – 1137, August 1998.
2. .Jia-Shen G. Hong & M.J. Lancaster, “Microstrip Filters
for RF/ Microwave Applications” John Wiley & Sons
Inc., 2001.
3. D.M.Pozar, “Microwave Engineering,” John Wiley, 2000.
4. G. Mathaei, L.Young & E.M.T. Jones, “Microwave Filter
impedance matching networks and coupling structures,”
Artech House, Norwood, MA, 1980.
5. IE3D Software Release – 8, Developed by M/S Zeland
Software Inc.
6. Tae-Yeoul Yun; and Kai Chang, “Transaction on
Microwave Theory and Techniques,” IEEE Trans, Vol 49,
No.3, March 2001.
7. M.Makimoto; and S.Yamoshita, “Microwave Resonators
and Filters, “IEEE Trans.Wireless communication Vol.2,
August 1986.
8. H.Ozalki and J.Ishii, “Synthesis of transmission -line
networks and the Design of UHF filters,” IRE Trans. On
circuit theory, vol. CT-2, pp.325- 336; December 1955.
9. E.N. Torogow, Senior member, IEEE, and G. E Collins,
“Band-Stop Filters for High-Power Applications”, IEEE
Trans. Vol 40, pp 185-196; September 1965
10. Jen-Tsai Kuo,,“ Parallel Coupled Microstrip Filters With
Suppression of Harmonic Response” In IEEE, Wei-Hsiu
Hsu, and Wei-Ting Huang Vol. 12, No. 10, Oct 2002 383
11. P. I. Richards, " Resistor-transmission-line circuits,"
Proc. IRE, Vol. 36, pp. 217-220, Feb. 1948.
12. M. Dishal, "Design of dissipative band-pass filters
producing desired exact amplitude-frequency
characteristics", Proc. Inst. Radio Engrs, 37, pp. 1050-69,
September 1949






CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO305-1


Abstract — In multiple-input multiple-output (MIMO) broadcast
channels, multiuser diversity is exploited by scheduling data
transmission to users with best channel conditions. To find the
best set of users, the base station requires information about user
channels, which may lead to a great increase in feedback
overhead. In multiple-antenna systems the multi-user capacity
depends greatly on whether the transmitter knows the channel
state information (CSI) or not. The Dirty Paper Coding (DPC)
scheme can be employed at the transmitter side to achieve the
optimal sum-rate capacity. In systems with large users, obtaining
full CSI from all users may not be practicablet and very complex
hardware implementations are necessary for DPC. On the other
hand, scheduling based on partial channel state information (CSI)
often results in a loss in sum-rate capacity. We consider an
efficient scheduling scheme for multi-user MIMO broadcast (BC)
channel with partial CSI feedback. In this paper, we propose a
scheduling scheme for a MIMO BC downlink channel with M
transmit antennas at the BS and N users at the receiver side with
K antennas each, which allocates independent data streams to
each of the transmit antenna to transmit to the most favourable
receiving antennas. The receiving antennas are selected based on
their signal-to-interference-plus-noise ratio (SINR). Out of total
NK receive antennas M antennas are selected with the highest
SINR values. The resultant throughput versus amount of
feedback is analyzed and it is observed that the proposed
scheduling algorithm provides more throughput than other
antenna selection algorithms with partial CSIT.

Index Terms —MIMO, Broadcast Channel, Channel State
information, Antenna selection.

I. INTRODUCTION
ultiple-Input Multiple-Output (MIMO) is one of the
important communication techniques that allow wireless
systems to achieve higher data rate [1] – [3]. Another
benefit of MIMO is the use of spatial diversity which can be
used to provide more reliable communication [4]. MIMO can

Swadhin Kumar Mishra is with the National Institute of Science and
Technology, Berhampur, Orissa, India (Phone: 9776644121; e-mail:
swadhin.mishra@gmail.com).

Prabina Pattanayak is with the National Institute of Science and
Technology, Berhampur, Orissa, India (Phone: 8018148448; e-mail:
prabina007@gmail.com).

A K Panda

is with the National Institute of Science and Technology,
Berhampur, Orissa, India. (Phone: 9437014147; e-mail: akpanda62@
hotmail.com).

achieve maximum spatial diversity gain by using Space-time
Block coding (STBC)[5].One key benefit of using MIMO is
the resultant high spatial multiplexing gain which can be
obtained by sending independent message streams over any
transmit- receive antenna pair simultaneously. The very well
known DPC (Dirty Paper Coding) has established the idea that
wireless systems can achieve data rate very close to the
fundamental capacity limit of the system by serving multiple
users simultaneously. In multi-user MIMO systems which have
many users at the receiver side, it is better that the transmitter
knows about the channel. So to have full Channel State
Information (CSI) at the base station, a huge amount of
feedback is required from the receivers to the transmitter (Base
Station). In case of DPC, full CSIT is required and the
implementation of it is very difficult. A relatively high
bandwidth of the uplink channel is being consumed to carry
such a huge amount of feedback. So, now the main aims of the
researchers are to find an efficient algorithm which will
consume less feedback but should give a throughput as much
as high it could be.
To reduce the amount of channel feedback and the
computation complexity, some work [6] – [10] has already
been done on the user selection criteria. The amount of
feedback is reduced in [11]. The algorithm uses limited
feedback. Each of the users at the receiver side selects a pair of
transmits and receive antenna which provide the maximum
Signal-to-Interference plus Noise Ratio (SINR). For large
number of users it is shown that this scheme achieves near
optimal sum-rate. It is also shown that by maintaining a
threshold on the SINR value, the amount of feedback can
significantly be reduced without much loss in the system
throughput.
Another SINR Feedback algorithm is discussed in [12]
where each user is having more than one receive antennas. The
basic idea behind the scheduling algorithm described here is
the each user firstly selects the most favorable transmit and
receive antenna for achieving the maximum SINR and then
feeds back the corresponding maximum SINR along with the
index of the transmit antenna. This scheme is similar to the
scheme described in [11]. But it differs from [11] in that
receive antenna selection is used for each user, whereas in [11]
every antenna is employed as an individual user. Since receive
antenna selection is performed at each user, the total feedback
load at each time slot is reduced without loss of throughput as
opposed to [11]. The disadvantage of the proposed scheme in
[12] is that only one antenna with the highest SINR is selected
per user. If an antenna of a user has SINR value, less than the
antenna of the same user with the highest SINR but more than
Antenna Selection Algorithm for MIMO BC
Channel Using Partial Side Information
Swadhin Kumar Mishra
1
, Prabina Pattanayak
2
, A K Panda
3

M
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO305-2

an antenna of another user with maximum SINR then the
selection algorithm will not choose the antenna. Hence using
this scheme not all the best antennas are selected. All these
scheduling schemes are implemented for independent and
identical distributed (IID) channels.
In this paper, we propose an algorithm for antenna selection
using partial feedback and compared the throughput
performance with that of [12]. In the proposed scheme spatial
multiplexing and receive antenna selection are both employed.
Here we propose that each user finds out a set of M highest
SINR values corresponding to the M transmit antennas, and
also the set of M indices of transmit and receive antennas
which achieve the highest SINR. The set of SINR values along
with the set of indices of transmit and receive antennas are
feedback to the Base Station. The base station transmits M
independent or same message streams to M receive antennas,
irrespective of user to whom they belong with the highest
SINR. It will be shown that our schemes provides more
throughput than [11], [12].
The rest of the paper is organized as follows. Section II
introduces the system mode. In Section III, a multiuser
scheduling algorithm is proposed for MIMO broadcast
channels and its achievable throughput is obtained. In Section
IV, results of different simulations are being presented. And
the different results are compared with the simulation results of
[12]. In Section V we presented the conclusion and scope of
future work.
The remainder of this paper is organized as follows.
System model is given in Section II. In Section III, we
describe the CLLL algorithm in details. In Section IV,
proposed MIMO BC channel scheduling based CLLL is
introduced. Simulation results and discussions are given in
Section V, and conclusions are drawn in Section VI.

II. SYSTEM MODEL
The wireless MIMO system that is considered here has M
transmit antennas at the base station and N number of users
with each users having K number of receive antennas as shown
in Fig. 1. So there are K x N number of receive antennas in all


Fig. 1. MIMO Broadcast Downlink Channel
from which the base station selects M for communication. The
received signal by the nth user at any timeslot is


n n n
w x H y (1)

where
n
y is an 1 K vector,
n
H is a M K complex
Gaussian channel matrix from the transmitter to the n
th
user,
x is an 1 M vector of the transmitted signal and
n
w is the
1 K noise vector ~ ) 1 , 0 ( CN . Each of the elements in the
channel matrix is identical and independent distributed (IID)
complex Gaussian with zero mean and unit variance i.e.
h
n
(k,m) ~CN (0,1) . h
n
(k,m) represent the complex gain from
transmit antenna m to the receive antenna k of user n. The
block fading channel model is assumed. We further assume
that the channel matrix is known at the receiver, but unknown
at the transmitter. It is assumed that the transmitter has a power
constraint P, i.e. P Tr E )} ( {
*
xx . We assume the total
transmit power is equal to unity. The total power is distributed
equally among all the antennas. Hence the average transmit
power per antenna is 1⁄M. The power constraint is assumed to
be per frame. The received signal at the
th
k antenna of user n
can be derived from (1) as

M
m
n n n
m w m x m k h k y
1
) ( ) ( ) , ( ) ( (2)
) (m x is the desired signal for user n transmitted through
Atenna m with ) (m x assumed to be independent for
individual transmit antennas. Then the signals
m m m x ), ( are considered as interference for the user.
Hence the SINR of ) (n y is given as

m m
k
n
k n
k m
m k h
M
m k h
SINR
2
2
) (
,
) , (
) , (
(3)
where
n
is the average SNR of user n.

III. SCHEDULING ALGORITHM
Here we propose an Antenna Selection algorithm where
each user feedbacks some information using the uplink
channel to the Base Station regarding the Channel State
Information (CSI). After receiving the Channel State
Information, the base station selects M number of receive
antennas and transmits M number independent message
streams from the individual transmit antennas to them.
A. Antenna Selection Algorithm
1. Let
n
m
S be the SINR measured by a particular receiving
antenna
n
mk
R of the n
th
user due to transmitting
antenna T
m
. The values of m, n and k are as follow :
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO305-3

K k N n M m 1 , 1 , 1
Initialize
n
mk
n
m
R S , and
m
T to zero.

2. Every user evaluates the Signal-to-Interference-plus-
Noise-Ratio (SINR)
n
m
S due to each of the
combination of transmitting antenna
m
T and receiving
antenna
k
R

by using (3). Each user have M×K SINR
values out of which it selects M maxiumum SINR
values corresponding to each of the transmit antenna.
The M maximum SINR values, index of the transmit
Atenna i.e.
m
T and index of the receive antenna i.e.
k
R are fed back by each user to the Base Station.

3. The Base Station (BS) after receiving all the information
feedback by the users, selects a receiving antenna with
maximum SINR for each of the transmit antenna
independent of the users as given below:
Arranging the received information in terms of each
transmit antennas.
for m=1:M

} ,......., , , {
3 2 1 N
m m m m
n
m
S S S S S
} ,......, , , {
3 2 1 N
mk mk mk mk
n
m
R R R R R
end
User selection and antenna selection.
for m=1:M


n
m n m
S n max arg
*


) ( *
*
m
n
mk m
R k
end
where
*
m
n is the user selected by the transmit antenna m and
*
m
k is the selected receive antenna for the corresponding user.
The basic idea behind the above scheduling algorithm is
that each user calculates SINR for every combination of
transmit and receive antenna and feedback the M maximum
SINR corresponding to M transmit antennas, index of
transmit antenna and receive antenna to the Base Station. At
the base station both user selection and antenna selection are
being performed. Each transmit antenna selects a user and also
the receive antenna of that user which provides the maximum
SINR. Since Antenna Selection is being performed at the Base
Station the overhead due to processing at the receiver side is
reduced. The total number of real values that will be feedback
at each time slot is N M real values (M maximum SINR
values from each user) plus M N M
2
log ) ( bits for T
m

and K N M
2
log ) ( bits for
n
mk
R where x denotes the
smallest integer larger than x.
For each transmit antenna, a set of SINR values feedback by
each user is created. Another cluster of sets are formed for
each transmit antenna consisting of the indices of receive
antennas feedback by each user.
B. Throughput Analysis
The sum-rate capacity achieved by the system using the
above Antenna Selection algorithm can be given as

) 1 (
*
*
,
m
n
n
k m
SINR C (4)
where
*
*
,
m
n
n
k m
SINR is the SINR of the selected user
*
m
n
scheduled by the
th
m transmit antenna and
*
m
k is the selected
receive antenna of user
*
m
n .
C. Feedback Load Analysis
As discussed in the scheduling algorithm, the total feedback
load at each time slot is N M real values plus
M N M
2
log ) ( bits for T
m
and K N M
2
log ) (
bits for
n
mk
R . Whereas the total feedback load at each time
slot for the antenna selection algorithm using limited feedback
in [12] is N real values plus M N
2
log bits of T
m.

IV. SIMUALTIONS AND RESULTS
In this section we present the simulation results for the
throughput analysis for the proposed scheduling algorithm
according to (4). The simulation is performed for M=2 and
K=2 over 10000 iterations. In Fig.2, the simulation result for
the achievable throughput versus number of users is shown for
0, 5 and 10 dB SNR respectively.

0 20 40 60 80 100 120 140
1
2
3
4
5
6
7
8
9
No. of Users, N
T
o
t
a
l

T
h
r
o
u
g
h
p
u
t

(
b
p
s
/
H
z
)


SNR=0dB (Proposed)
SNR=5dB (Proposed)
SNR=10dB (Proposed)

Fig. 2. Total throughput of the multiuser scheduling algorithm vs. the
number of users at various SNR 0,5 and 10dB for M=2 and N=2.




CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO305-4

0 5 10 15 20 25 30 35 40
1
2
3
4
5
6
7
8
No. of Users, N
T
o
t
a
l

T
h
r
o
u
g
h
p
u
t

(
b
p
s
/
H
z
)


SNR=0dB (Limited Feedback)
SNR=5dB (Limited Feedback)
SNR=10dB (Limited Feedback)
SNR=0dB (Proposed)
SNR=5dB (Proposed)
SNR=10dB (Proposed)

Fig. 3. Comparison of Total throughput vs. the number of users between
the proposed scheduling algorithm and the limited feedback
scheduling algorithm.


Fig. 3, presents a comparison between the proposed MIMO
broadcast scheduling algorithm and the limited feedback
scheduling algorithm described in [12]. It is shown that our
proposed scheduling algorithm performs better than the limited
feedback algorithm at the expense of slight increase in amount
of feedback. It is observed that the total throughput of both the
schemes attains nearly constant values as the number of users
increases. It is obvious as the number of users increases the
antenna selection algorithm has more number of receive
antenna to choose the best users.
It is observed that the overall performance of our algorithm
is better than the limited feedback scheduling algorithm.
Next we modify our scheduling algorithm to reduce the
overall feedback load and produce the performance of the
modified scheme. In the modified approach, instead of
feedback all the SINR values, we introduce a SINR threshold
at the receiver side. After calculating all the SINR value as (3),
the receiver side feedback the SINR values, along with the
corresponding indices, which are great than the set threshold.
A simulation is performed of the modified multiuser feedback
algorithm with M=2, N=2 at SNR of 10 dB for threshold
values of S
th
=4, 8 and 10 db. The simulation results are
compared with the original scheduling scheme as well as the
modified limited feedback algorithm in [12].Fig. 4 show the
performance of the modified scheme in terms of the total
throughput.
There are two outcomes which can be observed from the
simulation result. The modified approach of using a threshold
performs as good as the original approach for a threshold
value of S
th
= 4 dB. And it is also observed that the modified
approach implemented in our algorithm perform better than if
implemented in the limited feedback scheduling scheme of
[12].





0 5 10 15 20 25 30 35 40
0
1
2
3
4
5
6
7
8
No. of Users, N
T
o
t
a
l

T
h
r
o
u
g
h
p
u
t

(
b
p
s
/
H
z
)


lambda=4dB (Limited Feedback)
lambda=8dB (Limited Feedback)
lambda=10dB (Limited Feedabck)
lambda=4dB (Proposed)
lambda=8dB (Proposed)
lambda=10dB (Proposed)

Fig. 4. Comparison of Total throughput vs. Number of users using the
threshold Sth for the proposed method and the limited feedback method.



V. CONCLUSIONS
We have investigated various multiuser antenna selection
algorithms for MIMO broadcast channel. Our approach differs
from the other antenna selection in that both the users as well
as the antenna selection are done at the transmitter side
whereas in other antenna selection algorithms the antenna
selection is done at the user side. The total throughput
achieved by our approach is more than the other antenna
selection approach at the slight cost of feedback load. It is also
seen in the modified approach that the feedback load is greatly
reduced, by implementing a selection based on a threshold
value of SINR, with an insignificant loss in the total
throughput. All the discussed antenna selection algorithms
assume the channel to be IID. In [13], the antenna selection
algorithm of [12] is implemented for non- IID channels. The
proposed scheme can therefore be implemented for non-IID
channels. This scheduling algorithm will be having good effect
on the cases where a small increase in the feedback does not
affect the system a much.

REFERENCES
[1] I.E. Telatar, "Capacity of multi-antenna Gaussian channels”, European
Trans. Telecom., Vol. 10, No. 6, pp. 585-595, Nov. 1999.
[2] G. Caire, S.Shamai, “On the achievable throughput of a multiantenna
Gaussian broadcast Channel” IEEE Trans. Inf. Theory, Vol. 43,
pp.1691-1706, Jul 2003.
[3] S. Vishwanath, N. Jindal, A. Goldsmith, “Duality, achievable rates &
sum-rate capacity of Gaussian MIMO broadcast channel” IEEE Trans.
Inf. Theory. Vol. 49, No 10, pp 2658-2668, Oct 2003.
[4] Z.Tu and R.S.Blum, “Multiuser diversity for a dirty paper approach,”
IEEE Commun. Lett., vol. 7, pp. 370-372, Aug. 2003.
[5] S.M. Alamouti, “A simple transmit diversity technique for wireless
communication,” IEEE J.Select. Areas Commun., vol. 16, pp. 1451-
1458, Oct. 1998.
[6] K. K. Wong, R. D. Murch, and K. B. Letaief, “A joint channel
diagonalization for multi-user MIMO antenna systems,” IEEE Trans.
Wireless Commun., vol. 2,pp. 773-786, July 2003.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO305-5

[7] C. Swannack, E. Uysal-Biyikoglu, and G. W. Womell, “MIMO
broadcast scheduling with limited channel state information,” in Proc.
Annual Allerton Conf. on Commun., Control and Computing, Sep.
2005.
[8] A. Bayestech and A. Khandani, “On the user selection for MIMO
broadcast channels,” in proc. IEEE Int. Symp. Inf. Theory, Adela
Adelaide, Australia, Sept. 4-9, 2005, pp. 2325-2329.
[9] N.Jindal, W. Rhee, S. Vishwanath, S. A. Jafar and A. Goldsmith, “Sum
power iterative water-filling for multi-antenna Gaussian broadcast
channels via dual decomposition,” IEEE Trans. Inofrm. Theory, vol. 52,
pp. 754-759, Feb. 2006.
[10] N.Jindal, “MIMO broadcast channels with finite rate feedback,” in Proc.
IEEE Global Commun. Conft., St. Louis, MO, USA, Nov. 28-Dec. 2,
2005, vol. 3, pp. 1520-1524.
[11] M.Sharif and B.Hassibi, “On the capacity of MIMO boradcast channels
with partial side information,” IEEE Trans. Inform. Theory, vol. 51, no.
2, pp. 506-522, Feb. 2005.
[12] Wei Zhang, Khaled B. Lateif, “MIMO Broadcast Scheduling with
imited Feedback,” IEEE Select.Areas Commun, Vol. 25, pp. 1457–
1467, Sep.2007.S. M. Metev and V. P. Veiko, Laser Assisted
Microtechnology, 2nd ed., R. M. Osgood, Jr., Ed. Berlin, Germany:
Springer-Verlag, 1998.
[13] Swadhin Kumar Mishra, Sidhartha Panda, and C. Ardil, “MIMO
Broadcast Scheduling for Weighted Sum-rate Maximization”
International Journal of Electrical and Computer Engineering 4:15
2009, pp. 930-935, 2009.




Swadhin Kumar Mishra received his B. Eng. degree in
Electronics and communication engineering from Orissa Engineering
College, Bhubanswar, India, his M. Tech. degree in communications
system from Indian Institute of Technology,Madras in 2000 and 2009
respectively. He is presently working as an Assistant Professor in the
department of Electronic and Computer Engineering at National
Institute of Science and Technology, Berhampur, India. His current
research interests include multiuser MIMO, MIMO-OFDM, and network
coding.

Prabina Pattanayak received his B. Tech. degree in Electronics and
TeleCommunication engineering from National Institute of Science
And Technology, Berhampur, India. Presently he is pursuing M. Tech.
degree in Wireless Communication Technology at National Institute of
Science And Technology, Berhampur, India. His current research
interests include Wireless Communication, MIMO and different
scripting languages used for protocol testing.

A K Panda obtained his Ph.D degree in 1996 and is now working as
Professor at National Institute of Science and Technology. He was a
BOYSCAST fellow of Government of India to University of Michigan in
the year 1999-2000. He is a Regular Associate of Microprocessor
Laboratory of ICTP, Trieste, life member of IETE, ISTE and member
of VLSI Society of India and IEEE. He is also the coordinator of
TIFAC-CORE project on “3G-4G wireless communication”. His
research interest includes VLSI based Communication System.


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO306-1

HILBERT ANTENNAS IN MOBILE , RADAR AND
LIGHT COMBAT AIR CRAFT TECHNOLOGY
KUMAR ASHOK
[1]
, D.SUNITA
[2]
,SINGH PUSHPENDRA
[3]


[1]DEPARTMENT OF ELECTRONICS ENGINEERING,
KAMLA NEHRU INSTITUTE OF TECHNOLOGY, SULTANPUR
FORMERLY HARCOURT BUTTLER TECHNOLOGICAL INSTITUTE, KANPUR (INDIA)
e-mail- akprasad212@yahoo.co.in phone no +919415094287

[2]SCIENTIST B,ADA,MINISTRY OF DEFENCE,BANGALORE
e-mail- sunitadeviiet@yahoo.co.in phone no 9450333743
[3]HOD, DEPARTMENT OF ELECTRONICS ENGINEERING,CEST,LUCKNOW
e-mail- pushpendra_hbti05@yahoo.com phone no 9415157735


ABSTRACT
VLSI technology is boon to our communication
technology.With the widespread proliferation of
telecommunication technology in recent years, the
need for small size multi band antennas have been
increased manifold. Hilbert antenna is one of the
example for the same .By the invention of monolithic
I.C fabrication technology i.e V.L.S.I this antenna can
find many application applications in VHF/UHF
communication . A lot of Hilbert antennas can be
designed on single substrate. This yields to a
reduction of size which is the basic need of wireless
technology. As we know the Universal truth of
resistance ,that it use to to take maximum area in IC
fabrication technology which is the main motto of
LCC viz. TEJAS. Hence, for LCC the selected radome
and antennas should be such that , it has lesser
weight & highly strong ,so that It can become
aerodynamically stable
In this work, it has been proved that mutual coupling
in side by side configuration is highest than that of
collinear, stacked or echelon form and resistance
increases for the closure spacing of elements of
Hilbert antennas using Matlab software .It has also
been proved that mutual impedance is directly
proportional to directivity and radiation resistance of
transmitting and receiving antennas




Figure 1.1




CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011



INTRODUCTION
1.1 GENERAL

The role of space filling curve is most eminently used technology
for analyzing small antennas. The design of small antennas with
multi band characteristics has been the subject interest of of RF
communications in recent years. It is a continuous mapping from a
normalized one-dimensional [0,1] interval to a normalized two
dimensional 2D region [0,1].
There are many curves, which are dependent on space filling
principle. Hilbert has propped one most widely curves in
1891.Hilbert curve may provide an attractive solution for
miniaturization of Antennas since it offers a resonant structure
that may have a small footprint, as one increases the step order in
iterative filling of a 2D region.
Hilbert Antennas can be matched to a feed line with a given
characteristics impedance e.g. 50 ohms at its fundamental
resonant frequency. In this, feeder can be located closure to the
end rather than feeding it at the center.

1.2 NEED OF HILBERT ANTENNA
In the modern era of telecommunication technology the need for
small size multi band Antennas has increased manifold.
However an arbitrary reduction in the Antenna size would result in
a large reactance and deterioration of radiation efficiency. Hilbert
Antenna is very useful in VHF/UHF applications viz.
telecommunication, wireless technology etc.
1.3 ADVANTAGES OF HILBERT ANTENNA
The important advantage of using Hilbert curves/antenna is the
incorporation of its plane – filling characteristics to realize
resonant antennas with a smaller overall physical size e.g. Fourth
iterated fractal Hilbert curve geometry inscribed in a sqaure of side
7cm, is shown to result in a resonant frequency of 267Mhz, which
is much
lower than any other resonant antenna of similar size. There is no
need of balun.
1.4 PROBLEMS WITH HILBERT ANTENNA AND ITS IMPROVEMENT
The slight mismatch in this antenna is attributed to the feed
arrangement being unbalanced. The low values for the real part of
the impedance are consistent with other similar small antenna
such as Koch antenna & small meander-line antennas.
However, by using impedance matching
circuits, or even by charging the feed location, this can be
remedied. Bandwidth is narrowed for higher iteration orders while
the cross polarization level is decreased in the non-symmetrical
plane of antenna. For a Hilbert antenna occupying a fixed square
surface area of LxL and increasing ‘n’ from 1 to 5 significantly
lowers the resonant frequency by a factor of three or higher. This
yields to give reduction in bandwidth. The actual practical antenna
must satisfy the specific bandwidth, gain and pattern
characteristics, these miniature printed slot antennas with dual
band characteristics clearly show potentials for future mobile
wireless communication.
1.5 EFFFECT OF VARIOUS PARAMETERS IN HILBERT
ANTENNA
According to figure 1.2 there are following parameters, which use
to effect Hilbert antennas highly.
1.LENGTH With increase in antenna length L the two resonant
frequency shift lower. However antenna matching can be
improved by adjusting the micro strip length and substrate
thickness.
Poor impedance matches for narrower slots can be remedied by
adjusting the length of micro strip T-junction. However, with
increase in slot width there exist requirement of addition of a
shunt capacitance
2. DIELECTRIC CONSTANT The low and high resonant frequency shifts
lower as dielectric constant increases. For dielectric constant upto
4.4 the ‘S11’ characteristics are good. With further increase in
dielectric constant to 6 the ‘S11’ data degrades. Which can be
improved by proper choice of thickness and spreading (x).
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

3. SUBSTRATE THICKNESS Changing the substrate thickness ‘h’ has
significant effect on input impedance of the antenna e.g. if a
substrate thickness of 0.4mm is desirable ‘x’ needs to be increased
in order to get optimum matching. Similarly, when ‘h’ is equal to
Figure 1.2


1mm there exist requirement of decrement of ‘x’ slightly to
achieve a better high band matching.
4. SPREADING For smaller values of ‘x’ the high band matching is
superior to improve and for x=1mm both bands are fairly well
matched. It has been observed for ‘x=1.5mm the bandwidth in the
low and high bands are1.7% and 2.2% with in (-) 9.5db. Further
increase in ‘x’ deteriorates the high band performance. Increasing
‘x’ yields to better matching.


5. AXIAL LENGTH It is given by
Lax = 0.80 g
The equation used for this purpose are:
For 2.22 r 3.8
g/ o = 1.045-0.365 ln r + {6.3(w/h) r
0.945
/ (238.64+100w/h)}-
[0.148 – (8.81( r +0.95))/100 r] ln (h/ o)
For 3.8 r 9.8
g/ o =0.9217 –0.277 ln r + 0.0322 (w/h) [ r / (w/h +0.435)]
1/2

0.01 ln h/ o [4.6 – 3.65/ r
2
(w/ o)
1/2
(9.06 – 100w/ o)]
It is clear that for all the cases studied a third order Hilbert slot
antenna can be designed with an approximate axial length, Lax =
0.80 g where g is the guided wavelength given in [2].
While actual practical applications of such antennas must satisfy
the specific bandwidth, gain, and pattern characteristics, these
miniature printed slot antennas with dual-band characteristics
clearly show potentials for future mobile wireless applications.

2 MATHEMATICAL ANALYSIS
2.1MUTUAL IMPEDANCE OF TWO PARALLEL LINEAR
ANTENNAS
The mutual impedance of two coupled circuits is defined in circuit
theory as the negative of the ratio V21 induced in the output of the
circuit to the current I1 flowing in input of the circuit with output
open. Consider the circuit shown in figure 6. The mutual
impedance Z21 is then
Z21= -V21/I1. …………………………(4)
Geometry of the proposed Hilbert slot antenna
Parameters: substrate thickness=h, dielectric
constant=
r



Fig-2.1 COUPLED CIRCUIT OF TRANSFORMER

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


Fig-2.2 PARALLEL COUPLED ANTENNAS
Now instead of the coupled circuit of figure 2.1 Let us consider the
case of 2-coupled antennas 1 and 2 as shown in fig. 2.2 Suppose a
current I1 in antenna 1 induces an emf of V21 at the open terminals
of antenna 2. Thus
Z21= -V21/I1
If the generator moved to the terminals of antenna 2 then by
reciprocity theorem the mutual impedance –
V21/I1= Z21=Z12=-V12/I2 (5)
The emf V21 induced at the open terminals of antenna 2 by the
current in antenna 1 is given byV21=(1/I2) 0
L
I2E21 dz.
Where I2 is max. Current and Iz is the value at a distance z from
lower end of antenna 2 with its terminal closed and where E21 is
the electric field along antenna 2 produced by antenna 1.
Assuming that this current distribution is sinusoidal as given by Iz
=I2 sin z.
So, V21= 0
L
E21 sin z. dz Then
Z21= -V21/I1= -1/I1 0
L
E21 sin z. dz (6)
fig. 2.3Geometry of the Hilbert Curves for the first four Iterations


This is the general expression for mutual impedance of two thin
linear, parallel, center-fed antennas with sinusoidal current
distribution as shown in fig. 2.2.
2.3 PROPERTIES OF FRACTAL PATTERNS

The Fractal patterns consist of identical or similar elements
repeated in different magnifications, orientations and positions
most often in an interconnected passion to obtain the final
structure.
Once optimize for radiation characteristics these Antennas can
find many applications in UHF/VHF communications. These
Antennas resonate at frequencies in a near logarithmic interval.
The individual bands at these resonant frequencies are generally
small. However their relative positions can be controlled, by
perturbing the shapes of the Fractal geometry of the Antenna
configuration. The important features of these Fractal Antennas
include low profile, the possibility of multiple frequency band and
moderated gain.
2.4 FRACTAL HILBERT CURVES
If we put four copies of previous iteration, connected to additional
line segments, we get geometry of Hilbert Antennas. e.g. geometry
of order one (arranged in different orientations), connected to the
additional segments shown with dashed lines in figure no. 2.3
It would be interesting to identify the Fractal properties of this
geometry.
The plane filling nature is evident by comparing the first few
iteration of geometry shown in figure no. 2.3
It should be noted that geometry is not strictly self-similar since
additional connection segments are required. When an extra
iteration order is added to an existing one. But the contribution of
this additional length (shown by dashed lines) is small compared to
overall length of geometry, especially when the order the iteration
is large.
Hence, the small additional length can be disregarded which
makes the geometry self-similar.
2.5 HILBERT GEOMETRY
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

There are many Hilbert curves for first little iteration as shown in
figure no. 2.3 Let us consider side by side dimension of Hilbert
Antenna is ‘L’ and order ‘n’.
Fig.2.4 Axial length of the proposed Hilbert slot antenna

The length of each line segment ‘D’ is given by
D=L0+L1+L2+--------------------Ln = L/(2n-1)
(Geometrical Progression)…(1)
And the sum of all the line segments
S=(2
2n
-1) D=(2n+1) L …(2)
It should be noted that the resonant frequency is decreased as the
iteration order is increased.
2.6 DIMENSION OF HILBERT GEOMETRY
According to multiple copy algorithms the dimension D is defined
as
D=log (N)/log (1/f). …………(3)
Where N is the number of copies and f is the scale factor of
consecutive iterations. The dimension of the Hilbert curve is
D=log [(4n-1)/(4n-1-1)]/ log [(2n-1)/(2n-1-1)]
= (For large n) log (4n/(4n-1)/ log (2n)/(2n-1)
=Log4/log2
=2.
The similarity dimension of this curve approaches to an integer
value (2) because of the approximation involved when a large
Fractal order is considered.
2.7 MUTUAL COUPLING EFFECTS IN HILBERT
ANTENNAS
The major problem in Hilbert Antennas is mutual coupling
between two such Antennas. Two identical Hilbert antennas can
be located in following configurations (as shown in figure no.2.5)
1 Collinear
2 Side by side
3 Stacked (Echelon)
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO306-6



Fig2.5 Three Arrangements Of Two Parallel Elements Of Hilbert
Antenna





Fig. 2.6 Three Dimentional View Of Hilbert Antennas
In, Vertical Position, Horizontal Position, Rotated Position
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

2.8 MUTUAL IMPEDANCE OF PARALLEL SIDE BY SIDE
ELEMENTS OF HILBERT ANTENNA (MATHEMATICAL
ANALYSIS)
Let d be separation of antennas and L be the length of elements
(as shown in fig. 2.7)




Fig- 2.7 Parallel Coupled Antennas With Dimensions
By diagram,
r1= d
2
+z
2
and r2= d
2
+(L-z)
2

The mutual impedance then becomes,
Z21= R21+j X21=j30 0
L
[{exp (-j d
2
+z
2
)}/( d
2
+z
2
)+ {exp (-j d
2
+(L-
z)
2
}/( d
2
+(L-z)
2
)]Sin z. dz
=30[2 Ei(-j d)- Ei(-j ( d
2
+L
2
+L)- Ei(-j ( d
2
+L
2
-L)]
Where the exponential integral is given by,
Ei(+-jy)= Ci(y)+-jSi(y)
Thus the mutual resistance is given by,
R21=30[2Ci( d)-Ci { ( d
2
+L
2
+L)}- Ci { ( d
2
+L
2
-L)}] (7)
and
X21=-30[2Si( d)-Si { ( d
2
+L
2
+L)}- Si { ( d
2
+L
2
-L)}] (8)
IT should be noted thatZ21=Z12=R21+jX21=R12+jX12
Where L=N /2 For N Odd
The mutual resistance and reactance decreases with distance
between elements of Hilbert antenna. The exact expression where
the antenna length L is not restricted to an odd number of /2, the
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

mutual resistance and reactance are given by BROWN and KING as
[5]
R21=30[2(2+Cos L) Ci d-4Cos2 ( L/2){Ci ( 4d
2
+L
2
- L)/2+
Ci ( 4d
2
+L
2
+ L)/2}+ Cos L {Ci ( d
2
+L
2
- L) + Ci ( d
2
+L
2
+
L)}+Sin L {Si ( d
2
+L
2
+ L)- Si ( d
2
+L
2
-L)-2 Si ( 4d
2
+L
2
+ L)/2+
2Si ( 4d
2
+L
2
- L)/2}]/Sin2 ( L)/2 . (9)
And
X21=30[-2(2+Cos L) Si d-4Cos2 ( L/2){Si ( 4d
2
+L
2
-
L)/2+Si ( 4d
2
+L
2
+ L)/2}+ Cos L {Si ( d
2
+L
2
- L) + Si ( d
2
+L
2
+
L)}+Sin L {Ci ( d
2
+L
2
+ L)- Ci ( d
2
+L
2
-L)-2 Ci ( 4d
2
+L
2
+
L)/2+ 2Ci ( 4d
2
+L
2
- L)/2}]/Sin2 ( L)/2 .
(10)
In the special case of L=N /2 for N is Odd the above expressions
reduce to relations given previously i.e. equation 7and 8
2.9 MUTUAL IMPEDANCE OF PARALLEL COLLINEAR
ELEMENTS OF HILBERT ANTENNA
According to Figure 2.5(b),
Spacing s=h-L, CARTER gives the mutual resistance and reactance
as
R21 = -15Cos h [-2Ci2 h +Ci2 (h -L) + Ci2 (h +L) -ln {(h
2
-L
2
)/ h
2
}]+15Sin( h)[2Si2 h -Si2 (h -L) - Si2 (h +L)] .
…(11)
X21=15Sin h [2Ci2 h -Ci2 (h -L) - Ci2 (h +L) -ln {(h
2
-L
2
)/ h
2
}]-
15Cos( h)[2Si2 h -Si2 (h -L) - Si2 (h +L)] .
…(12)
Curves for R21 & X21 of parallel collinear /2 antennas and L= /2
are presented as function of spacing s. It has been proved that, the
mutual resistance and reactance decreases with distance between
elements of Hilbert antenna.
32..10 MUTUAL IMPEDANCE OF PARALLEL ELEMENTS OF
HILBERT ANTENNA IN ECHELON/STACKED FORM
For this is let us suppose each element is an odd number of /2
long. CARTER and KING give the mutual resistance andreactanceas



Figure 2.8c
R21 = -15Cos h [-2CiA-2CiA’+CiB+CiB’+CiC+CiC’+ +15Sin h [+2SiA-2SiA’-SiB+SiB’-
SiC+ SiC’+ . …(13)
X21=-15Cos h *+2SiA+2SiA’-SiB-SiB’- SiC- SiC’ ++15Sin h[ 2CiA-2CiA’-
CiB+CiB’-CiC+CiC’+ …..(14)
Where,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

A= [( d
2
+h
2
) +h]
A’= [( d
2
+h
2
) -h]
B= [{ d
2
+(h-L)
2
} +(h-L)]
B’= [{ d
2
+(h-L)
2
} -(h-L)]
C= [{ d
2
+(h+L)
2
} +(h-L)]
C’= [{ d
2
+(h+L)
2
}-(h-L)]
It has been proved that, the mutual resistance and reactance
decreases with distance between elements of Hilbert antenna.
The stacked or Echelon arrangement is most general solution of which
the side-by-side position and the collinear position are the special
cases and it is highly dependent on orientation.
2.11 MUTUAL IMPEDANCE OF ELEMENTS OF HILBERT
ANTENNA IN INCLINED FORM
Referring to the shown fig.2.9 Consider the length of the element is L
and L’ separated by a distance R with orientation angles and ’ as
indicated. Let us assume L <<1, L ’<<1 and r >>1. The mutual
impedance Z21 is given by

Fig-2.9 Two Elements Of Hilbert Antenna In Oriented Form
Z21=V21/I1=60 L /r sin exp (-j r) L’ sin ’
…………(15)
Z21=(Magnitude)(Orientation)(Periodic function)
Z21=(60 L L ’/r ) (sin sin ’)(sin 2 r +jcos 2 r )
With maximum value
Z21 (max)= 60 L L ’/r
………………(16)
Hence, in the inclined form it has been proved that, the mutual
impedance depends on three factors: -
(1) Magnitude
(2) Orientation
(3) Periodic function
2.12 MUTUAL IMPEDANCE IN TERMS OF DIRECTIVITY AND
RADIATION RESISTANCE
Consider a transmitting antenna and a receiving antenna separated by
a distance r. The power delivered to the receiving antenna load under
matched conditions is
Pr=(Vr/2)
2
/Rr (W) ………..(17)
Where
Vr=Voltage induced at terminals of receiving antenna,
Rr=Radiation resistance of receiving antenna
The power transmitted is given by
Pt=It
2
Rt (W) .
……………… (18)
Where
It=Current at terminals of transmitting antenna
Rt=Radiation resistance of transmitting antenna
From (1) & (2) the magnitude of mutual impedance of antenna is
Mod Zm=Vr/It=2 (PrRrRt)/Pt. …………(19)
From the FRIIS transmission formula
Pr/Pt=ArAt/r
2 2
……………….(20)
Where
Ar=Effective aperture of receiving antenna
At= Effective aperture of transmitting antenna
From 19
Zm (max)=2 (ArAtRrRt/r
2 2
)
……………(21)
Substituting, directivity D=4 A/
2
The maximum mutual impedance can be expressed as
-------------(22)
Mod Zm (max)= ( DrDtRrRt)/2 r
Dr =Directivity of receiving antenna
Dt = Directivity of transmitting antenna
Rr = Radiation resistance of receiving antenna
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

Rt = Radiation resistance of transmitting antenna
r = Separation in wave lengths
Fig. 2.10 Configurations Of Different Hilbert Antennas
Thus, mutual impedance of two antennas is a function of their
directivities, aperture and radiation resistance. It has been assumpted
as r>> .
3 SIMULATION AND RESULTS
The expression discussed in the previous articles
for the calculation of mutual resistance and reactances have been
simulated by MATLAB software. The results obtained have been
attached in the succeeding papers.
On the basis of the results obtained it has been proved that at
closure spacing of Hilbert antennae the mutual resistance and
reactance increases rapidly. This yields to higher gain and higher
efficiency of Hilbert antenna. It has been also found that mutual
resistance and reactance is maximum in the case of side-by-side
arrangement.
4. CONCLUSION AND DISCUSSION
On the basis of analytical studies and simulation results the effects of
spacing of Hilbert antennae in different configurations have been
studied numerically using MATLAB software. It has been found that
mutual resistance and reactance increases for closure spacing of
Hilbert antennae. It has been proved that at larger spacing of the
elements of Hilbert antenna the total radiation resistance and
reactance decreases rapidly. However, the stacked, echelon form the
radiation resistance and reactance is highly dependent on orientation
The important advantage of Hilbert antenna is incorporation of its
plane filling characteristics to realize resonant antenna with smaller
overall physical size.
This antenna is simple & self -similar, there curves have the additional
property of approximately filling a plane. These properties are
exploited in realizing a small resonant antenna. This approach has
resulted in an antenna size smaller than /10 and still resonant with
performance comparable to a dipole whose resonant length is close to
/2.
With wide spread proliferation of telecommunication technology in
recent years, the need of small size multi band antennas has increased
many fold. Once optimize for radiation characteristics Hilbert antennas
can find many applications in UHF/VHF communication antennas,
which is most eagerly wanted technology in this telecommunication
era.

Figure 3
By the invention of monolithic IC fabrication technology these
antennae can be design in more compact form on single substrate.
This research will provide guidelines to new researchers for further
research in area of fabrication and design of many Hilbert antennae on
single substrate. This will yield to reduction in area, which is the basic
need of wireless and microwave communication systems. (As shown in
figure 3)
8.Future aspects
VLSI technology is boon to our electronics
technology.Radomes are used in spacecrafts and aircrafts.
In India Department of Defence is developling a light
combat craft (lcc)since 1980. Till now Dep’t. Has
developed three high combat crafts. The main issue is of
lesser weight of crafts. There exists developments of
RS422 interface which use to convert the parameters of
outer environment in to different voltage level (like
temperature, humidity etc.)The different levels are
voltage, resistance temperature detector, thermocouple,
temperature pressure scanner, voice, synchronous angle,
fine events etc. There exist a technology by which
resistance of a cut can be decreased. As we know the
Universal truth of resistance ,that it use to to take
maximum area in IC fabrication technology which is the
main motto of LCC viz. TEJAS. Hence, for LCC the
selected radome and antennas should be such that , it has
lesser weight & highly strong ,so that It can become
aerodynamically stable. By the invention of monolithic
I.C fabrication technology i.e V.L.S.I this antenna can
find many application applications in VHF/UHF
communication . A lot of Hilbert antennas can be
designed on single substrate. This yields to a reduction of
size which is the basic need of wireless technology.

5.REFERENCES
1 ZHU JINHUI, HOORFAR AHMAD, E.NADER
“Bandwidth, cross polarization, feed point
characteristics of matched Hilbert antennas”.
2 SAYEM.T.M and ALI. M. “Characteristics of micro
strip fed miniature printed Hilbert slot antenna”.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

3 VINAY.J.K, JOSE. K.A., VARADAN. V.K and
VARADAN. V.V “Hilbert curve Fractal antenna “.
4 4 GONZALEZ-ARBESU J.M., BLANCH S. and
ROMEU J. ‘The Hilbert curve as small self
resonant monopole from practical point of view”.
5 J.C. EDWARD &B.G. KEITH, “Electromagnetic waves
and Radiating systems” PHI Second Edition.
6 BROWN G.H. and R. KING, “High frequency models
in antenna investigation”.
7 CARTER P.S. “Circuit relations in radiating systems
and applications to antenna problems”.
8 KING H.E., “Mutual impedance of unequal length
antennas in Echelon
91Raghavan V “Material science and Engg.” 3
rd

edition PHI ND 1993
10-Skolnic I Merrill “Introduction to radar system”
3
rd
edition TMH.
11-J. Decker “Engineering Materials”TMH
12-S.Reuven “Scattering effect of seams on
sandwich radome performance”
13- Chassell Rick “The effect of rain on microwave
landing system antenna radomes ” IEEE symposium
on antenna technology & applied electromagnetic
Winnipeg Canada,Aug 15-17,1990
14-Boumens Morcel and Wagner Ulrike ,”spherical
near field radome test facility for nose mounted
radomes of commercial traffic aircraft”Orbit /Fr-
Europe Gmblt Johann –Seb astion –batch ,Str
11,85591 Vaters stetten ,Germany
15-Griffith lances “A fundamental and technical
review of radome“, May 2008
16-“Electric field on Earth “The physics Fact book
edited by Glenn Elert and Atmospheric electricity
by ELAT.Institute National de Pesqesus Spacious
June 2005.
17-Some related websiteswww.sciam.com,
www.hypertextbook.com,
www.mpdigest.com/issues/articals/may2008
18. Chan. K. K, Chang.P.R & Hsu.F, “Radome design
by simulated annealing technique” IEEE Trans
07803-07305/92.1992
19. Nie.X-C, Yuan.N, Li.W.L, Yeo.T.S, Gan.Y.B. “Fast
analysis of electromagnetic transmission through
arbitrarily shaped airborne radome uses
precorrcted FET method”, process in
electromagnetic research, PIER 54, 37-59, 2005.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011





X=spacings
Y=Ohms

Graph-1
MUTUAL RESISTANCE OF TWO PARALLEL SIDE-BY-SIDE ELEMENTS OF HILBERT ANTENNAS
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO306-13




X=spacings
Y=Ohms

GRAPH-2
MUTUAL REACTANCE OF TWO PARALLEL SIDE-BY-SIDE ELEMENTS OF HILBERT
ANTENNAE
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO306-14


X=spacing=Ohms

Graph 3
MUTUAL RESISTANCE OF TWO PARALLEL CO-LINEAR ELEMENTS OF
HILBERT ANTENNAE









CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO306-15

GRAPH-4
MUTUAL RESISTANCE OF TWO PARALLEL CO-LINEAR ELEMENTS OF HILBERT ANTENNAS

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO401-1

Error modelling of BFSK over Rayleigh Fading
Channel – A Statistical Frame Work
Vicky Singh
1
, Amit Sehgal
2
, Rajeev Agrawal
3

1,2,3
ECE Department, G.L. Bajaj Institute of Technology and Management, Gr. Noida

1
vicky.glbitm@gmail.com,
2
amitsehgal09@gmail.com,
3
rajkecd@gmail.com

Abstract- Frequency flat, fast Rayleigh fading may be
considered the most critical disturbance in a wireless
communication system. In its most general form, it is modeled
as a multiplicative time continuous random (zero mean
complex Gaussian) distortion of the transmitted signal. In
order to obtain a reliable wireless link for time continuous
Rayleigh fading channel, each part of the system must be
carefully designed. The special properties of the channel are
used as starting points in the design. Over the last 10 years a
tremendous amount of work has been presented about
communication over the Rayleigh fading channel. It seems
that much of the research has used techniques developed for
Additive White Gaussian Noise (AWGN) channel as the
starting point. In several cases this has put unnecessary
restrictions on the system design and has led to poor
performances. This means that less works use the Rayleigh
fading channel as the starting point. In this paper, we present
a statistical frame work for estimation and calculation of
average bit error rate (BER) of binary frequency shift keying
(BFSK) under Rayleigh fading channels. A detailed approach
to obtain BER expressions for BFSK in presence of AWGN
channel is available in several published material. A little or
no material is available for fading channels from the
perspective of undergraduate students. The statistical
framework presented is a step by step procedure to derive the
expressions for BER. Two different approaches have been
investigated in this paper to obtain these expressions. The
expressions thus obtained have been used for comparative
study of BPSK and BFSK in terms of BER through the
simulation. Following the same approach BER of other shift
keying techniques can be easily derived and investigated.
Keywords —Gaussian Q function, Error function, MGF, Bit
Error Rate, Rayleigh Fading, BFSK.

I. INTRODUCTION
Regardless of the branch of science or engineering,
theoreticians have always been enamored with the
notion of expressing their results in the form of
closed-form expressions. But the elegance of closed-
form solution is often neglected. First due to
complexity of its form and the second difficulty in
computing it numerically. Under such a scenario,
one becomes motivated to search instead for a
solution that is simple in form and also simple to
evaluate. Another cause for this motivation is that
the methods used to derive these alternative simple
forms must also be applicable to situations where
the closed-form solutions are not easily obtainable.
There are several performance parameters of a
digital communication system like SNR, amount of
fading, average outage duration, outage probability,
packet error rate etc. but even till date BER is
assumed to be one of the most important and critical
parameter for analysis of communication systems. It
is the one that is most revealing about the behavior
of a communication system. The search for and
ability to find a statistical frame work for problems
dealing with the evaluation of BER over generalized
fading channels motivated us to write this paper.

Two different approaches have been discussed in
this paper to obtain mathematical model of BER
estimation and calculation for BFSK under Rayleigh
fading channel. In the first method error function is
used while the second approach is based on Moment
Generating Function (MGF) using Gaussian Q
function. Starting from the basics of the Rayleigh
fading we develop the final expression in a lucid
manner so that undergraduate students may get an
exposure to the mathematics behind the expressions
of the BER over fading channels.
A similar work for BPSK has been presented by the
authors of this paper and is available in [4].

II. ERROR FUNCTION BASED APPROACH

In a multipath channel, received signal consists of a
large number of plane waves whose complex low
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO401-2

pass signal can be modeled as circular symmetric
Gaussian random variable and is of the form,

re im
j | | | = + (1)
If none of the multipath component is dominant,
re
| and
im
| are Gaussian processes of 0 mean
independent and identically distributed Gaussian
random variable of 0 mean and variance
2
o .
The magnitude | is a Rayleigh random variable
which has a probability density given as [6, page no.
524].

2
2 2
( ) exp
2
p
| |
|
o o
| | ÷
=
|
\ .
(2)
This represents the Rayleigh fading channel model
which mitigates the signal propagation in Non Line
of Sight (NLOS) channel suffering from multipath
fading.
When Rayleigh fading is present, the received
carrier amplitude is attenuated by the fading
amplitude | , which is a random variable with mean
square value
2
| and probability density function
(PDF) dependent on the nature of the fading
channel.
Now the received instantaneous signal power is
attenuated by
0 b
E N and therefore instantaneous
Signal to Noise Ratio (SNR) is given as [6, page no.
527]

2
0 b
E N ¸ | A (3)
Average SNR is given as

2
0 b
E N ¸ | A (4)
Due to such a condition imposed by fading, Bit
Error Probability (BEP) of any modulation scheme
is obtained by replacing
0 b
E N by ¸ in the
expression for AWGN performance. This is known
as conditional BEP and it is denoted by ( ; )
b
p E ¸ .
The BER of BFSK over AWGN channel is given by
[6, page no. 529] as

0
1
( ) ( )
2 2
b
b
E
p E erfc
N
= (5)
However in the presence of Rayleigh fading
amplitude | , the effective bit energy to noise ratio
is
2
0
b
E
N
|
. So the conditional BEP for given value of
| is
2
0
1 1
( ; ) ( ) ( )
2 2 2 2
b
E
p b erfc erfc
N
|
¸
¸ = = (6)
Where
2
0
b
E
N
|
¸ =
To find the error probability over all random values
of
2
| , we have to evaluate the conditional PDF
( ; ) p E ¸ over the PDF of¸ .
For Rayleigh fading, | is a Rayleigh distributed
random variable, therefore
2
| is a chi-square
distributed with two degrees of freedom. Since
2
| is chi-square distributed, ¸ is also chi-square
distributed. The PDF of ¸ is given as

1
( ) p e
¸
¸
¸
¸
¸
÷
=

0 ¸ > (7)
Now to find the BEP over Rayleigh fading channel,
we have to evaluate conditional BEP over PDF of
¸ [page no252, 7]

0
( ) ( ; ) ( )
b
p E P b p d
¸
¸ ¸ ¸
·
=
í
(8)
Or
0
1
( ) ( )
2 2
b
p E erfc e d
¸
¸
¸
¸
¸
·
÷
=
í
(9)
Consider

1
2
I erfc e d
¸
¸
¸
¸
÷
| |
=
|
|
\ .
í
(10)
Now putting
2
x
¸
= and b ¸ = (11)
eq. (10) can be written as

2
1
2 ( )
x
b
I erfc x e dx
÷
=
í
(12)
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO401-3

Or
1 2
2 I I = (13)
Where
2
2
( )
x
b
I erfc x e dx
÷
=
í
(14)

2
3
( )
2 2
x
b
b b
erfc x e I
t
÷
÷ | |
= ÷
|
\ .
(15)
Where
2
1
2
3
b
x
b
I e x dx
+ | | ÷
÷
|
\ .
=
í
(16)
Now making the substitution
2 b
x y
b
+ | |
=
|
\ .
, eq. (16)
can be computed as

1
2
3
2
y
b
I e y dy
b
÷
÷
=
+
í
(17)
At this stage we make use of a pre-established result
given as

1
2
( )( )
y
e y dy erfc y t
÷
÷
=
í
(18)
Now substituting eq. (18) in eq. (17), we have

3
( ) ( )
2
b
I erfc y
b
t =
+
(19)

( 2)
( )
2
b b x
erfc
b b
t
| |
+
=
|
|
+
\ .
(20)

Putting eq. (20) in eq. (15), we have
0
1 2
( ) [ ( )
2 2
b
p E erfc y e erfc
¸
¸
¸ ¸
¸ ¸
¸ ¸ ¸
÷
·
| |
+
| = ÷
|
+
\ .
(21)

From eq. (13) and making us of eq. (21) we have

1
2
2 2 2
I erfc e erfc
¸
¸
¸ ¸ ¸ ¸
¸ ¸
¸ ¸
÷
| |
| |
+
| = ÷ ÷
|
|
|
+
\ .
\ .


(22)
From eq. (9) and making use of eq. (22), the final
expression is
0
1 2
[ (
2 2 2
( ) )
b
p erfc e erfc E y
¸
¸
¸ ¸ ¸
¸ ¸
¸ ¸ ¸
÷
·
+
= ÷
+
| |
|
|
\ .

(23)


This finally results into desired result as shown
below

1
( ) 1
2 2
b
P E
¸
¸

= ÷
+


(24)

III. MGF BASED APPROACH


The BEP of BFSK over AWGN in terms of
Gaussian Q function is given by [7, page 229]

0
( ) ( )
b
b
E
p E Q
N
= (25)
Therefore conditional BEP can be given as
( ; ) ( ) p E Q ¸ ¸ = (26)
When characterizing the performance of coherent
digital communication, the generic form of
expression for BEP involves Gaussian Q function
with an argument proportional to square root of
instantaneous SNR of received signal. In case of
communication over a slowly varying fading
channel, the instantaneous SNR per bit is a time-
varying random variable with a PDF
( ) P
¸
¸ .
To compute the average BEP, we must evaluate an
integral consisting of above mentioned Gaussian Q
function and fading PDF, that is

0
( ) ( ) I Q a P d
¸
¸ ¸ ¸
·
=
í
(27)
Where a is a constant which depends on the specific
modulation or detection scheme.
Now using classical definition of Gaussian Q
function in eq. (27), we have

2
2
2
0 0
1
exp( )( ) ( )( )
2sin
a
I d P d
t
¸
¸
u ¸ ¸
t u
·
÷
=
í í
(28)
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO401-4


2
2
2
0 0
1
[ exp( ) ( ) ]
2sin
a
P d d
t
¸
¸
¸ ¸ u
t u
·
÷
=
í í
(29)
Now the inner integral is in the form of Laplace
Transform with respect to variable¸ . Since Moment
Generating Function (MGF) of ¸ is

0
( ) ( )
s
M s e P d
¸
¸ ¸
¸ ¸
·
=
í
(30)
From eq. (29) and eq. (30), we have

2
2
2
0
1
( )
2sin
a
I M d
t
¸
u
t u
÷
=
í
(31)
The Laplace Transform of Rayleigh PDF is given as
[7, pp. 125]

1
( )
1
M s
s
¸
¸
÷ =
+
, 0 s >

(32)

Now putting eq. (32) into eq. (31), we have

2
2
1
2
0
(1 )
2sin
a
I d
t
¸
u
u
÷
= +
í
(33)
The above integral reduces to following form [7,
page 125]
Or
2
2
1
2
(1 )
2
1
2
a
I
a
¸
¸
= ÷
+

(34)
Now comparing eq. (8) and eq. (27) and making use
of eq. (34), the final expression of bit error
probability is given as below which is similar to the
result obtained by earlier discussed approach.

1
( ) (1 )
2 1
b
p E
¸
¸
= ÷
+
(35)

IV. RESULTS

The results obtained above have been simulated
using MATLAB and following graph is obtained
which compares the performance of BFSK over
AWGN channel and Rayleigh fading channel.
0 5 10 15 20 25 30 35
10
-4
10
-3
10
-2
10
-1
Eb/No, dB
B
i
t

E
r
r
o
r

R
a
t
e
BER for BFSK modulation in Rayleigh channel
AWGN-Theory
Rayleigh-Theory
Rayleigh-Simulation

[Figure 1: BER of BPSK for Rayleigh fading
channel]

When the performance of BFSK over AWGN
[Fig.1] is compared with the performance over
Rayleigh fading channel, it is found that about
22.5dB degradation due to multipath channel
occurs.
The following graph compares the performance of
BPSK over AWGN and Rayleigh fading channel
which the authors obtained in [1] using MATLAB.
From the graph [Fig.2] it is clear that about 26dB
degradation occurs in case of BPSK due to
multipath channel.

0 5 10 15 20 25 30 35
10
-4
10
-3
10
-2
10
-1
Eb/No, dB
B
i
t

E
r
r
o
r

R
a
t
e
BER for BPSK modulation in Rayleigh channel
AWGN-Theory
Rayleigh-Theory
Rayleigh-Simulation

[Figure 2: BER of BPSK for Rayleigh fading
channel]
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO401-5

Now having seen the performance of BFSK and
BPSK, it becomes clear that due to degradation over
multipath channel we have to spend so much energy
in order to get a reliable wireless link. This is not a
good situation because spending so much energy
will cause an increase n global warming which has
already become a threat to our existence.

V. CONCLUSIONS

We have briefly demonstrated in this paper two
approaches to evaluate the performance of BFSK
over Rayleigh fading channel. A sincere effort has
been done by the authors in explaining the
mathematics behind these two approaches which is
not easily available. Also the performance of BFSK
is compared with BPSK and it has been found that
in case of BFSK about 3.5dB less degradation occur
over multipath channel as compared to BPSK.

It is the authors hope that when the reader reach the
end of this paper, they will fully appreciate the
power behind the two approaches discussed and
will generate for themselves an insight into finding
new and exiting applications.

REFERENCES

[1] G. D. Forney, L. Brown, M. V. Eyuboglu, J. L. Moran III,
“The V.34 High-Speed Modem Standard”, IEEE Comm.
Mag., Vol. 34, No. 12, pp. 28-33, Dec.1996.

[2] A. Acampora, “Wireless ATM: A Perspective on Issues
and Prospects”, IEEE Pers. Comm. Mag., Vol. 3, No. 4, pp. 8-
17, Aug. 1996.

[3] Digital communication by Proakis fourth edition, chapter
5, page no. 257.

[4] Vicky Singh, Amit Sehgal “ Error modelling of BPSK
over Rayleigh fading channel- A Statistical Pedagogic
Approach”, National Conference on Advancement of
Electronics and Communication Engineering, IIMT college of
engineering, Gr.Noida, February 2011.

[5] M.K. Simon and M.S. Alouini, “Simplified noisy
reference loss evaluation for digital communication in the
presence of slow fading and carrier phase error,” IEEE Trans.
Vet. Tech., vol. 50, no. 2, pp. 480-486, Mar. 2001] “Digital
Communication over fading channels”, alouini.

[6] Digital modulation techniques, fuqin xiong, 2000 Artech
house, London.

[7] “Digital Communication over Fading Channels” second
edition 2005 by M.K. Simon and M.S.Alouini.

[8] B. Sklar, “Rayleigh Fading Channels in Mobile Digital
Communication Systems”, Part II: Mitigation”, IEEE Comm.
Mag., Vol. 35, No. 7, pp. 102-109, July 1997.

[9] K. Brayer, Ed., Data Communications via Fading
Channels, IEEE Press, 1975.

[10] D. C. Cox, “Wireless Personal Communications: What Is
It?” IEEE Pers Comm. Mag., pp. 20-35, April 1995.

[11] J.K. Cavers, “On the Validity of the Slow and Moderate
Fading Models for Matched Filter Detection of Rayleigh
Fading Signals”, Can. J. of Elect. &
Comp. Eng., Vol. 17, No. 4, pp. 183-189, 1992.

[12] W.F. Walker, “The Error Performance of a Class of
Binary Communications Systems in Fading and Noise”, IEEE
Trans. Comm. Systems, pp. 28-45, March 1964.

[13] Q. Dai, E. Schwedyk, “Detection of Bandlimited Signals
Over Frequency Selective Rayleigh Fading Channels”, IEEE
Trans. Commun., Vol. 42, No. 2/3/4, pp. 941-950,
Feb./Mar./April 1994.

[14] X. Yu, S. Pasupathy, “Error Performance of Innovations-
Based MLSE for Rayleigh Fading Channels”, IEEE Trans.
Vehic. Techn. Vol. 45, No. 4, pp. 631-642, Nov. 1996.

[15] S. Sampei, T. Sunaga, “Rayleigh Fading Compensation
for QAM in Land Mobile Radio Communications”; IEEE
Trans. Vehic. Techn. Vol. 42, No. 2, pp. 137-147, May 1993.

[16] S. A. Al-Semari, T. E. Fuja, “I-Q TCM: Reliable
Communication Over the Rayleigh Fading Channel Close to
the Cutoff Rate”, IEEE Trans. Inform.
Theory, Vol. 43, No. 1, pp. 250-262, Jan. 1997.

[17] P. Y. Kam, H. M. Ching, “Sequence Estimation over the
Slow Nonselective Rayleigh Fading Channel with Diversity
Reception and its Application to Viterbi Decoding”, IEEE J.
Select. Areas Comm., Vol. 10, No. 3, pp.562-570, April 1992.

[18] A. D. Polyanin and A. V. Manzhirov, Handbook of
Integral Equations, Second Edition, Updated, Revised and
Extended, Chapman & Hall/CRC Press, 2008, 1144 pages.






CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO401-6



CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO402-1

PERFORMANCE EVALUATION OF VARIOUS ERROR CORRECTING CODES
ON PLC SYSTEM
Priyank Sharma, Student M.Tech (Comm.),
Shobhit University,
Meerut (UP)
sh_priyank@yahoo.com

Ravikant Saini, Astt. Prof., (ECE)
Shobhit University,
Meerut (UP)
sravikantk 2@rediffmail.com
Abstract- In this paper, performance of DMT has
been studied for a PLC system. The power line has
been modeled as an access link channel. To observe the
effect of impulsive noise, a simulation model based on
statistical properties of impulsive noise has been used.
Performance has been observed with AWGN only and
AWGN and impulsive noise both with different error
correcting codes. System has been observed by 256 no
of sub carriers and Message Transfer Unit (MTU) size
280 bits.
Keywords Component-Power Line Channel, Access
Link Channel, DMT, MTU.

I. INTRODUCTION
The liberalization of telecommunications and the
deregulation of electricity utilities have added new
dimensions to the potential application of the electricity
infrastructure for the most efficient use of the local loop
[1]. The main focus is and will continue to be on the
connection between house and transformer as a solution for
the last dirty mile problem to compete with other last mile
technologies like digital subscriber line (DSL),wireless
local loop, or telephone lines. Furthermore, new interest
arises due to recent developments regarding in-house
networking.
Due to the fact that power lines were not conceived for
communication purposes but to deliver electrical energy
(which involve signals of much lower frequency and
higher power), the medium characteristics are certainly
adverse. The channel is time and frequency variant, and
exhibits remarkable differences between locations,
according to the network topology; kind of wire installed
and connected loads. Even in a specific power circuit the
conditions change depending on the selected transmission
path (i.e. sockets where the transmitter and receiver are
plugged)
Optimization of a transmission scheme can only be done
after an accurate channel model is made available.
Although some models have been proposed [2-6] ,but there
is no widely accepted model similar to mobile radio
channels. DMT as a multicarrier modulation is the best
technique for combating the frequency selective fading [7,
8]. Adaptivity in terms of bit loading is an added feature
against the time variation of the multipath channel. Strong
forward error correction strategy is required to overcome
the hostile noise environment, which includes background
noise, colored noise and a combination of impulsive noises
[9- 12].

II. CHANNEL MODELS
The modeling problem for PLC, which consists of
investigating the characteristics of the power network as a
communication channel , is a serious one .Power line grid
has tree like structure. Signal faces reflections at every
branching, empty socket and connected loads. Thus a
complex echo scenario arises leading to frequency
selective fading.


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO402-2

ACCESS LINK
The model for Access link is based on physical signal
propagation effects in mains networks including numerous
branches and impedance mismatching. Besides multipath
propagation accompanied by frequency selective fading, signal
attenuation of typical power cables increasing with length and
frequency is considered [5]. Model describes the transfer function
of the channel in terms of very few parameters. The transfer
function

Fig(1).Basic DMT transceiver structure

p
fd j d f a a
N
i
i
v e e g f H
i i
k
/ . ) (
2
.
) . (
1
1 0
   





where N is the number of the relevant multipaths,a0 and a1
are attenuation parameters and k is an exponent taking
value in the range of 0.5 to 1; is the path length and vp is
the speed of EM wave in the material.
III DISCRETE MULTITONE
DMT efficiently divides an input serial data bit stream
into a large number of low-data-rate streams and assigns
them to a set of parallel and orthogonal subcarriers
according to the SNR of respective subcarriers [8]. SNR of
each subcarrier is estimated in an initial training session.
Fig 1 shows the block diagram representation of a basic
DMT transceiver.
The data bits assigned to each subcarrier are mapped onto
a Quadrature Amplitude Modulation (QAM) constellation
to form a complex sample, and then the resulting N
complex samples from the N subcarriers are extended to be
a 2N point complex –conjugate symmetric sequence. This
2Npoint complex sequence is transformed using inverse
fast Fourier transform (IFFT) to generate 2N point real
samples for transmission through the channel. In order to
mitigate ISI and maintain the continuity between
consecutive DMT symbols, the last 25% of each 2N-
sample block are circularly wrapped to prefix the block
itself. After receiving the transmitted signal, the receiver
discards the prefixed samples; then the remaining received
samples are demodulated into complex symbols by the 2N
point FFT. These complex symbols are decoded using a
QAM decoder. The in phase and quadrature components of
the complex symbols are used by the QAM decoder to
generate the received data per subcarrier. Finally a parallel
to serial conversion recovers the original transmitted data
bit stream. For the better results different algorithm may be
used[13-16]. The vital components of a DMT transceiver
system are the IFFT and FFT at the transmitter and
receiver , respectively.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO402-3

IV SIMULATION & RESULTS
The noisy environment in power line has been represented
as a combination of background noise and impulsive noise.
Background noise has been modeled as AWGN and a
model based on statistical properties of impulsive noise
has been used for impulsive noise.

Performance of DMT is evaluated by below methods:
1) 256 no of subcarriers & MTU size 280 bits.
2) 128 no of subcarriers & MTU size 112 bits.

Following five Error Correcting Codes have been studied:
Convolution, Hamming, BCH, CRC, & RS.

In order to study the performance degradation due to
impulsive noise, two noise environments have been
considered.
1. AWGN only
2. AWGN and Impulsive noise.
Performance of DMT for Access Link Channel
Fig.(2) & Fig(3) shows the Result 256_280 AWGN &
Impulsive noise

Fig.(4) & Fig(5) shows the Result 128_112 AWGN &
Impulsive noise




Figure (2) 256_280 With AWGN Only

Figure (3) 256_280 With AWGN & Impulsive Noise

Figure(4)128_112 With AWGN Only

Figure (5) 128_112 With AWGN & Impulsive Noise
V. CONCLUSION
Performance of DMT system has been evaluated for the
Access link. In impulsive noise analysis the error rate is
higher than the AWGN environment as expected. This is
because of the burst errors caused by Impulsive noise. In
order to improve the performance further higher error
correcting codes can be used.
REFERENCES:
[1] N. Pavlidou, A. J. H.Vinck, J. Yazdani, and B.Honary,
“Power line communications ”State of the art and future
trends,” IEEE Commun. Mag.,vol. 41, pp.34–40, Apr.
2003.
[2] O. Hooijen, “A channel model for the residential
power circuit used as a digital communications medium,”
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO402-4

IEEE Trans. Electromagn. Compat.,vol. 40, pp. 331–336,
1998.

[3] H. Meng, S. Chen, ”Modeling of Transfer
Characteristics for The Broadband Power Line
communication channel”, IEEE Transactions on
Power Delivery, Vol. 19, No 3, pp. 529-551. Jul. 2004.
[4]L.T.Tang,P.L.So,E.Gunawan,Y,L.Guan,S.Chenand
T.T.Lee,”Characterization and modeling of In-Buiding
Power Lines for High speed Data Transmission,”IEEE
Trans.Power Delivery,vol-18,No-1,pp.31-40,Jan.2003.
[5] M. Zimmermann and K. Dostert. "A Multipath Model
for the Powerline Channel,'' IEEE Trans. Commun..
vol.50. no. 4. Apr. 2002, pp. 553-59.
[6] A. Mori, Y. Watanabe and M. Tokuda & K.
Kawamoto “Transmission Characteristics of an OFDM
Signal for Power Line Communication System with High
Bit Rate”.IEEE Communication Magazine, July 2007.
[7] E. Biglieri. Coding and modulation for a horrible
channel. IEEE Communications Magazine, 41:92–98,
May 2003.
[8] S. Baig and N. D. Gohar. A discrete multitone
Transceiver at the heart of the PHY layer of an in-home
power line communication local area network. IEEE
Communications Magazine, 41:48–53, Apr. 2003.
[9] M. Zimmermann and K. Dostert. An analysis of the
broadband noise scenario in power line networks. In IEEE
International Symposium on Power Line Communications
and Its Applications ISPLC ’2000, pages 131–138, 2000.
[10] M.Zimmermann and K.Dostert, “Analysis and
modeling of impulsive noise in broadband powerline
communications,”IEEE Trans Electomagn. Compat.,
vol.44,No.1,pp.249-258,Feb.2002.
[11]R. M. Vines, M. J. Trussel, L. J. Gales, and J. B.
O’Neal, Jr., “Noise on residential power distribution
circuits,” IEEE Trans. Electromagn.Compat., vol.
EMC-26, pp. 161-168, Nov. 1984
[12] D. Anastasiadou and T. Antonakoupoulos,
”Multipath characterization of Power-Line Networks”,
IEEE Trans. On Power Delivery, vol. 20,pp. 90-99, 2005.
[13] E.Guerrini,G.Dell’ Amico,P.Bisagila, L.Guerrieri,”
bit-loading algorithm and SNR estimate for Homeplug
AV” IEEE ISPLC’07,pp419-424,26-28 March 2007.
[14]F.S.Muhammad, J.-Y.Baudais, J.-F.Helard and
M.Crussiere,” Coded adaptive linear precoded discrete
multitone over PLC Channel,” ISPLC’08,April 2008
[15] A.M.Wyglinski,F.Labeau and P.Kabal,”Bit loading
with BER constraint for multicarrier systems,” IEEE
Trans. On wireless comm.., vol.4,pp. 1383-1387,july
2005
[16] M. H. L. Chan and R.W. Donaldson, “Amplitude,
width, and interarrival distributions for noise impulses on
intrabuilding power line 3, pp. 320–323, Aug. 1989.





CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO421-1

Design and Analysis of Microstrip Antennas for W-LAN and Ku-band Applications

1
Ravindra Kumar Yadav and
2
Ram Lal Yadava
1
Department of Electronics and Communication Engineering,
I.T.S. Engineering College, Greater Noida, Uttar Pradesh, India
2
Department of Electronics and Communication Engineering,
Galgotia’s College of Engineering & Technology, Greater Noida, Uttar Pradesh
ravipusad@gmail.com

Abstract

This paper deals with the design analysis methods for microstrip antennas (MSA) meant for W-LAN and Ku
Frequency band applications. The design analysis methods for MSAs can be broadly classified into two
groups. In the first group, the methods are based on equivalent magnetic current distribution around the
patch edges; The Transmission Line Model and the MNM. The second group methods are based on the
electric current distribution on the ground plane (similar to dipole antennas), and the popular methods are
Method of Moments, Finite Element method, Spectral Domain Technique and Finite-difference Time
Domain method. In this work, both the techniques have been implemented to analyze the fabrication and
performance of the antenna. The Transmission Line Model and the Method of Moments has been used.
Three antennas structure have been considered namely; proximity and aperture coupled, and slot coupled
MSAs.
Keywords: Microstrip antenna, Ultra wideband coupled patch antenna and circularly polarized MSA.
I. INTRODUCTION

Since there has been an increase in the demand of
Microstrip Antennas, these are manufactured in a
large scale, but there are limitations with the design
of these antennas. This is because proper analysis is
required before fabrication and then simulation for
characteristics investigations [1-5]. Therefore
present paper aims to describe the design analysis of
Microstrip Antennas using the Transmission line
model and the Method of Moments and study the
performance characteristics as well. For this, IE3D
software is used, and since the transmission line
model is an approximate model, genetic algorithm
is also being implemented to optimize the design.
The Method of Moment model is used to determine
the current distributions and field distributions of
the antenna. An aperture coupled microstrip antenna
is chosen for the design using transmission line
model and a wire antenna is chosen for the Method
of Moments model [6-10].



II .THEORY

Microstrip antennas are used due to its
several advantages like small size, ease of
fabrication, low cost and many others. Polarization
is an important part of the study of microstrip
antennas. Usually the polarization of an antenna is
defined as the wave radiated by the antenna in a
particular direction. This is usually depends on the
feeding technique [4-5]. However in general,
orientations of the electric field vector of a plane
wave relative to direction of propagation define the
polarization of wave. If an electric field vector of an
EM wave is parallel to x-axis, wave is said to
linearly x-polarized. This wave could be generated
by a straight wire antenna parallel to x-axis.
Similarly y-polarized waves can also be defined and
generated. If two plane waves of equal amplitude
and orthogonally polarized are combined with
phase – difference, the resulting wave is circularly
polarized and electric field vector describe a circle
centered on propagation vector. The field vector
will rotate by for every wave length traveled.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO421-2

Circular polarization generated as either RHCP or
LHCP. Right hand circularly polarized wave
describes a wave with E-field vector rotating
clockwise when looking in the direction of
propagation [7]. Bandwidth: The band width of
antenna is usually defined as the frequency range
within the performance of the antenna with respect
to a certain characteristics. It is expressed as the
percentage of the difference between upper and
lower frequency to the center frequency;
% 100
o
u
f
f f
BW

.
Bandwidth of an antenna can also be defined in
terms of radiation patterns or VSWR/reflected
power as follows;
VSWR Q
VSWR
BW
1
.
Where, Q is a quality factor. Feeding techniques:
Popular feeding techniques are the coaxial probe
feed, microstrip line feed, aperture coupled and the
proximity fed antenna.

Types of antennas analyzed:

Aperture coupled Microstrip Antenna
consists of 2 layers of dielectric separated by
a ground plane. Feed is given in the lower
layer and a slot is made between the 2 layers
in the ground plane which couples the power
from lower dielectric to upper one (Fig 1).


Fig 1. Geometry of aperture coupled MSA.

Proximity fed Microstrip Antenna is shown
in the Fig 2. Proximity coupling offers some
opportunity to reduce feed line radiation
while maintaining a relatively thick rotation.
The main advantage of Circular Polarization
is that regardless of substrate for the
radiation patch. The input impedance of the
antenna is affected by the overlap of the
patch and the feed line, and by the substrate.
This feature adds degree of freedom in the
design, but may complicate the task of
selecting an optimum design.

Fig 2. Geometry of proximity coupled MSA.

3
rd
antenna analyzed is a rectangular slot
antenna for ultra wideband applications. It is
shown below (Figs 3 and 4).


Fig 3. Geometry of slot coupled MSA (Top view).



Fig 4. Geometry of slot coupled MSA (Side view).



CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO421-3

Brief introduction of methodologies used:

IE3D is simulation software used to
simulate various geometries of antennas.
It is commercially available software.
For design optimization process genetic
algorithms are used which work on the
theory of natural reproduction process.
For this optimization process a fitness
function is programmed using the
transmission line model of the geometry.
Method of Moments is another method
which gives idea about the current and
field distribution with respect to
normalized wavelength.

Design Specifications

Silicon feed substrate = 100 um
Silicon antenna substrate = 200 um
Cavity height = 150 um
Substrate thickness = 1.5748 mm
Substrate dielectric constant = 2.17

III. RESULTS

Results from the simulations

The simulation results for the proximity coupled
antenna are tabulated as;

Table 1. VSWR, Gain and resonance frequency (f
r
)

The antenna geometry was again designed with a
substrate provided and the results are;

Ls
mm
Ws
mm
VSWR Bandwidth
MHz
2.828

1.414 1.0275 223


Therefore there is increase in bandwidth. The
simulation results for the Rectangular slot antenna
(resonating at 2.44 GHz) with patch substrate for
ultra wideband applications and phased array
systems are shown below in table 2.

Table 2. VSWR, Gain and BW

W
1

mm
W
2

mm
L
1

mm
L
2

mm
L
3

mm
VSWR Gain BW
MHz
44 36 16 5

4 1.0382 3.191 690

The overall performance of the simulation software
(IE3D) for the aperture coupled microstrip antenna
(substrate thicknesses: 1.2 mm (lower), 1.5748 mm
(upper),) (dielectric constant: 1.9 lower, 3.9 upper is
listed in Table below in table 3.

Table 3. VSWR, Gain and resonance frequency (fr)

Patch
Length
mm
Patch
Width
mm
Slot
Length
mm
Stub
length
mm
VSWR Bandwidth
MHz
28.88

17.85

9.367

5.274

1.34 14

6
th
round results of Genetic Algorithm
optimization

Here substrate thickness = 1.27 mm
Substrate dielectric constant = 6.33
Bandwidth: = 690 MHz
6
th
round (all dimensions in meters)

L = 0.01 m to 0.032 m
W = 0.01 m to 0.029 m
La = 0.005 m to 0.013 m
Ls = 0.001 m to 0.009 m

IV. CONCLUSIONS

Both transmission line model and the
method of moment’s model have been
programmed. In case of transmission line
model, we observe that by varying the
length, width and other dimensions of the
patch, we can vary parameters like VSWR
and Frequency of operation.
In case of method of moment’s model we
can get the detailed explanation of current
and field distribution provided we have the
L
1
mm
L
2
mm
VSWR Gain
dB
BW
MHz
f
r
GHz
16.4

16.4

1.0275 1.95

200 15.6
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO421-4

Green’s function available for that antenna
structure.
Aperture coupled microstrip antenna has a
higher gain than the rectangular slot antenna
with patch stub but the latter has a higher
bandwidth. The proximity coupled antenna
showed very lee gain but a high bandwidth.
Genetic Algorithms are very useful to
implement the transmission line model when
we use the VSWR as the fitness function.

Table 4. Optimized results of VSWR.

Gen
erati
on
L
(patch
length)
W
(patch
width)
Ls
(aperture
length)
Ls
(stub
length)
VSWR
35 0.03126 0.01451 0.00433 0.00834 1.824
41 0.03131 0.01547 0.00411 0.0064 1.654
52 0.03197 0.01149 0.00574 0.00461 1.214
54 0.03131 0.01561 0.00412 0.00867 1.636
55 0.03137 0.01391 0.00453 0.00661 1.317
56 0.03137 0.01328 0.00456 0.00563 1.650
57 0.03142 0.01309 0.00461 0.00472 1.568
82 0.03161 0.01147 0.00556 0.00553 1.648
93 0.03181 0.01666 0.0041 0.00651 1.320
100 0.03128 0.01155 0.00507 0.00479 1.064

When we fabricated the antenna, GAs are
very useful to provide us with the
dimensions of the antenna. IE3D is
commercially available software that
enables us to simulate a particular antenna
geometry which gives us an idea about the
gain, bandwidth and other parameters of the
antenna which are vital for its performance
and use.

Therefore it is summarized the gain of
aperture coupled microstrip antenna is
found to be high; however the bandwidth
was very low. The proximity coupled
antenna showed a better bandwidth, which
can further be enhanced by providing a stub.
However the rectangular patch antenna
showed a very high bandwidth of 700MHz
but gain was less. The resonant frequencies
for the 3 antennas are found to be 2.4 GHz
for Aperture coupled and 15.5GHz
Proximity however rectangular patch shows
2.4 GHz. The optimized results of VSWR
vary from 1.064 to 1.824.

Acknowledgment

The authors are very thankful to Prof. K.V.
Karthikayen, IIT-R, for his motivation and
support to complete this research work in
time.
REFERENCES
[1] David M Pozar, “Microwave Engineering”, John
Willey & Sons Inc; 1998.
[2] Constantine A. Balanis, “Antenna Theory, Analysis
and Design”, John Willey & sons Inc; 2002.
[3] W.L. Stutzman, G.A. Thiele, “Antenna Theory and
Design,” John Wiley Sons, 1981.
[4] Girish Kumar and K.P. Ray, “Broad band Microstrip
Antenna”, Artech House, Inc, 2003.
[5] I.J.Bahl et al, “Microstrip Antennas”, Artech House
Inc; 1980.
[6] P. Mazumdar, Elizabeth M.Rudnick, “Genetic
Algorithms for VLSI design, Layout and test
automation”, Prentice Hall, PTR, 1999.
[7] R. A. Saeed and Sabira Khatun, “Design of
Microstrip Antenna for WLAN,” IEEE, Trans.,
Antennas and Propagate, Vol. AP-33, No.2, pp-131-
137, Feb 1987.
[8] D. M.Pozar,” A Review of Aperture Coupled
Microstrip Antennas: History, Operation,
Development, Applications”, IEEE, Trans.,Antennas
and Propagate, Vol. AP-40, No.2, pp-138-143, Feb
1992.
[9] P. K. Singhal and Laxmi Shrivastava, “On
Investigations of a wide band Proximity Fed Bow
Tie Shaped Microstrip Antenna”,
http//www.ctystems.com/zeland.
[10] http://en.wikipedia.org/woki/Gemetic_algorithm
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO422-1


Abstract— There are several Optical Orthogonal Codes are
commonly used as signature codes for Optical Code Division
Multiple Access (OCDMA) communication systems. In this paper
some of Optical Orthogonal Codes have been discussed with their
algorithm of generation and comparison by the MATLAB results
corresponding to auto-correlation and cross-correlation with a
new OOC coding scheme. These codes have been assigned to
users as a signature sequence and the maximum value of auto-
correlation and cross-correlation of the codes will decides the
Multiple Access Interference for the optical CDMA system.

Index Terms — OCDMA-Optical Code Division Multiple
Access, OOC -Optical Orthogonal Codes,
a
ì -auto-correlation
constraint,
c
ì -cross-correlation constraint.

I. INTRODUCTION
ptical fiber has a capability to transfer data information at
a very high speed. Multiple users can transfer
information with its huge bandwidth that is called by multiple
access. For optical domain multiple access is known by
OCDMA.
In OCDMA each user is distinguish by a specific optical
code sequences. An encoding operation optically transforms
each data bit before transmission. At receiver end decoding
operation is done to retrieve the original information. The data
information‟s each bit is multiplied by a code sequence that
may be either in one-dimensional time domain, wavelength
domain and two-dimensional time-wavelength domain [1].
The nature Optical Orthogonal Codes is either in unipolar
optical orthogonal codes or bipolar optical orthogonal codes.
The unipolar codes are formed by „1‟ and „0‟. If there is an
optical pulse it means weight „1‟ is coded and if there is no
pulse „0‟ is coded. The unipolar codes are used in an
incoherent optical CDMA system and they are formed in two
ways that are one dimensional and two dimensional OOCs.
One dimensional OOCs
An OOC is usually represented by quadruple
) , , , (
c a
w n ì ì

where n is the sequence code length, w is the
number of ones in a sequence code,
a
ì auto correlation
constraint and
c
ì is the cross-correlation constraint. These
codes have to satisfy two properties.
1) The Auto Correlation Property:
¯
÷
=
©
1
0
n
t
t t
x x
t
s
a
ì (1)
X = (x
0
,x
1
,x
2
,……..x
n-1
);
e code set C and any integer t , 0
<t < n
2) The Cross Correlation Property:
¯
÷
=
©
1
0
n
t
t t
y x
t
s
c
ì (2)
X = (x
0
,x
1
,x
2
,……..x
n-1
), Y = (y
0
,y
1
,y
2
,…..y
n-1
) ; X,Y
e code set
C and any integer t , 0 s t < n

All subscripts in eq. (1) and (2) are in modulo n addition.

If maximum of our auto-correlation and cross-correlation
constraints is ì then by Johnson bounds the calculation of
maximum number of codes.
max(
a
ì ,
c
ì ) = ì ;1 s ì sw-1.

By Johnson‟s bound [2] the number of optical orthogonal
codes generated by the given values of sequence length n,
weight w, and ì the maximum of autocorrelation and cross-
correlation constraint is
( )( ) ( )
( )( ) ( )
1 2 ............
( , , )
1 2 ............
A
n n n
J n w
w w w w
ì
ì
ì
| | ÷ ÷ ÷
=
|
|
÷ ÷ ÷
\ .

II. VARIOUS ONE DIMENSIONAL OOCS DESIGNING METHOD
A. Prime Sequence OOCs
Prime sequences are the code set which contains „1s‟ and
„0s‟ each one of the p prime sequences is taken on the basis
for a group of CDMA sequences where p is a prime number.
A group is generated by rotating the original prime sequence.
This method allows forming p groups of p codes resulting in a
total length of p
2
.
Galios Field is used to construct the prime sequence.
i.e. GF(p) = (0,1,2,….p-1)
so the code sequence in the form of bit positions
P
X
C ={ (0)
p
x
c , (1)
p
x
c , (2)
p
x
c …… ( 1)
p
x
c p ÷ }
where
( )
p
x
c j =x.j (mod (P)) for x, j eGF(p);
The binary code word
C
P
X
={ ) 0 (
p
xi
c , ) 1 (
p
xi
c , ) 2 (
p
xi
c ,….., ) 1 ( ÷ p c
p
xi
} with
A Proposed Design of One Dimensional Optical
Orthogonal Coding Scheme and Analysis with
Others
1
Ratnesh Kumar,
2
R.C.S. Chauhan,
3
G.P.Bagaria
1
Department of Electronics Engineering, H.B.T.I., Kanpur, U.P., India
O
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO422-2

¹
´
¦ + =
=
otherwise
p jp j d i
i c
p
x p
xi
0
)) (mod( ) ( 1
) (
2
[3]

where, j= (0,1,2,….p-1)
This scheme generates the set of optical orthogonal codes
) , , , (
c a
w n ì ì for any prime number p, such that weight w =
p, code length p
2
, auto-correlation constraint
a
ì = p-1 and
cross correlation constraint
c
ì = 2. The numbers of optical
orthogonal codes in the set are given by p.

Example for prime taken p=7
Galios Field, GF(p) = (0,1,2,3,4,5,6)
D following positions of „1s‟
}, 0000100000 0010000010 1000001000 0000100000 {100000000
}, 1000010000 0000010000 0001000000 0001000010 {100000000
}, 0010001000 0100000000 0000000100 0010001000 {100000000
}, 0000000100 0000100100 0100100000 0100000000 {100000000
}, 0000000010 1000000001 0000000010 1000000001 {100000000
}, 0100000001 0001000000 0000010000 0000000100 {100000001
}, 0001000000 0000001000 0010000001 0000010000 {100000010
1}, 2 3 4 5 6 {0 ) 6 (
2}, 4 6 1 3 5 {0 ) 5 (
3}, 6 2 5 1 4 {0 ) 4 (
4}, 1 5 2 6 3 {0 ) 3 (
5}, 3 1 6 4 2 {0 ) 2 (
6}, 5 4 3 2 1 {0 ) 1 (
0}, 0 0 0 0 0 {0 ) 0 (
7
6
7
5
7
4
7
3
7
2
7
1
7
0
7
7
7
7
7
7
7
=
=
=
=
=
=
=
=
=
=
=
=
=
=
x
x
x
x
x
x
x
x
x
x
x
x
x
x
C
C
C
C
C
C
C
d
d
d
d
d
d
d

Generated codes have n = 49, w = 7 = p,
a
ì = 6 = p-1,
c
ì = 2.
(
a
ì and
c
ì are the maximum values in a code-set).

B. Quasi prime OOCs
Quasi prime is the extension of the OOC code set based on
prime sequences. A quasi prime code
qp
xk
C is the time shifted
and extended (or contracted) versions of prime sequences
code
p
x
C . It is given as, with q number of „1‟s
) (i c
qp
xk
= ) ] ([
n
p
x
kp i c + ; where i = 0,1,….qp-1 [3][4]
each quasi-prime code then comprises qp chips taken
cyclically from a shifted prime code, and contains q ones.code
set ) , , , (
c a
w n ì ì code length n = qp, weight w = q
where, (r-1)p < q < rp,
q and r are positive integers.
auto-correlation constraint
a
ì = (p-1) r, cross-correlation
constraint
c
ì = 2r and the number of code-set N = p.

Example for p =5, q=4, k=3
Code-set contains the contracted version of prime codes are
00001}, 10000 01000 {00100
00010}, 10000 00100 {00001
00100}, 10000 00010 {01000
01000}, 10000 00001 {00010
10000}, 10000 10000 {10000
4
3
2
1
0
=
=
=
=
=
qp
k
qp
k
qp
k
qp
k
qp
k
C
C
C
C
C

Generated codes have n = 20, w = q = 4,
a
ì = 3,
c
ì = 3.

Example for p =5, q=7, k=3
Code-set contains the extended version of prime codes
01000}, 00100 00010 00001 10000 01000 {00100
00100}, 00001 01000 00010 10000 00100 {00001
00010}, 01000 00001 00100 10000 00010 {01000
00001}, 00010 00100 01000 10000 00001 {00010
10000}, 10000 10000 10000 10000 10000 {10000
4
3
2
1
0
=
=
=
=
=
qp
k
qp
k
qp
k
qp
k
qp
k
C
C
C
C
C
Generated codes have n = 35, w = q = 7,
a
ì = 6,
c
ì = 4.
(
a
ì and
c
ì are the maximum values in a code-set).

C. Quadratic Congruence OOCs
The optical orthogonal code
P
X
C ={ ) 0 (
p
x
c , ) 1 (
p
x
c , ) 2 (
p
x
c ,
….., ) 1 ( ÷ p c
p
x
is based on quadratic placement operator
) (k y
x
, such that
¹
´
¦ = +
=
otherwise
p i kp k y if
i c
x p
x
0
) mod( ) ( 1
) (
where,

p
i
k and p x = ÷ = } 1 ,... 2 , 1 {

| | )) (mod( 1 ) ( ) 1 (
2
p k k y k y
x x
+ + = +
)) (mod(
2
) 1 (
) ( p
k xk
k y
x
+
=
[5]

and for y
x
0 ) 0 ( = 1 0 ÷ s < p x
here the orthogonal code-set ) , , , (
c a
w n ì ì is constructed for
the code length, n = p
2
; weight, w = p ;
a
ì = 2 ;
c
ì =4 or3
(for p=5,7).
For example, p=7, the quadratic operators are given as
1
k
y ={0 1 3 6 3 1 0},
2
k
y ={0 2 6 5 6 2 0},
3
k
y ={0 3 2 4 2 3 0},
4
k
y ={0 4 5 3 5 4 0},
1000000}, 0000100 0000010 0001000 0000010 0000100 {1000000 =
1000000}, 0001000 0010000 0000100 0010000 0001000 {1000000 =
1000000}, 0010000 0000001 0000010 0000001 0010000 {1000000 =
1000000}, 0100000 0001000 0000001 0001000 0100000 {1000000 =
7
4
7
3
7
2
7
1
C
C
C
C
Generated codes have n = 49, w = p = 7,
a
ì = 2,
c
ì = 3.
(
a
ì and
c
ì are the maximum values in a code-set).

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO422-3

D. Hadamard Matrix OOCs
A Hadamard matrix is a square matrix whose entries are
either 1 or 0 and rows are may be mutually orthogonal.
Hadamard matrix order 2 is

=
0 1
1 1
2
H

Its order must be in multiple of 2. 2n Hadamard order can be
given as

=
n n
n n
n
H H
H H
H

Generation of orthogonal codes a matrix of order n-1 can be
constructed from Hadamard matrix of order n by deleting first
row and first column and The rows of the matrix of order n – 1
form a code set. From the same code set the repeated or
cyclically shifted code are included only once to form the
optical orthogonal code set of length n =4t-1; weight w =2t-1,
a
ì = t-1,
c
ì = t, where t is any positive integer. [6]

E. Skolem Sequence OOCs
The Skolem sequences of order M, can be written as
collection of ordered pairs
{ } i a b M i b a
i i i i
= ÷ s s , 1 : ) , (

and
{ } { } M b a
M
i
i i
2 ,....., 2 , 1 ,
1
=
=


M is the number of code-set and the length of code-set n =
6M+1. The orthogonal codes of weight w =3 can be written as
{ }
3 2 1
, ,
i i i
x x x , x
ij
represents the j
th
position of bit „1‟ in the
i
th
code word for 1 M i s s . [7]
x
1 i
= 0 for all i,
x
2 i
= i for all i,
x
3 i
is obtained from the Skolem sequence as in example.

Example for Skolem sequence of order M =4 is given as
S = {(1, 2) (7, 5) (6, 3) (8, 4)}
For i =1, x
i3
= M+2 = 6;
For i =2, x
i3
= M+5 = 9;
For i =3, x
i3
= M+3 = 7;
For i =4, x
i3
= M+4 = 8;
The optical orthogonal code set is
{(0,1,6), (0,2,9), (0,3,7), (0,4,8)}
The corresponding codes are
C
1
= (1100001000000000000000000)
C
2
= (1010000001000000000000000)
C
3
= (1001000100000000000000000)
C
4
= (1000100010000000000000000)
Generated codes have n = 31, w = 3,
a
ì = 1,
c
ì = 1.
Similarly other code words of length n =6M+1, weight w = 3
and
a
ì = 1,
c
ì =1 can be generated using Skolem sequences
of order M. (
a
ì and
c
ì are the maximum values in a code-set).

F. Quadratic Residues OOCs
Quadratic residues are the code-set which is defined by „a‟
as
a = x
2
mod (p)
where, x is integer

The prime „p‟ has total following residues (0,x
1
,x
2
,…….,x
(p-1)/2
)
the QR sequence is Q
1
= (q
1
, q
2
, q
3
,……q
p
) with
q
1
= q
p
= 0 ;
q
2
= q
p-1
= x
1
;
q
3
= x
2 ;
and
q
k
= q
p-k+1
for 1 s k s p.

The j
th
QR sequence is obtained by multiplying Q
1
by j with
all elements are given under mod(p) for j = 1 to p-1.These (p-
1) QR sequence when considered as weighted positions of the
binary code with length n represents the orthogonal code set.
[8]

For example p=5, the quadratic residues are (0, 1, 4)
QR sequence Q
1
= (0, 1, 4, 1, 0)
QR sequence Q
2
= (0, 2, 3, 2, 0)
QR sequence Q
3
= (0, 3, 2, 3, 0)
QR sequence Q
4
= (0, 4, 1, 4, 0)
the orthogonal code set ) , , , (
c a
w n ì ì = ( p
2
, p, 2, 2)
The corresponding code-set (25, 5, 2, 2) are
C
1
= (10000 01000 00001 01000 10000)
C
2
= (10000 00100 00010 00100 10000)
C
3
= (10000 00010 00100 00010 10000)
C
4
= (10000 00001 01000 00001 10000)
Generated codes have n = 25, w = p = 5,
a
ì = 2,
c
ì = 3.
(
a
ì and
c
ì are the maximum values in a code-set).

G. Table of Primes OOCs
The elements of Galois Field GF(p), p is prime, are
(1,2,3,….p-1). Suppose o is primitive root for prime p, then
all the elements of GF(p) can be represented by { o
x
, for x
= (0,1,2….p-2)}. The prime sequence code
p
S
1
= (o
0
, o
1
,……., o
2 ÷ p
)mod(p);
p
S
2
= 2
p
S
1
(mod (p));
p
S
3
= 3
p
S
1
(mod (p));
. .
. .

p
p
S
1 ÷
= (p-1)
p
S
1
(mod (p)).
For example with p=5 and its primitive root o =2 and GF(5)
= (1 2 3 4) with
2
0
= 1
2
1
= 2
2
2
= 4
2
3
= 8 mod(5) =3
The prime sequence code
p
S
1
= (1 2 4 3)
p
S
2
= (2 4 8 6) mod (5) = (2 4 3 1)
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO422-4

p
S
3
= (3 6 12 9) mod (5) = (3 1 2 4)
p
S
4
= (4 8 16 12) mod (5) = (4 3 1 2)
The orthogonal code set with weight positions is given as
follows
p
C
1
= ( 1 p+2 2p+4 3p+3 ) = ( 1 7 14 18 )
p
C
2
= ( 2 p+4 2p+3 3p+1 ) = ( 2 9 13 16 )
p
C
3
= ( 3 p+1 2p+2 3p+4 ) = ( 3 6 12 19 )
p
C
4
= ( 4 p+3 2p+1 3p+2 ) = ( 4 8 11 17 )
The corresponding codes of length n = p
2
-p =20, weight w =
p-1= 4,
a
ì = 2,
c
ì =

1 and code words M= p-1 = 4;
C
1
= (01000 00100 00001 00010);
C
2
= (00100 00001 00010 01000);
C
3
= (00010 01000 00100 00001);
C
4
= (00001 00010 01000 00100);

Similarly for other prime p and root o , the GF(p), its
elements, prime sequence codes and then orthogonal code set
can be generated.
III. PROPOSED ONE DIMENSIONAL CODES
A. Difference of Positions
In binary the „0‟ and „1‟ constitute a binary sequence, if
sequence is orthogonal then it is called by optical orthogonal
codes. In DOP representation „0‟ and „1‟ are changed in the
numbers. From binary sequence the weight (1‟s) position is to
be find. From these positions distance is calculated that will
gives the DOP. For an example n=9, w=3 and suppose the
code is 110010000. [9]

Shifted Seq. Positions of „1‟s DOP
110010000 (0, 1, 4) (1, 3, 5)
100100001 (0, 3, 8) (3, 5, 1)
001000011 (2, 7, 8) (5, 1, 3)
010000110 (1, 6, 7) (5, 1, 3)
100001100 (0, 5, 6) (5, 1, 3)
000011001 (4, 5, 8) (1, 3, 5)
000110010 (3, 4, 7) (1, 3, 5)
001100100 (2, 3, 6) (1, 3, 5)
011001000 (1, 2, 5) (1, 3, 5)

This shows that the code 110010000 can be represented as
(1, 3, 5) or the cyclic shifting of it i.e. (3, 5, 1) or (5, 1, 3)
instead of (0, 1, 4)) or (0, 3, 8) or (2, 7, 8) or (1, 6, 7) or (0, 5,
6) or (4, 5, 8) or (3, 4, 7) or (2, 3, 6) or (1, 2, 5).

B. Extended Matrix
Extended matrix is used to design for the calculation of
a
ì
and
c
ì constraints. A DOP can be reformed in the matrix that
is known as extended matrix. A DOP has elements (C
1
, C
2
, . . .
C
w-1
, C
w
) then the extended matrix for the dimension (w, w-1)
1 1 2 1 2 w-1
2 2 3 2 3 w-2
w w 1 w 1 w+z
C C +C ....... (C +C .....+C )
C C +C ...... (C +C ......+C )
: : :
C C +C ...... (C +C .....+C )









C. Auto-Correlation
For a DOP‟s extended matrix, the elements are checked for
the repetition in its row, suppose a code is (1, 3, 5). The for
this DOP extended matrix is (1 4,3 8,5 6) take 1, 4 and check
to similar number in its row. As in this 1, 4 in not found so the
auto correlation for this DOP is 1.
a
ì = maximum common elements + 1
D. Cross- Correlation
The cross-correlation constraint
c
ì of any two orthogonal
code is also can be observed in the extended difference of
positions representations of code 1 and code 2.

Extended code1 Extended code2
c
ì
(1 2,1 6,5 6) (1 3,2 6,4 5) 2
(1 2,1 6,5 6) (1 4,3 6,3 4) 2
(1 2,1 6,5 6) (2 3,1 5,4 6) 2
(1 2,1 6,5 6) (2 4,2 5,3 5) 2

In any two codes in extended matrix DOP representation.
Comparing the rows of one code‟s extended matrix with other
code‟s extended matrix the maximum common element gives
calculation for
c
ì . ie
c
ì = maximum common elements with other code + 1.

E. Grouping of Codes
Grouping of codes means to form a set of codes having
same valued auto-correlation and cross-correlation. From the
correlation matrix, all codes have value
a
ì are selected so that
a new correlation matrix forms that is of single valued
a
ì and
different valued
c
ì .
For particular code in new correlation matrix similar
c
ì
codes are selected (S). So the different groups are formed by
S
K
C
where K is maximum codes for group calculated by jhonson‟s
bound.
Mutually equal
c
ì codes are the code group for optimal
optical orthogonal code. Similarly others codes group can be
formed.



CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO422-5

IV. COMPARISON TABLE
In the comparison table the algorithms for generating one
dimensional optical orthogonal codes has been compared on
the basis auto-correlation and cross correlation constraints
,
a c
ì ì and number of codes generated.

Type of
algorithm
Auto-
Correlati
on
(
a
ì
)
Cross-
Correlat
ion
(
c
ì
)
Code
length
(n)
Weight
(w)
No. of
codes
generated
Prime
sequence
p-1 2
2
p

p p
Quasi prime Q - 1 2r qp q p

Quadratic
congruence
2 4
2
p

p p-1
Hadamard
matrix
t-1 T 4t - 1 2t - 1

< n
Skolem
sequence
1 1 6M+1 3 M
Quadratic
residues
3 1
2
÷ p

2
p

p p-1
Table of
prime
p - 1 p – 1 p p ÷
2

p - 1 p-1
Proposed
algorithm
w-1 w-1
user
defined
w<<n
Johnson‟s
bound
q varies from (r-1) p<q< rp where r, q is the positive integer,
p is a prime number, m is an integer, M is Skolem sequence
order
V. CONCLUSION
The Optical CDMA system employing optical orthogonal
codes as the optical signature sequence to the OCDMA users.
One dimensional optical orthogonal codes are designed on the
basis of some algorithms found in literatures [ 3-5, 7-9 ] using
MATLAB based programs along with the calculations of auto-
correlation and cross correlation constraints ,
a c
ì ì
respectively. In this paper the algorithms of generating one
dimensional optical orthogonal have been compared on the
basis of some parameter like code length „n‟, code weight „w‟,
and auto-correlation and cross correlation constraints ,
a c
ì ì
which are responsible for BER performance of OCDMA
system. In future these coding schemes can be improved for
generating the groups of one dimensional optical orthogonal
codes with maximum possible codes.

REFERENCES
[1] A. Stok and E. H. Sargent, “The Role of Optical CDMA in Access
Networks,” IEEE Commun. Mag., vol. 40, no. 9, pp. 83–87, 2002.
[2] Fan R.K. Chung, Jawad A. Salehi and Victor K. Wie, “Optical
Orthogonal Codes: Design, Analysis, and Application,” IEEE Trans.
Information Theory, vol. 35, no. 3, may 1989.
[3] A.A.Shaar and P.A.Davis, “Prime sequences and Quasi-optimal
sequences for channel code division multiplexing,” Electronics Letters,
vol. 19, pp. 888-889,October 1983.
[4] Andrew S. Holmes and Richard R.A. Syms, “All Optical CDMA using
Quasi Prime Codes,” IEEE Journal of lightwave technology, vol. 10, no.
2, Feb 1992.
[5] M. Srinivasan, G. Geetha, M. Meenakshi, “Comparision of Prime Code,
Extended Prime Code and Quadratic Congruence Code using a
Normalized Throughpu Metric,” ICTON, pp.168-176, 2004.
[6] L.C.Tran, J. Sebem, B. J. Wysocki, T. A. Wysocki, T. Xia and Ying
Zhao, “Complex Orthogonal Sequences from Amicable Hadamard
Matrices,” Vehicular Technology IEEE Conference, Vol-3, pp 1490-
1493, 2004.
[7] Abrham, J. Kotzig, “Skolem Sequences and Additive Permutations,”
Book Discrete Math, pp 143–146, 1981.
[8] Manoj Choudhary, P. K. Chatterjee and Joseph John, “Optical
Orthogonal Codes using Quadratic Residues,” in Proc. of National
Conference on Communication, IIT Kanpur, 2001.
[9] R.C.S.Chauhan, Rachna Asthana and Y.N.Singh, “Uni-polar Orthogonal
Codes : Design, Analysis and Applications,” HiPC 2010.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO431-1

PERFORMANCE of SOFT DECODING VERSES HARD DECODING
Archana Sharma, Mohd Murtaja, SCRIET, CCS.University Campus, Meerut
Prof. (Dr.) Rajeev Kapoor, H.O.D. of EC, Delhi Technical University, Delhi
Archna.mtech@gmail.com,murtaza_amu@yahoo.com, rajeevkapoor@dce.ac.in

Abstract— We introduce the hard
decision decoding and soft decoding. In
soft decoding scheme the probability of
error is low comparative the hard
decoding scheme. In hard decoding
scheme output is threshold comparison
based we decided the middle value of
maximum signal to minimum signal. In
other words, a firm or hard decision is
made on each bit at the threshold
detector. With soft decision decoding
[1][2][6] , the received codeword is com-
pared in total with the known code words.
I. INTRODUCTION
With hard decision decoding, the output
from the optimum demodulator is passed to
a threshold detector. Using triple
redundancy again as an example, the two
codeword would be 111 and 000. For binary
polar signals, these might be represented by
voltage levels 1V 1 V 1V and -1V -1V -1V.
The threshold level for the threshold
detector would be set at 0 V. If now the
sampled signal from the optimum
demodulator is 0.5 V 0.7 V -2V, the output
from the threshold detector would be 1 V 1
V -1 V, and the decoder would decide that
this was a binary 110 codeword and produce
a binary 1 as output. In other words, a firm
or hard decision is made on each bit at the
threshold detector.
With soft decision decoding
the received codeword is compared in total
with the known code words in the set
111and 000 in this example the comparison
is made on the basis of minimum distance
(the minimum distance referred to here is a
Euclidean distance as described shortly. This
is not the same as the minimum distance
introduced in code). To illustrate this,
consider the first two points in an x, y, z
coordinate system. Let point P
l
have
coordinates x
l
y
l
z
l
and point P
2
have
coordinates x
2
, y
2
z
2
from the geometry of
the situation, the distance d between the
points is obtained from
d
2
= (x
1
- x
2
)
2
+ (y
1
-y
2
)
2
+ (z
1
-z
2
)
2
II. HARD-DECISION-DECODING
ALGORITHM
The algorithm will be explained on the basis
of majority vote [1][3]. An error free
received codeword would be e.g. c = [1 0 0
1 0 1 0 1]. Let’s suppose that we have a
BHC channel and the received the codeword
with one error – bit c1 flipped to 1.
1. In the first step all v-nodes ci send a
”message” to their c-nodes fj containing the
bit they believe to be the correct one for
them. At this stage the only information a v-
node ci has, is the corresponding received i-
th bit of c, yi[3][4].
That means for example, that c0 sends a
message containing 1 to f1 and f3, node c1
sends messages containing y1 (1) to f0 and
f1, and so on.
2. In the second step every check nodes fj
calculate a response to (fj ci) every
connected variable node. The response
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO431-2

message contains the bit that fj believes to
be the correct one for this v-node ci
assuming that the other v-nodes connected
to fj are correct. This will be the case if all
check equations [2][1]are fulfilled. We will
later see that the whole algorithm contains a
loop, so another possibility to stop would be
a threshold for the amount of loops [2][4].
3. Next phase: the v-nodes receive the
messages from the check nodes and use this
additional information to decide if their
originally received bit is OK. A simple way
to do this is a majority vote. When coming
back to our example that means, that each v-
node has three sources of information
concerning its bit. The original bit received
and two suggestions from the check nodes.
Table 2 illustrates this step. Now the v-
nodes can send another message with their
(hard) decision for the correct value to the
check nodes.
4. Go to step 2. Loop In our example, the
second execution of step 2 would terminate
the decoding process since c1 has voted for
0 in the last step. This corrects the v-nodes
use the answer messages from the c-nodes to
perform a majority vote on the bit value.
The transmission error and all check
equations are now satisfied.

Table 1. Comparision of error
III. SOFT INPUT SOFT OUTOUT
ALGORITHMS
The source block delivers information by the
mean of sequences which are row vectors x
of length K. The encoder block delivers the
codeword c of length N, which is the coded
version of x. The code rate is defined by the
ratio R = K/N. The codeword c is sent over
the channel and the vector y is the received
word: a distorted version of c.
Hence, the channel is represented by a
discrete time equivalent model. The channel
is a non-deterministic mapped between its
input c and its output y. We assume that y
depends on c via a conditional probability
density function (pdf) p(y|c). We assume also
that the channel is memoryless:
p(y|c)=
For example, if the channel is the binary-
input additive white Gaussian noise (BI-
AWGN), and if the modulation is a binary
phased shift keying (BPSK) modulation with
the 0= +A, 1 = −A mapping, we have:
p(yn|cn)=


where Es = A
2
is the energy of the symbol
sent over the channel.
Soft Input Soft Output (SISO)
algorithms [1][2]are the fundamental
building blocks of iterative decoders. A SISO
is in general a block that accepts messages
(to be defined later) about both the input and
output symbols of an encoder and provides
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO431-3

extrinsic messages about the same symbols.
The extrinsic message is generated by
considering the a priori constraints that exist
between the input and output sequences of
the encoder. In this section we will give more
precise definitions of the “message” and on
the input output relationships of a SISO
block. Moreover we will show efficient
algorithms to perform SISO for some special
types of encoders.
A. Definition of the input and output metrics
A SISO module generally works
associated to a known mapping (encoding)
between input and output alphabets [4].
A SISO module is a four port device
that accepts some messages of the input and
output symbols of the encoder and provides
homologous output extrinsic messages. We
will consider the following two types of
normalized messages:
Likelihood ratio (LR)
LLR information L(x ) = ln ,
L(x) represents the ratio between the
likelihood of the symbol being 1 and 0 "_

IV. SIMULATION AND ANALYSIS
RESULTS
If we consider the power spectral density 5,
and no of message bits n as in table and
normalized bit energy E
b
6 then if the error
is in 2 bit then calculate the probability of
error in soft decoding and hard decoding
scheme are shown in table in table.2
Table- 2


Fig.-1 decoding comparisons
The probability of error is low and higher
decoding throughput in soft decoding
scheme then hard decoding scheme as
shown in fig.-1





DECISION n=1 n=2 n=3 n=4 n=5 n=6 n=7 n=8 n=10
P_e_hard 0.0239 0.8744 2.903 5.2896 7.5817 9.6382 11.4406 12.034 15.3502
P_e_soft 0.0000 0.0047 0.0623 0.2333 0.5246 0.9107 1.3607 1.8489 2.4600
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COMO431-4




REFERENCES
[1] R. G. Gallager, Low Density Parity
Check Codes. MIT Press, 1963.
[2] D. MacKay, “Good Error-Correcting
Codes Based on Very Sparse Matrices,”
Information Theory, IEEE Transactions on,
vol. 45, 1999.
[3] C. Berrou, A. Glavieux, and P.
Thitimajshima, “Near optimum error
correcting coding and decoding: Turbo-
Codes,” in Communications,
1993.Technical Program, Conference
Record, IEEE International Conference on,
May 1993
[4]Y. Dai, N. Chen, and Z. Yan, “Memory-
efficient decoder architectures for LDPC”
accepted and to appear in IEEE
Trans.Circuits and Systems.
[5]P. Gopalan, R. Lipton, and Y. Z. Ding,
“Codes, adversaries, and information:
[6]SIMAON HAYKIN, Digital
Communications McMaster University
JHON WILEY &SONS.
[7] M. Mansour and N. Shanbhag, “High-
throughput LDPC decoders,” IEEE Tran.
Very Large Scale Integration Systems, vol.
11, pp. 976 – 996, 2003.
[8] B.P. Lathi, “Modern Digital and Analog
Communication System” Oxford University
Press Inc,1998.
[9] Taub Schilling, “Principles of
Communication Systems” Tata McGRAW
HILL,1991.

[10] P. Gopalan, R. Lipton, and Y. Z. Ding,
“Codes, adversaries, and information: A
computational approach,” paper, 2001,
submitted for publication.
[11] Clark, G. C. Jr. and J. Bibb Cain.,
Error-Correction Coding for Digital
Communications, New York, Plenum Press,
1981.
[12] Gitlin, R. D., J. F. Hayes, and S. B.
Weinstein, Data Communications
Principles, New York, Plenum, 1992.
[13] Heller, J. A. and I. M. Jacobs, "Viterbi
Decoding for Satellite and Space
Communication," IEEE Transactions on
Communication Technology, Vol. COM-19,
October 1971, pp 835–848.
[14] Yasuda, Y., et. al., "High-rate
punctured convolutional codes for soft
decision Viterbi decoding," IEEE
Transactions on Communications, Vol.
COM-32, No. 3, pp 315–319, March 1984.
[15] Haccoun, D., and Begin, G., "High-
rate punctured convolutional codes for
Viterbi and sequential decoding," IEEE
Transactions on Communications, Vol. 37,
No. 11, pp 1113–1125, Nov. 1989.




CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O1-1


A Critical Analysis on Prosthetic Arm


Anil kumar
B.Tech (EI 3 yr)
Galgotia‟s college of Eng. & Tech.
Greater Noida
anil00699@gmail.com

Abstract – Prosthetic limbs are a real
time challenging and demanding
biomechatronics field due to more and
more day-to-day basis work objectives for
amputees. This paper investigates the
design of various prosthetic arms to
mimic its human counterpart. The
comparative analysis of available
technology and research work in field of
bionic arms are being done. In this
analysis, the merits and demerits of
various prosthetic hand for real time
application is being highlighted. In-depth
analysis shows that further enhancement
could be done in order to augment DOF,
grasp control, available strength of the
artificial arm and signal processing
techniques.
Keywords- Prosthetic Arm, Robotic Arm,
Artificial hand

INTRODUCTION
Prosthesis is an artificial extension that
replaces a missing body part. Prostheses are
typically used to replace parts lost by injury
or missing from birth or to supplement
defective body parts. In addition to the
standard artificial limb for every-day use,
many amputees have special limbs and
devices to aid in the participation of sports
and recreational activities. Myoelectric
Dr. Monika Jain
Professor-Electronics & Instrumentation
Galgotia‟s college of Eng. & Tech.
Greater Noida
monika_jain24@rediffmail.com

prosthetic hand is used as natural hand by
amputees (who lost their natural hand) by
amputation. Initially, hooks are being used
for this purpose. In recent time, with the
advancement in technology, artificial smart
arm are being used. In myoelectric hands,
electromyography (EMG) signals are used
as control input. EMG signals are
bioelectrical signal released by muscles.
They are normally of 10-500 Hz frequency
and up to 10 mV. The main part of signal is
near 450Hz. conventionally, surface
myoelectrodes are being used in myoelectric
prosthetic arms, however with help of SRI
EMG technique, signal strength and
accuracy of signal would increase. This can
be use to control individual fingers. If
prosthetic hand can be made of same
mechanical properties as a neuromuscular
control system of hand then amputee can use
it more effectively. A lot of work is done in
this field to improve prosthetic arms but
feedback is not used effectively. By using
feedback connected at artificial fingers like
pressure sensor, accelerometer, temperature
sensor, hand‟s working can be improved.
Also DOF and grasping capacity of artificial
arm can be improved. There is no doubt that
microelectronics offers numerous
opportunities for improvement in the
engineering of substitutes for the human
limb. The possibilities become exciting as
soon as we move away from the all-
mechanical prostheses which have been the
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O1-2


conventional prescription until today. We
have described a relatively simple
application of hybrid microelectronics in
this novel field.
The purpose of this project is to develop a
myoelectric controlled individual finger,
multi-DOF adaptive grasping technique
equipped commercial hand capable to fast
response. Also amputee can feel touch,
temperature and force applied by prosthetic
hand on the joint of prosthetic and natural
hand.

LITERATURE SURVEY
In the field of prosthetic arm, lots of
research work is in progress. Maximum
prosthetic hands are simple grippers with
one or two degrees of freedom (DOF). They
are using smart hooks (passive fingers and
thumb) like Otto Bock [1] they have 2 or 3
point of contact so more force is required for
grasping. Also they are capable of gripping
as in myohand variplus speed [2] but not
capable of some more works that include
finger, like door opening, rotating car key
etc. Kenzo Akazawa [3] developed a hand
using dynamic behaviour of antagonist
muscles, flexor muscles and extensor
muscles but is similar as in [1]. Also its
response is slow and a large training session
is required to train the amputee. For fast
processing, Isamu [4] has presented
evolvable hardware chips with genetic
algorithm. It takes less time to train
amputee. But it restricts the DOF of hand.
Also dedicated chip is required for
prosthetic hand which reduces versatility to
use high speed microcontrollers.
Microprocessor and high torque motor based
hand is being made by Ryuhei [5], however,
this uses only two surface EMG signals, due
to which less DOF is obtained. Also, more
training required for amputee. Another main
drawback in above is that they uses EMG
signal as stationary signal but for multi DOF
it is required to process it as non stationary
signal. Real time EMG pattern recognition
[6] is done to improve the hand functioning.
But still signal is not sufficient so fingers
can‟t be controlled separately. For this
purpose SRI EMG (surface recorded
intramuscular EMG [7] signals can be used
they are more accurate than surface EMG
signals. Prosthetic hand with individual
finger movement will be more similar to
natural hand. Prosthetic hand approximately
similar to natural hand is made by Bryan
Christie [8] and Dean Kamen [9] with the
help of DARPA. But it is connected by
nerve system so not easy to use also it‟s still
in experimental state. Chappell [10] has
presented an approach having an artificial
hand with sensors allowing for the inclusion
of automatic control loops, freeing the user
from the cognitive burden of object holding
which is similar to the natural low level
spinal loops that automatically compensate
for object movement. Force, object slip and
finger positions are variables that need to be
measured in a hand designed for the
physically impaired person. It shows that a
high specification sensor is required for
designing an arm. Also it must be designed
separately for each amputee. But latest
technology provides adaptive signal
processing helping to adapt the amputees‟
requirements. An electrically driven locking
mechanism has been built by Law and
Hewson [11], which is controlled by the
electromyogram (EMG) of the surviving
muscles in the upper arm. Hybrid
technology is used for the construction of
the associated electronic circuitry. Many
similar applications are now being
considered in an attempt to improve the
performance of upper-limb prostheses using
latest researches. Development, testing and
experimentation work of a device for the
hand rehabilitation was done by Mulas et al
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O1-3


[12]. The system designed is intended for
people who have partially lost the ability to
control correctly the hand musculature, for
example after a stroke or a spinal cord
injury. Based on EMG signals the system
can "understand" the subject volition to
move the hand and actuators can help the
fingers movement in order to perform the
task. This paper describes the device and
discusses the first results conducted on a
healthy volunteer. It requires a number of
actuators to increase the DOF of prosthetic
arm. Also EMG processing done is not
sufficient to provide significant
performance. Massa et al [13] designed a
hand to augment the dexterity of traditional
prosthetic hands while maintaining
approximately the same dimension and
weight. This approach is aimed at providing
enhanced grasping capabilities and natural
sensory-motor coordination to the amputee,
by integrating miniature mechanisms,
sensors, actuators, and embedded control. A
biomechatronics hand prototype with three
fingers and a total of six independent DOFs
has been designed and fabricated. This
research work is focused on the actuators
system, which is based on miniature
electromagnetic motors. However, still not
using better EMG processing technologies
which can dramatically increase
performance of the hand also the grasping
force is low because of the limited torque
generated by the miniature actuators used
for the hand (which are among the best
available on the market in that range of
size). Embedded control architecture for the
action and perception of an
anthropomorphic 16 degree of freedom, 4
degree of actuation prosthetic hand for use
by transradial amputees has also been
reported. The prosthetic hand is provided
with 40 structurally integrated sensors useful
both for automatic grasp control and for
biofeedback delivery to the user through an
appropriate interface (either neural or non-
invasive). Cipriani et al [14] briefly
describes the mechatronic design of the
prosthesis, the set of sensors embedded in
the hand and finally focuses on the design of
the control architecture that allows action
and perception for such a sophisticated
device. It consists of 8-bit architecture
microcontrollers, however, not using the
available signal processing techniques.
Herrera et al [15] designed and constructed
a prosthesis that will be strong and reliable,
while still offering control on the force
exerted by the artificial hand. The design
had to account for mechanical and electrical
design reliability and size. These goals were
targeted by using EMG in the electrical
control system and a linear motion approach
in the mechanical system. The prosthetic
gripper uses EMG to detect the amputee's
intended movement. The control system
requires an adaptation mechanism for each
amputee's characteristics. Gordon et al [16]
used proportional myoelectric control of a
one-dimensional virtual object to investigate
differences in efferent control between the
proximal and distal muscles of the upper
limbs. Restricted movement was allowed
while recording EMG signals from elbow or
wrist flexors/extensors during isometric
contractions. Subjects used this proportional
EMG control to move the virtual object
through two tracking tasks, one with a static
target and one with a moving target (i.e., a
sine wave). Eriksson et al [17] studied the
neural network feasibility for categorizing
patterns of EMG signals. The signals
recorded by the surface electrodes are
sufficient to control the movements of a
virtual prosthesis. The presented method
offers great potential for the development of
future hand prostheses. A signal processing
system based on RAM as a look-up table
(LUT) has been presented by Torresen et al
[18]. This provides a fast response besides
being compact in size. Several algorithms
for programming it have been proposed For
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O1-4


the given data set used in the following
experiments, the time used for programming
the RAM was approximately equal to the
time used for training a feed-forward neural
network – solving the same problem in the
best possible way. However, the main
advantage of this scheme is the fast runtime
speed. Ferguson [19] described the
development of a system that will allow
complex grasp shapes to be identified based
on natural muscle movement. The
application of this system can be extended to
a general device controller where input is
obtained from forearm muscle, measured
using surface electrodes. This system
provides the advantage of being less
fatiguing than traditional input devices. V.
Tawiwat et al. [20] applied a mouse„s roller
with a gripper to increase the efficiency of a
gripper could lead to material handling
without slipping. To apply a gripper, the
optimization principle is used to develop
material handling by use of a signal for
checking a roller mouse that rotates. In case
the roller rotates, meaning that the material
slips. A gripper will slide to material
handling until the roller does not rotate. In
an attempt to improve the functionality of a
prosthetic hand device, a new fingertip has
been developed, that incorporates sensors to
measure temperature and grip force and to
detect the onset of object-slip from the hand.
The sensors have been implemented using
thick-film printing technology and exploit
the piezoresistive characteristics of
commercially available screen printing
resistor pastes and the piezoelectric
properties of proprietary lead-zirconate-
titanate (PZT) formulated pastes. The force
sensor exhibits a highly linear response to
forces. The force sensor response is also
extremely stable with temperature. The
ability of the piezoelectric PZT vibration
sensor to detect small vibrations of the
cantilever, indicative of object slip, has also
been demonstrated [21]. Externally powered
upper extremity prosthesis has been
considered as a system in which the
necessary components to design a better
prosthetic arm are viewed and divided into
four subsystems: input, effecter, feedback
and support. Current research is reviewed in
terms of these subsystems. Each subsystem
performs its own task, but they are related to
each other and together they function to
make up a prosthetic upper extremity, which
provides the movement to the amputee [22].
Hands such as the All Electric Prosthetic
Hand utilize a series of gears to transmit the
motion of motors housed in the forearm to
the relevant fingers [23]. Other designs have
the actuators transmitting the power directly
to the joint. An example of this is the
Anthroform Arm, which uses pneumatic
„muscles‟ mimicking the muscles of the
human arm connected directly to the „bones‟
it moves [24]. Shape Memory Alloy wires
are also used, to both provide the force and
transmit the motion. SMA wires contract
when heated and return to their initial shape
when cooled [25]. This method of actuation
is utilized in the Shape Memory Alloy
Activated Hand constructed by DeLaurentis
et al [26].

CONCLUSIONS

A lot of work is done for developing the
prosthetic arm but still more precision work
could be done. Grasping technique can be
enhanced so that amputees can work more
effectively, response time of available arms
is not sufficient. Still the available prosthetic
arms are not comparable with natural hand
in terms of DOF. The main limitations are
less space available for motors and their low
size to torque ratio. Power consumed is also
more if number of motors increases. The
lack of enhanced signal processing
techniques in artificial hands and associated
controllers has limited their functionality
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O1-5


and technological progress. Still a lot of
work can be done to improve the available
prosthetic arms. Use of hydraulics in place
of motors can provide great amount of
controlled force in lesser space. Due to this
available control of force applied by the
hand on the object, amputees can handle soft
or brittle items easily. By effective use of
hydraulics, DOF can also be increased
without any increased power consumption.
Also use of adaptive signal processing
techniques can improve overall performance
of artificial hand.

REFERENCES
[1]Available on:
http://www.ottobock.com.au/cps/rde/xchg/o
b_au_en/hs.xsl/384.html
[2Available on:
http://www.ottobock.com.au/cps/rde/xchg/o
b_au_en/hs.xsl/19932.html
[3] Kenzo Akazawa, Ryuhei Okuno and
Masaki Yoshida “Biomimetic EMG-
prosthetic-hand” 18th Annual International
Conference of the IEEE Engineering in
Medicine and Biology Society Amsterda,
pp. 535-536.
[4] Isamu Kajitani and Tsukuba Masahiro
Murakawa “An Evolvable Hardware Chip
for Prosthetic Hand Controller”
Microelectronics for Neural, Fuzzy and Bio-
Inspired Systems, 1999. MicroNeuro' 99.
Proceedings of the Seventh International
Conference on pp. 179-186.
[5] Ryuhei Okunol, Masahiro Fujikawa',
Masaki Yoshida2, Keno Akazawal
“Biomimetic hand prosthesis with easily
programmable microprocessor and high
torque motor” Engineering in Medicine and
Biology Society, 2003. Proceedings of the
25th Annual International Conference of the
IEEE Vol.2 pp.1674-1677.
[6] Jun-Uk Chu, Inhyuk Moon, Member,
IEEE, and Mu-Seong Mun “A Supervised
Feature Projection for Real-Time
Multifunction Myoelectric Hand Control”
International Conference of the IEEE
Engineering in Medicine and Biology
Society pp.282-290.
[7] Nikolay S. Stoykov, Madeleine M.
Lowery, Charles J. Heckman, Allen Taflove,
and Todd A. Kuiken “Recording
Intramuscular EMG Signals Using Surface
Electrodes” Proceedings of the 2005 IEEE
9th International Conference on
Rehabilitation Robotics June 28 - July 1,
2005, Chicago, IL, USA pp. 291-294.
[8] Available on:
http://spectrum.ieee.org/robotics/medical-
robots/winner-the-revolution-will-be-
prosthetized/2
[9] Available on:
http://spectrum.ieee.org/biomedical/bionics/
dean-kamens-luke-arm-prosthesis-readies-
for-clinical-trials
[10] P H Chappell, “A fist full of sensors”
Journal of Physics: Conference Series, vol.
15, 2005, pp. 7-12.
[11] H.T. Law and J. J. Hewson, “An
Electromyographically Controlled Elbow
Locking Mechanism for an Upper Limb
Prosthesis”, Electro component Science and
Technology, vol. 10, 1983, pp. 87-93.
[12] Marcello Mulas, Michele Folgheraiter
and Giuseppina Gini, “An EMG-controlled
Exoskeleton for Hand Rehabilitation”, IEEE
9th International Conference on
Rehabilitation Robotics, Chicago, IL, USA,
June 28 - July 1, 2005, pp 371-374.
[13] M. C. Carrozza, S. Micera, R.
Lazzarini, M. Zecca and P. Dario,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O1-6


“Development of a Novel Prosthetic Hand”
Ongoing Research and Preliminary
Results‖ IEEE/ASME Transactions on
Mechatronics, vol. 7, 2002, pp. 108-114.
[14] Christian Cipriani, Marco Controzzi,
Fabrizio Vecchi and M. Chiara Carrozza,
“Embedded Hardware Architecture Based
on Microcontrollers for the Action and
Perception of a Transradial Prosthesis”,
International Conference on Biomedical
Robotics and Biomechatronics Scottsdale,
October 19-22, 2008, AZ, USA, pp. 848-
853.
[15] Andres Herrera, Andres Bernal, David
Isaza and Malek Adjouadi, “Design of an
Electrical Prosthetic Gripper Using EMG
and Linear Motion Approach”, National
Science Foundation Grants EIA-9906600
and HRD-0317692.
[16] Keith E. Gordon, Daniel P. Ferris,
“Proportional Myoelectric Control of a
Virtual Object to Investigate Human
Efferent Control”, Exp Brain Res, vol. 159,
2004, pp 478–486.
[17] Lars Eriksson, Fredrik Sebelius,
Christian Balkenius, “Neural Control of a
Virtual Prosthesis”, IEEE 8th International
Conference on Artificial Neural Networks,
vol. 2, 1998, pp. 905-910.
[18] Jim Torresen, Vidar Engh Skaugen, “A
Signal Processing Architecture Based on
RAM Technology”, 16th European
Simulation Multiconference on Modelling
and Simulation, ISBN:90-77039-07-4, pp.
317-319.
[19] Simon Ferguson, G Reg Dunlop,”Grasp
Recognition from Myoelectric Signal”,
Australasian Conference on Robotics and
Automation, Nov 27-29, 2002, Auckland,
pp.83-87.
[20] V. Tawiwat, and S. Sarawut, “Optimal
Design for Grip Force of Material
Handling”, Proceedings of World Academy
of Science, Engineering and Technology,
July, 2008, pp 864-868.
[21] A. Cranny , D.P.J. Cotton, P.H.
Chappell, S.P. Beeby, N.M. White, “Thick-
film Force and Slip Sensors for a Prosthetic
Hand”, Sensor and Actuator 2005, pp. 162-
171.
[22] HaeOck Lee, Dawnlee J. Roberson, “A
Systemic View of an Upper Extremity
Prosthesis”, IEEE International Conference
on System of Systems Engineering, April
2007, pp. 1-6.
[23] I. Kato and K. Sadamoto (ed.),
“Mechanical Hands Illustrated”, Survey
Japan (1982).

[24] G. K. Klute, J. M. Czerniecki and B.
Hannaford, “McKibben Artificial Muscles:
Pneumatic Actuators with Biomechanical
Intelligence”, [Online, accessed 26 March
2001]
URL:
http://brl.ee.washington.edu/BRL/devices/m
ckibben/aim99v6.pdf

[25] J. L. Banks, “Anthropomorphic Robotic
Finger Platform Based on Shape Memory
Alloy”, Master‟s Thesis, MIT Artificial
Intelligence Laboratory, [Online, accessed 4
May 2001]
URL:
http://www.ai.mit.edu/projects/humanoid-
robotics-group/Abstracts2000/banks.pdf

[26] K. J. DeLaurentis, C. Mavroidis and C.
Pfeiffer, “Development of a Shape Memory
Alloy Activated Robotic Hand”, [Online,
accessed 19 March 2001] URL:
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O1-7


http://www.cronos.rutgers.edu/~mavro/pape
rs/act2000.pdf

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-1

Analysis of Bio-inspired Algorithms–
BFO, PSO and FFA
Vipul Singhal
Vipul.vanshi@gmail.com
Delhi College of Engineering, Delhi University, Delhi, India

ABSTRACT

Engineering problems can be defined as search for
one near optimal description among many
possibilities, under real time constraints. Search
techniques such as Bacterial foraging, Particle swarm
and Firefly based optimization algorithms are now
among the latest research topics. This paper throws
light on these bio-inspired evolutionary algorithms
that are used to optimize the objective function, given
constraints. Optimization as an engineering problem
has been exposed as their one application along with
the several advancements made in this field. These
techniques are explained in detail along with work
and research that has been done in that area. It also
provides the insight of modification and
improvements, these algorithms continue to enjoy
performing better for specified problem.


1. INTRODUCTION

Optimization has been an active area of research
for several decades. As many real-world
optimization problems become increasingly
complex, better optimization algorithms are
needed. Over the past few decades, many
biologically inspired computational methodologies
have been invented, such as evolutionary
algorithms which include genetic algorithms,
bacterial foraging, particle swarm firefly algorithms
and many others. These tend to enjoy a regular
attention by scientist in several research areas of
soft computing and others as well.
Here in this paper we have taken review of three
basic algorithms and have seen how they can be
applied to several problems. Section 1 contributes
to explain basic and classic algorithm of BFA
which is then extended to provide its further
applications and research areas. It tells about how
the bacteria forage by maximising their energy
intake in minimum span of time. This concept then
is used to keep them moving and later reproduce to
give next generation. It also provides that those
with poor foraging strategy tend to eliminate.
Section 2 explains basics and classic particle
swarm optimization technique. It also throws light
on the work done in that area and its applications.
Section 3 then explains firefly algorithm and its
emergence. Advancements and research areas then
exposed.

2. BACTERIAL FORAGING OPTIMIZATION:

Bacterial foraging algorithm is inspired by the
pattern exhibited by bacterial foraging behaviours
[1]. They tend to maximise energy per unit time.
Bacteria have the tendency to gather to the nutrient-
rich areas by an activity called “chemotaxis”. It is
known that bacteria swim by rotating whip-like
flagella driven by a reversible motor embedded in
the cell wall. E. coli has 8-10 flagella placed
randomly on a cell body. When all flagella rotate
counter clockwise, the move is called Run. When
the flagella rotate clockwise, they all pull on the
bacterium in different directions, which causes the
bacteria to Tumble.


FIG. 1.1 SWIM AND TUMBLE OF A BACTERIUM [1]
.
The foraging algorithm is based on the behaviour
of the E. coli bacteria, which has tree fundamental
steps: chemotaxis, reproduction, and elimination
and dispersion. Chemotaxis is used in order to
move and search for food, this step has much
resemblance with a biased random-walk model [2].
In order to maximise energy in minimum time
bacteria tend to follow different search strategies as
described in fig. 1.2


CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-2


FIGURE 1.2 SEARCH STRATEGIES FOR FORAGING
BACTERIA [1]

The chemotactic operator employed in BFOA is
supposed to guide the swarm to converge toward
optima. Reproduction is used to make an identical
copy; and elimination and dispersion is used to
avoid noxious elements that can kill or spread a
group of bacteria in the search space.


2.1. CLASSICAL ALGORITHM:

When bacteria get food in sufficient amount, they
increase in length, and in presence of suitable
temperature, they break in the middle to form an
exact replica of themselves. This phenomenon
inspired Passino[1] to introduce an event of
reproduction in BFOA. Due to the occurrence of
sudden environmental changes or attack, the
chemotactic progress may halt, and a group of
bacteria may move to some other places. This
constitutes the event of elimination–dispersal in the
real bacterial population, where all the bacteria in a
region are killed or a group is dispersed into a new
part of the environment. As said, bacteria try to
move to areas which are rich in nutrients and free
of noxious substances [4]. If θ is a position and J(θ)
is a function which is the more negative the more
the nutrients prevail over the noxious substances,
then we can imagine that bacteria naturally try to
find those, where J (θ) has its minima.
A bacterium position after a tumble can be
determined through (1), where the position at a
given instant is calculated in terms of the position
at the previous instant and the step size C(i) applied
in a random direction ϕ(j), generated by the
bacterium tumble

θ
i
(j+1,k,l)= θ
i
(j,k,l)+C(i)* ϕ(j,k,l) (1)

Where i indicate the bacterium in a total population
of S individuals, j the chemotactic step, the
reproductive step k, and the elimination and
dispersal step l.
The health function is calculated at these positions
of bacteria as (2)

J(i,j,k,l)=J(i,j,k,l)+J
cc
(θ’(j,k,l),P(j,k,l)). (2)

Here second expression shows the effect of
attraction and repulsion among the bacteria. The
exact expression of J
cc
(θ’(j,k,l),P(j,k,l)) is in [1].
And this then is compared with the previous health
function as

J (θ
i
(j+1,k,l))<J(θ
i
(j,k,l))

If it’s true then the bacterium will keep moving in
the same direction (“run” behaviour) for a given
maximum number of steps, after which a “tumble”
will occur anyway. The least healthy bacteria
eventually die while each of the healthier bacteria
(those yielding lower value of the objective
function) asexually split into two bacteria, which
are then placed in the same location. This keeps the
swarm size constant. This is done through
reproduction step. Some bacteria are liquidated at
random with a very small probability while the new
replacements are randomly initialized over the
search space.

2.2. PSEUDOCODE:

1. Initialize parameters n, N, Nc, Ns, Nre, Ned,
Ped,
C(i)(i=1,2…N), i =1,2,---
Where,
n: Dimension of the search space,
N: The number of bacteria in the population,
NC : Chemotactic steps,
Nre: The number of reproduction steps,
Ned : the number of elimination-dispersal events,
Ped: Elimination-dispersal with probability,
C (i): the size of the step taken in the random
direction specified by the tumble.
2. Elimination-dispersal loop: l=l+1
3. Reproduction loop: k=k+1
4. Chemotaxis loop: j=j+1
[a] For i =1,2…N, take a chemotactic step for
bacterium i as follows.
[b] Compute fitness function, J (i, j, k, l).
Let,J (i, j, k, l) =J (i, j, k, l)+J
cc

i
( j, k, l), P( j, k, l))
(i.e. add on the cell-to cell attractant–repellent
profile to simulate the swarming behaviour )
.Where, J
cc
is defined in .
[c] Let J
last
=J (i, j, k, l) to save this value since we
may find a better cost via a run.
[d] Tumble: generate a random vector Δ(i)€ Rn
with each element ∆
m
(i),m =1,2,..., p, a random
number on
[-1, 1].
[e] Move:
Let
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-3

This results in a step of size C(i) in the direction of
the tumble for bacterium i.
[f] Compute J (i, j +1, k, l) and
Let J(i, j,k,l)=J(i, j,k,l)+J
cc
( θ
i
( j,k,l),P( j,k,l))
[g] Swim
i) Let m=0 (counter for swim length).
ii) While m<N
s
(if have not climbed down too long).
• Let m=m+1.
• If J (i, j +1, k, l)<J
last
( if doing better), let
J
last
= J (i, j+1, k, l) and
Let

And use this θ
i
(i+1, j, k) to compute the new J (i,
j+1, k, l) as we did in [f]
Else, let m= N
s
. This is the end of the while
statement.
[h] Go to next bacterium (i, 1) if i≠N (i.e., go to [b]
to process the next bacterium).
5. If j<Nc, go to Step 3. In this case, continue
chemotaxis, since the life of the bacteria is not
over.
6. Reproduction:
[a] For the given k and l, and for each i =1,2,...,N,
Let
be the health of the bacterium i (a measure of how
many nutrients it got over its lifetime and how
successful it was at avoiding noxious substances).
Sort bacteria and chemotactic parameters C(i) in
order of ascending cost health J (higher cost means
lower health).
[b] The Sr bacteria with the highest J
health
values die
and the remaining Sr bacteria with the best values
split (this process is performed by the copies that
are made are placed at the same location as their
parent).
7. If k<Nre, go to Step 3. In this case, we have not
reached the number of specified reproduction steps,
so we start the next generation of the chemotactic
loop.
8. Elimination-dispersal: For i=1,2...,N, with
probability Ped, eliminate and disperse each
bacterium, and this result in keeping the number of
bacteria in the population constant. To do this, if a
bacterium is eliminated, simply disperse one to a
random location on the optimization domain. If
l<Ned , then go to Step. 2 ; otherwise end.


2.3. ADVANCEMENTS IN BFO AND ITS APPLICATION AND
RESEARCH AREAS:

Vast applications have been found where BFO has
shown remarkable results and has been modified
for different problems according to the objective
function. Initial applications of evolutionary
algorithm were meant for static optimization
problems but in recent years the emergence of
another member of the EA family[5]– bacterial
foraging algorithm (BFA), the self adaptability of
individuals in the group searching activities has
attracted a great deal of interests including dynamic
problems. W. J. Tang and Q. H. Wu have
contributed their work by proposing DBFA, which
is especially designed to deal with dynamic
optimization problems, combining the advantage of
both local search in BFA and a new selection
scheme for diversity generating. They used the
moving peaks benchmark (MPB) [6] as the test bed
for experiments. The performance of the DBFA is
evaluated in two ways. The first is concerned with
the convergence of the algorithm in random
periodical changes in an environment, which are
divided into three ranges from a low probability of
changes to a higher one. The second is testing a set
of combinations of the algorithm parameters which
are largely related to the accuracy and stability of
the algorithm. All results are compared with the
existing BFA [1], and show the effectiveness of
DBFA for solving dynamic optimization problems.
It is worth mentioning that the diversity of DBFA
changes after each chemotactic process rather than
the dispersion adopted by the BFA after several
generations. The DBFA utilizes not only the local
search but also applies a flexible selection scheme
to maintain a suitable diversity during the whole
evolutionary process. It outperforms BFA in almost
all dynamic environments. The results are shown in
[5]. They have further given solution for global
optimization given in [7].
The novel BSA has been proposed for global
optimization. In this algorithm, the adaptive tumble
and run operators have been developed and
incorporated, which are based on the understanding
of the details of bacterial chemotactic process. The
operators involve two parts: the first is concerned
with the selections of tumble and run actions, based
on their probabilities which are updated during the
searching process; the second is related to the
length of run steps, which is made adaptive and
independent of the knowledge of optimization
problems. These two parts are utilized to balance
the global and local searching capabilities of BSA.
Beyond the tumble and run operators, attraction
and mutation operations have also been developed.
A.ABRAHAM, A. BISWAS, S. DASGUPTA AND S.
DAS have shown [8] that the major driving forces
of Bacterial Foraging Optimization Algorithm
(BFOA) is the reproduction phenomenon of virtual
bacteria each of which models one trial solution of
the optimization problem.
BFO and PSO have been used in combination and
their combined performance has been utilised to
incorporate the merits [9] of two bio-inspired
algorithms to improve the convergence for high-
dimensional function optimization. It is assumed
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-4

that the bacteria have the similar ability like birds
to follow the best bacterium (bacterium with the
best position in the previous chemotactic process)
in the optimization domain. The position of each
bacterium after every move (tumble or run) is
updated according to (3)

θ
i
(j+1,k,l)=θ
i
(j+1,k,l)+C
cc

b
(j,k,l)-θ
i
(j,k,l), (3)

if J
i
(j+1,k,l)>J
min
(j,k,l)

where θ
b
(j,k,l) and J
min
(j,k,l) are the position and
fitness value of the best bacterium in the previous
chemotactic process respectively, C
cc
is a new
parameter, called attraction factor, to adjust the
bacterial trajectory according to the location of the
best bacterium.


3. PARTICLE SWARM OPTIMIZATION:

Particle swarm optimization is a high-performance
optimizer that is very easy to understand and
implement. It is similar in some ways to genetic
algorithms or evolutionary algorithms, but requires
less computational bookkeeping and generally only
a few lines of code [10]. Particle swarm
optimization originated in studies of synchronous
bird flocking and fish schooling, when the
investigators realized that their simulation
algorithms possessed an optimizing
characteristic[11]-[13]. As the particles traverse the
problem hyperspace, each particle remembers its
own personal best position that it has ever found,
called its local best. Each particle also knows the
best position found by any particle in the swarm,
called the global best. Overshoot and undershoot
combined with stochastic adjustment explore
regions throughout the problem hyperspace,
eventually settling down near a good solution. This
process can be visualized as a dynamical system,
although the behaviour is extraordinarily complex
even when only a single particle is considered with
extremely simplified update rules. This new
optimization technique has much promise, and
electromagnetic researchers are just beginning to
explore its capabilities.


3.1. CLASSICAL ALGORITHM:

The particle swarm concept originated as a
simulation of a simplified social system. The
original intent was to graphically simulate the
graceful but unpredictable choreography of a bird
flock. Initial simulations were modified to
incorporate nearest-neighbour velocity matching,
eliminate ancillary variables, and incorporate
multidimensional search and acceleration by
distance (Kennedy and Eberhart 1995, Eberhart and
Kennedy 1995). At some point in the evolution of
the algorithm, it was realized that the conceptual
model was, in fact, an optimizer. Through a process
of trial and error, a number of parameters
extraneous to optimization were eliminated from
the algorithm, resulting in the very simple original
implementation (Eberhart, Simpson and Dobbins
1996).
PSO emulates the swarm behaviour and the
individuals represent points in the -dimensional
search space. A particle represents a potential
solution. The velocity V
i
d
and position X
i
d
of the d
th

dimension of the i
th
particle are updated as follows
(4) & (5).

V
i
d
← V
i
d
+ C
1
* rand1
i
d
*(pbest
i
d
- X
i
d
) + C
2
*
rand2
i
d
*(gbest
d
- X
i
d
) (4)

X
i
d
← X
i
d
+ V
i
d
(5)

Where X
i
= (X
i
1
, X
i
2
, … , X
i
D
) is the position of i
th

particle V
i
= (V
i
1
, V
i
2
, … , V
i
D
) represents velocity
of the particle i. pbest =( pbest
i
1
, pbest
i
2
, … ,
pbest
i
D
) is the best previous position yielding the
best fitness value for their i
th
particle; and gbest =
(gbest
i
1
, gbest
i
2
, … , gbest
i
D
) is the best position
discovered by whole population[14]. C
1
and C
2
and
are the acceleration constants reflecting the
weighting of stochastic acceleration terms that pull
each particle toward pbest and gbest positions,
respectively. rand1
i
d
and rand2
i
d
are two random
numbers in the range [0, 1].

3.2. PSEUDOCODE:

1:Generate the initial swarm by randomly
generating the position and velocity for each
particle;
2: Evaluate the fitness of each particle;
3: repeat
4: for each particle i do
5: Update particle i according to (1) and (2);
6: if (x
i
) <f(x
pbesti
) then
7: x
pbesti
:=x
i
;
8: if f (x
i
) <f(x
gbest
) then
9: xgbest := x
i
10: end if
11: end if
12: end for
13: until the stop criterion is satisfied


3.3. ADVANCEMENTS IN PSO AND ITS APPLICATION AND
RESEARCH AREAS:

APPSO (Agent based parallel PSO) is based on two
types of agents: one coordination agent and several
swarm agents. The swarm is composed of various
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-5

sub-swarms, one for each swarm agent. The
coordination agent has administrative and
managing duties. All the calculations are done by
the swarm agents (see Figure 2).
In order to gain benefit from the large knowledge
and insights achieved in the research field of
sequential PSO it is important to modify the
swarm’s behavior as little as possible. The
inevitable changes to the algorithm due to the
parallelization should also lead to positive effects


FIG. 2 STRUTURE OF APPSO NETWORK SYSTEM [14]

with respect to solution quality. This is realized by
using strategically niching where according to the
requirements the sub-swarms either work together
searching for one global optimum or spread over
the search space to find various global/local
optima, e.g. on multimodal functions.
To address the problem of space locus searching, a
slowdown particle swarm optimization (SPSO) is
proposed to improve the convergence performance
of particle swarm from the position viewpoint. The
particle swarm in SPSO is divided into many
independent sub-swarms to guarantee that particles
converge to different position, since space locus
has multiple optimal solutions and requires the
convergence of both fitness and position of particle.
Furthermore, particle velocity is updated by half
according to fitness to achieve the position
convergence [14].
Several PSO algorithms have been recently
proposed to address DOPs of which using multi
swarms seems a good technique. The multi swarm
method can be used to enhance the diversity of the
swarm, with the aim of maintaining multiple
swarms on different peaks.
One of the reasons that particle swarm optimization
is attractive is that there are very few parameters to
adjust. One version, with very slight variations (or
none at all) works well in a wide variety of
applications. Particle swarm optimization has been
used for approaches that can be used across a wide
range of applications, as well as for specific
applications focused on a specific requirement. In
this brief section, we cannot describe all of particle
swarm’s applications, or describe any single
application in detail. Rather, we summarize a small
sample.
Generally speaking, particle swarm optimization,
like the other evolutionary computation algorithms,
can be applied to solve most optimization problems
and problems that can be converted to optimization
problems. Among the application areas with the
most potential are system design, multi-objective
optimization, classification, pattern recognition,
biological system modelling, scheduling
(planning), signal processing, games, robotic
applications, decision making, simulation and
identification. Examples include fuzzy controller
design, job shop scheduling, real time robot path
planning, image segmentation, EEG signal
simulation, speaker verification, time-frequency
analysis, modelling of the spread of antibiotic
resistance, burn diagnosing, gesture recognition
and automatic target detection, to name a few.

4. FIREFLY OPTIMIZATION

Fireflies, also called lighting bugs, are one of the
most special and fascinating creatures in nature.
These nocturnal luminous insects of the beetle
family Lampyridae (order Coleoptera), inhabit
mainly tropical and temperate regions, and their
population is estimated at around 1900 species.
They are capable of producing light thanks to
special photogenic organs situated very close to the
body surface behind a window of translucent
cuticle. Bioluminescent signals are known to serve
as elements of courtship rituals, methods of prey
attraction, social orientation or as a warning signal
to predators (in case of immature firefly forms
commonly referred to as glow worms). The
phenomenon of firefly glowing is an area of
continuous research considering both its
biochemical and social aspects.
Mechanisms of firefly communication via
luminescent flashes and their synchronization has
been imitated effectively in various techniques of
wireless networks design, dynamic market pricing
and mobile robotics.

4.1. CLASSICAL ALGORITHM

The flashing light of fireflies is an amazing sight in
summer sky in the tropical and temperate regions.
There are about two thousand firefly species, and
most fireflies produce short and rhythmic flashes.
The pattern of flashes is often unique for a
particular species. The flashing light is produced by
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-6

a process called bioluminescence. However, two
fundamental functions of such flashes are to attract
mating partners and to attract potential prey. In
addition it may also serve as a protective warning
mechanism to remind potential predators of the
bitter taste of fireflies.
The rhythmic flash, the rate of flashing and the
amount of time form the part of the signal system
that bring both sexes together. We know that the
light intensity at a particular distance r from the
light source obeys the inverse square law. That is to
say, the light intensity I decreases as the distance r
increases in terms if I . Furthermore, the air
absorbs light which becomes weaker as the
distance increases. These two combined factors
make most fireflies visual to a limit distance,
usually several hundred meters at night, which is
good enough for fireflies to communicate.
The flashing light can be formulated in such a way
that it is associated with the objective function to
be optimized, which makes it possible to formulate
new optimization algorithms.
Assumptions:
1. All fireflies are unisex so that one firefly will be
attracted to the other fireflies regardless of their
sex.
2. The brighter one will move towards the less
bright one. The attractiveness is proportional to the
brightness and they both decrease as the distance
increase. And if there is no brighter one than a
particular firefly, it will move randomly;
3. The brightness of a firefly is affected or
determined by the landscape of the objective
function.

4.1.1 LIGHT INTENSITY AND THE ATTRACTIVENESS:

In the firefly algorithm, there are two important
issues: the variation of light intensity and
formulation of the attractiveness. For simplicity, we
can always assume that the attractiveness of a
firefly is determined by its brightness which in turn
is associated with the encoded objective function.
In the simplest case particular location x can be
chosen as I(x) . However attractiveness β is
relative.
In simplest form, the intensity I(r) varies according
to inverse square law
I(r) =
Where I
s
is the intensity at the source. For a given
medium with light absorption coefficient , the light
intensity I varies with the distance r. That is

,

where I
0
is the original light intensity. In order to
avoid the singularity at r=0 in the expression I
s
/r
2
,
combined effect of both the inverse square law and
the absorption can be approximated as the
following Gaussian form.
I(r) =
Thus of the firefly can be defined as

Where β
0
is attractiveness at r=0
The movement of a firefly i is attracted to another
more attractive firefly j is determined by
= + (x
i
- x
j
) +α ,
where second term is due to the attraction. The
third term is randomization with α being the
randomization parameter, and is a vector of
random numbers drawn from a Gaussian
distribution or uniform distribution. For eg. The
simplest form is and can be replaced by
rand(1/2).
The algorithm presented here makes use of a
synergic local search. Each member of the swarm
explores the problem space taking into account
results obtained by others, still applying its own
randomized moves as well. The influence of other
solutions is controlled by value of attractiveness. It
can be adjusted by modifying two parameters: its
maximum value β
0
and an absorption coefficient γ.
The first parameter describes attractiveness at r
j
= 0
i.e. when two fireflies are found at the same point
of search space S. In general β
0
[0; 1] should be
used.

SPECIAL CASES:

For γ→0, the attractiveness is constant β =β
0
, this is
equivalent to say that the light intensity does not
decrease in an idealized sky.
When γ→∞ leads to β(r)→δ(r), which means that
the attractiveness is almost zero in the sight of
other fireflies.

.4. 2. PSEUDOCODE:

1. Input:
f(z), z = [z1; z2; :::; zn]T
m, β
0
, γ, min u
i
, max u
i
(parameters)




2. Output:
X
imin





CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-7

Begin
for i=1 to m do
x
i
← Generate Initial Solution ()
end
3. repeat
i
min
← arg min
i
f(x
i
)
x
i
min
← arg min
xi
f(x
i
)
4. for i=1 to m do
for j=1 to m do
if f(x
j
) < f(x
i
) then
r
j
← Calculate Distance (x
i
, x
j
)
β←β
0
e
-γrj

u
i
← Generate Random Vector (min u
i
, max u
i
)
for k=1 to n do
x
i,k
←(1- β)x
i,k
+ βx
i,k
+ u
i,k
end
end
end
end
5. u
imin
← Generate Random Vector (min u
i
, max u
i
)
for k=1 to n do
x
imin,k
← x
imin,k
+ u
imin,k

end
until stop is condition true
6.end





4.3. ADVANCEMENTS IN FFA AND ITS APPLICATION AND
RESEARCH AREAS:


Many bio-inspired algorithm exist that use simple
rules and local information to create a global
consensus of a single parameter, such as clock time
in the flashing firefly algorithm, or location in the
slime algorithm. In this paper we identify a control
loop as the core structure of several bio-inspired
algorithms. Two ways of combining these loops are
identified, serially, and nested. These methods of
combination can be used to combine algorithms to
allow the control of multiple parameters (one per
algorithm) on a global level without the use of a
centralised point of control.
Thresholding is one of the most important
techniques for performing image segmentation. It is
generally simple and computationally efficient. The
MEFFT algorithm simulates the behaviour of
fireflies and the phenomenon of bioluminescent
communication to develop the algorithm to select
the adequate thresholds for image segmentation.
The segmentation results of MEFFT algorithm are
promising and it encourages further researches for
applying this algorithm to complex and real-time
image analysis problems such as target recognition,
complex document analysis and biomedical image
application [16]. The other application include
synchronisation of wireless network which be seen
in [17]. An application in pulse-coupled oscillator’s
model addresses the issue of synchronization of the
oscillators with different frequencies. As a
continuation of this research, it is important to
study how the topology and communication delay
of the distributed oscillators system affect the
synchronization process.

5. CONCLUSION

While studying and working on these algorithms
we have seen that performance of any technique is
objective dependent. There are various parameters
that define the aptness of the algorithm. So it
depends on the requirements of the objective
function and on the constraints imposed. For
example if we take several benchmark problems
and compare the results of three algorithms then it
has been seen each would perform differently. Not
only their individual performance vary even they
tend to become less or more stable in terms of the
parameters defining the quality of performance
such as elapsed time, convergence rate, optimum
value (maximised/minimized), mean and standard
deviation etc. So depending upon our requirement
we may define our objective function and the
performance of algorithm for that particular
function is based on how aptly the required
objective is received, which could be in terms of
final value, elapsed time, mean etc.
Thus having witnessed these variations one can
generalise the performance on overall terms and
state any of the algorithm as best among the three
taken.

6. REFERENCES

[1] K. M. PASSINO, “Biomimicry of bacterial foraging for
distributed optimization and control,” IEEE Control Syst., vol.
22, no. 3, pp. 52–67, Jun. 2002.
[2]. S. DAS, S. DASGUPTA, A. BISWAS and A. ABRAHAM,
“On Stability of the Chemotactic Dynamics in Bacterial-
Foraging Optimization Algorithm.” IEEE Trans. on Systems,
Man, and cybernetics, part: A: Systems and Humans, vol. 39,
no. 3, pp. 670-679, May 2009.
[3]. B. D. Hughes, Random Walks and Random Environments.
London, U.K.: Oxford Univ. Press, 1996.
[4]. L. DOS SANTOS COELHO, C. DA COSTA SILVEIRA,
C.A. SIERAKOWSKI and P. ALOTTO, “Improved Bacterial
Foraging Strategy Applied to TEAM Workshop Benchmark
Problem.” IEEE Trans. On Magnetics, Vol. 46, no. 8, pp. 2903-
2906, Aug. 2008.
[5] W. J. TANG and Q. H. WU, “Bacterial Foraging Algorithm
For Dynamic Environments.” Proc. of IEEE Congress on
Evolutionary Computation Sheraton Vancouver Wall Centre
Hotel, Vancouver, BC, Canada, pp. 1324-1330, July 16-21,
2006.
[6] R. W. MORRISON and K. A. DE JONG, “A test problem
generator for non stationary environments”, Proc. of IEEE
Congress on Evolutionary Computation, IEEE Press, pp. 2047–
2053, 1999.
[7]. W. J. TAN and Q. H. WU, “A Bacterial Swarming
Algorithm For Global Optimization.” Proc. of IEEE Congress
on Evolutionary Computation, pp. 1207-1212, 2007.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO1O2-8

[8]. A. ABRAHAM, A. BISWAS, S. DASGUPTA AND
SWAGATAM DAS, “Analysis of Reproduction Operator in
Bacterial Foraging Optimization Algorithm.” Proc. of IEEE
Congress on Evolutionary Computation, pp. 1476-1483, 2008.
[9]. YING CHU, HUA MI, HUILIAN LIAO, ZHEN JI, AND
Q. H. WU, “A Fast Bacterial Swarming Algorithm For High-
dimensional Function Optimization.” Proc of IEEE Congress on
evolutionary computation, pp. 3135-3140, 2008.
[10]. DANIEL W. BOERINGER AND DOUGLAS H.
WERNER, “Particle Swarm Optimization Versus Genetic
Algorithms for Phased Array Synthesis.” Trans. Of IEEE on
Antennas and Propagation, vol. 52, no. 3, pp. 771-779,March
2004.
[11] R. EBERHART AND J. KENNEDY, “A new optimizer
using particle swarm theory,” in Proc. 6th Int. Symp. Micro
Machine and Human Science (MHS ’95), pp. 39–43, 1995.
[12] J. KENNEDY AND R. EBERHART, “Particle swarm
optimization,” in Proc. IEEE Int. Conf. Neural Networks, vol. 4,
pp. 1942–1948, 1995.
[13]. J. J. LIANG AND A. K. QIN, “Comprehensive Learning
Particle Swarm Optimizer for Global Optimization of
Multimodal Functions.” Trans of IEEE on Evolutionary
Computation, vol. 10, no. 3, June 2006.
[14]. LI ZHI-JIE, LIU XIANG-DONG, DUAN XIAO-DONG,
WANG CUN-RUI, “An Improved Particle Swarm Algorithm
for Search Optimization.” Proc. of IEEE Global Congress on
Intelligent System, pp.154-158, 2009.
[15]. MICHAEL BREZA, JULIE MCCANN, “Can Fireflies
Gossip and Flock?: The possibility of combining well-knowbio-
inspired algorithms to manage multiple global parameters in
wireless sensor networks without centralised control.
[16]. MING-HUWI HORNG AND TING-WEI JIANG,
“Multilevel Image Thresholding Selection based on the Firefly
Algorithm.” Symposia and Workshops on Ubiquitous,
Autonomic and Trusted Computing, pp. 58-63, 2010.
[17]. LIN CUI, HONGPENG WANG, “Reachback Firefly
Synchronicity with Late Sensitivity Window in Wireless Sensor
Networks.” Proc. of IEEE Ninth International Conference on
Hybrid Intelligent Systems, pp. 451-456, 2009.


BIOMETRIC AUTHENTICATION USING IMAGE PROCESSING TOOLS

Neha Mittal Ruchi Gupta Pratibha Singh
Department of Electronics and communication
J.P Institute of Engineering & Technology, Meerut
Email: nehamittal2008@gmail.com
COPO1O1-1



Abstract
This paper presents the study of biometric
authentication using various digital image
processing techniques and some future
possibilities in this field. The position of
biometrics in the current field of security has
been depicted in this work. For biometrics
recognition the Region of interest (ROI) is
extracted from the biometric modalities.
Various different features are extracted from
ROI. These features can be fuzzy features,
wavelet features, Energy and entropy
features. These sets of features are used for
the authentication of users using Euclidean
Distance, Chi square measure and fuzzy
Support Vector Machines as classifiers.
Recognition score of 82% is obtained with
wavelet feature on palmprint database of
PolyU. We have also outlined the
comparative study and usability of biometric
authentication systems in this paper.

Keywords: wavelet features, multimodal
biometric, ROI, Euclidean Distance, chi-
square,

1. INTRODUCTION
Computer aided personal authentication is
becoming more and more reliable in this
information era. As in order to deal with
security, Authentication plays an important
role. Biometrics is one of the most reliable in
this field. A wide variety of systems require
reliable personal recognition schemes to
either confirm or determine the identity of an
individual requesting their services. The
purpose of such schemes is to ensure that the
rendered services are accessed only by a
genuine user, and not anyone else. By using
biometrics it is possible to confirm or
establish an individual’s identity.
The most widely used biometric feature is the
fingerprint while most reliable one is the
iris[1].Palmprint, as a biometric feature
gaining acceptance in the field of biometric
recognition compared with other modalities:
low cost capture device, easy self positioning,
cannot be easily stolen, large area to get large
information.

In this paper, an authentication of biometrics
is studied. Selection of region of interest,
feature extraction, and classification of these
features is explored.

The organization of the paper is as follows:
Section 2 presents the extraction of ROI. The
extraction of features is described in Section
3. Results are described in Section 4 and
finally multimodal biometrics systems and
conclusions are drawn in Section 5 and 6.

2. ROI EXTRACTION
The useful information content of an image of
any biometric modality can be abstracted by
different image parameters; this is nothing but a
region of interest. ROI is simply a subimage of
interest. An algorithm for abstracting ROI of
Palmprint is given in [8].


BIOMETRIC AUTHENTICATION USING IMAGE PROCESSING TOOLS



COPO1O1-2




Fig. 1: A typical Palmprint Fig. 2: The extracted ROI of
palmprint

3. FEATURE EXTRACTION
The issue of choosing the features to be
extracted should be guided by some concerns
like the features should carry enough
information about the image and should not
require any domain-specific knowledge for
their extraction. They should be easy to
compute in order for the approach to be
feasible for a large image collection and rapid
retrieval.

Finger print matching techniques can be
placed into two categories. One of them is
Minutiae based and the other one is
Correlation based. Minutiae based techniques
find the minutiae points first and then map
their relation placement on the finger.
Correlation based techniques require the
precise location of a registration point and are
affected by image translation and rotation [6,
11,12].

The iris of the eye is colored area surrounded
by the pupil. Iris patterns are unique and are
obtained through video based image
acquisition system. Each iris structure is
featuring a complex pattern. This can be a
combination of specific characteristics known
as corona, crypts, filaments, freckles, pits,
furrows, striations and rings [4].

The palmprints possess random textures. A
palmprint is a rich source of texture. It
consists of zigzag and crisscrossing lines in
addition to three principal lines. Palmprint
also contains creases, wrinkles and ridges.
Texture features can be extracted using
statistical measures (entropy, homogeneity),
wavelet, and fractals.

In [8] the ROI is divided in to non-
overlapping windows. First the membership
function is calculated and from this
membership function different features (i.e.
energy feature, fuzzy feature, Entropy feature)
can be calculated. Some features are as
follows:

3.1 Wavelet Features
Wavelets are immensely suitable for texture
analysis. In [1] and [8] Haar wavelets are
effective on the palmprints. Simplicity of
these wavelets is an overriding factor in the
choice of these wavelets though other
celebrated wavelets such as Daubechies
wavelets, Symlets are also tried but the
improvement accrued from the latter ones is
not proportionate with the effort.

A two-dimensional discrete wavelet transform
is applied on the ROI of a palmprint resulting
in four subbands: HL1, LH1 and HH1, which
correspond to fine-level coefficients and LL1,
which corresponds to coarse-level
coefficients. The HL1, LH1 and HH1together
contribute to a detail image. Each detail
image represents a high frequency structure,
i.e. lines, creases, wrinkles, ridges on a
palmprint. Fig. 3 shows one level of
decomposition. LL1 can be further
decomposed into four sub bands and the
process is repeated till the information from
LLn at the nth decomposition is insignificant.

In any decomposition, the detail images
include horizontal, vertical and diagonal parts
BIOMETRIC AUTHENTICATION USING IMAGE PROCESSING TOOLS



COPO1O1-3



of the decomposed image. These three detail
images are superimposed in [8] to yield
composite mage. It may be noted that as the
decomposition level increases the size of the
detail images also decreases.
The results of authentication correspond to
the composite image.

From each detail images, energy feature is
calculated as [1]:


2
1 1
( ( , )) , 1, 2,..., 5
M N
d
i i
x y
E S x y i

(1)


Fig. 3: Two dimensional one level DWT
decomposition


Fig. 4: Three-level DWT decomposition

where i is the decomposition level. H
i
,V
i
and
D
i
are the detail images in horizontal, vertical
and diagonal directions respectively. Other
features are under investigation.In [8] all the
detail images are super imposed and then
energy is calculated.

4. Results and implementation

In [8] a 100% recognition score is obtained
using fuzzy feature. A simple Euclidean
distance measure is used to find out the
recognition rate. In [8] the database used, is
created in the biometrics lab of IIT Delhi.
ROI is divided in to non-overlapping
windows. Features are calculated from these
windows. The size of the window is varied
and the recognition score is obtained.
Given two data sets of features corresponding
to the training and testing samples, a
matching algorithm determines the degree of
similarity between them. A Euclidean
distance is adopted as a measure of
dissimilarity for the palmprint matching using
both wavelet and fuzzy features.

The wavelet feature in [8] is applied on polyU
database [15] on 50 users and 5 images of
each user (total 250 images).4 images are
taken for training data and 1 image for testing
data. A recognition score of 82% is obtained.
Euclidean distance measure is taken as a
classifier.ROC plot is shown in fig.5

BIOMETRIC AUTHENTICATION USING IMAGE PROCESSING TOOLS



COPO1O1-4




Fig 5: ROC plot for PolyU database

5. MULTIMODAL BIOMETRIC
SYSTEM
Multimodal biometric systems are capable of
utilizing more than one physiological or
behavioral characteristic for authentication. In
order for the biometrics to be ultra-secure and
to provide more-than-average accuracy, more
than one form of biometric identification is
required. Hence the need arises for the use of
multimodal biometrics. This uses a
combination of different biometric
recognition technologies.
In certain situations, the user might find one
form of biometric identification is not exact
enough for identification. This can be the case
with fingerprints, where at least 10% of the
population have worn, cut or unrecognizable
prints.
Multimodal biometric technology uses more
than one biometric identifier to compare the
identity of the person. Therefore in the case of
a system using say three technologies i.e. face
mimic and voice. If one of the technologies is
unable to identify, the system can still use the
other two to accurately identify against.
By using more than one means of biometric
identification, the multimodal biometric
identifier can retain high threshold
recognition settings. The system administrator
can then decide the level of security he/she
requires. For a high security site, they might
require all three biometric identifiers to
recognize the person or for a lower security
site, only one or two of the three. With this
methodology, the probability of accepting an
impostor is greatly reduced.
The unimodal biometric systems can have the
variety of limitations which are as follows
[13]:
(a) Noise in sensed data: A fingerprint
image with a scar or a voice sample
altered by cold, are examples of noisy
data. Noisy data could also result from
defective or improperly maintained
sensors (e.g., accumulation of dirt on a
fingerprint sensor) or unfavorable ambient
conditions (e.g., poor illumination of a
user’s face in a face recognition system).
(b) Intra-class variations: These variations
are typically caused by a user who is
incorrectly interacting with the sensor
(e.g., incorrect facial pose), or when the
characteristics of a sensor are modified
during authentication (e.g., optical versus
solid-state fingerprint sensors).
(c) Inter-class similarities: In a biometric
system comprising of a large number of
users, there may be inter-class similarities
(overlap) in the feature space of multiple
users.
(d) Non-universality: The biometric system
may not be able to acquire meaningful
biometric data from a subset of users. A
fingerprint biometric system, for example,
may extract incorrect minutiae features
BIOMETRIC AUTHENTICATION USING IMAGE PROCESSING TOOLS



COPO1O1-5



from the fingerprints of certain
individuals, due to the poor quality of the
ridges.
(e) Spoof attacks: This type of attack is
especially relevant when behavioral traits
such as signature or voice are used.
However, physical traits such as
fingerprints are also susceptible to spoof
attacks. These systems are able to meet
the rigorous performance requirements
imposed by various applications. They
address the problem of non-universality,
since multiple traits ensure sufficient
population coverage. They also deter
spoofing since it would be difficult for an
impostor to spoof multiple biometric traits
of a genuine user simultaneously.



Fig 6: A multimodal biometric system

Limitations of unimodal biometrics are
overcome by the multimodal biometric
system.

6. CONCLUSIONS
Biometric recognition can offer a high degree
of security, they are far from perfect solution.
As the fingerprint can give the high accuracy
of recognition but it have draw backs as
discussed. So one solution to this problem is
to integrate two or more modalities, to
provide high security.
Multimodal biometric system also prevent
from spoofing as it is difficult for an impostor
to spoof multiple biometric modalities of a
genuine user simultaneously.

References
[1] Xiang-Qian Wu, Kuan-Quan Wang, David
Zhang, ”Wavelet Based Palmprint
Recognition”, First international
conference on Machine learning
andcybernetics, Beijing, pp-1253-1257, 4-
5 Nov 2002.
[2] R. C. Gonzalez, R. E. Woods, Digital
Image Processing, Addison Wesley
publishers, 1993.
[3] L. Hong and A. K. Jain, “Integrating Faces
and Fingerprints for Personal
Identification”, IEEE Trans. on Pattern
Analysis and Machine
Intelligence, Vol. 20, No. 12, pp. 1295-
1307, Dec1998.
[4] Sanjay R. Ganorkar, Ashok A. Ghatol,
“Iris Recognition: An Emerging Biometric
Technology”, In Proc. of the 6th WSEAS
International Conference on Signal
Processing, Robotics and Automation,
Greece, Feb. 2007, pp. 91 – 96.
[5] N. Duta, A.K. Jain, and K.V. Mardia,
“Matching of Palmprint,” Pattern
Recognition Letters, Vol.23, No. 4, pp.
477–485, 2001.
[6] David Zhang, Wai-Kin Kong, Jane You,
and Michael Wong, “Online Palmprint
Identification”, IEEE Trans. Pattern
Analysis &Machine Intelligence, Vol. 25,
No. 9, pp. 1041 – 1050,September 2003.
[7] Ying-Han Pang, Andrew Teoh Beng Jin,
David Ngo Chek Ling, "Palmprint based
Cancelable Biometric Authentication
BIOMETRIC AUTHENTICATION USING IMAGE PROCESSING TOOLS



COPO1O1-6



System", Int. J. of Signal Processing,
Vol.1, No.2, pp.93-99, 2004.
[8] M. Hanmandlu, H M Gupta, Neha Mittal,
S. Vasikarla,”An authentication system
based on palmprint”,ITNG-2009,pp-399-
404
[9] Lei Zhang, ZhenhuaGuo, Zhou Wang,
DavidZhang, "Palmprint Verification
Using Complex Wavelet Transform", IEEE
Intl. Conf. on ImageProcessing, San
Antonio, TX, Sept.16-19,2007.
[10] Li Fang, Maylor K.H. Leung , Tejas
Shikhare, Victor Chan, Kean Fatt Choon ,
“Palmprint Classification”, IEEE Int. Conf.
on Systems, Manand Cybernetics, Vol. 4,
8-11 October 2006, pp.2965-2969.
[11] L. Hong and A. K. Jain, “Integrating
faces and fingerprints for personal
identification”, IEEE Trans. Pattern
Anal. Mach. Intell., Volume 20, No. 12,
Dec. 1998, pp. 1295–1307.
[12] R. Cappelli, D. Maio, D. Maltoni, J. L.
Wayman, and A. K. Jain, “Performance
evaluation of fingerprint verification
systems”, IEEE Trans. Pattern Anal. Mach.
Intell., Volume 28, issue 1, Jan. 2006, pp.
3–18.
[13] M. Indovina, U. Uludag, R. Snelick,
A. Mink, and A. K. Jain, “Multimodal
biometric authentication methods: A
COTS approach,” in Proc. of Workshop on
Multimodal User Authentication, (Santa
Barbara, CA), pp. 99–106, Dec 2003.
[14] Wen-Shiung Chen, Yao-Shan Chiang
and Yen-Hsun Chiu, ”Biometric
Verification by Fusing Hand Geometry and
Palmprint”, IIH-MSP 2007
[15] H. K. Polytechnic University,
“Palmprint database”, Biometric Research
Center Website. 2005.
http://www4.comp.polyu.edu.hk/~biometri
cs/.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-1
Medical Image Enhancement In Spatial Domain: Statistical Analysis
Sanjeev Kumar Gupta
1
, Sonika Singh
2
& Rahul Singh
3

1
Assistant Professor ,
2,3
B.Tech Scholar, ECE Department, Faculty of Engineering and
Technology, R.B.S. College, Bichpuri Campus, Agra
ABSTRACT
Image quality and utility become crucial issues for engineers, scientist, doctors, patients,
insurance companies and lawyer whenever there are changes in the technology by which
medical images are achieved. Examples of such changes include analog to digital conversion,
lossy compression for transmission and storage, image enhancement, and computer aided
methodology for diagnosis of disease in medical images. Edit an image so that it is more suitable
for a specific application than the original image is termed as image enhancement technique.
Image is defined as a two dimensional function f(x, y), where x and y are spatial coordinates,
that bears information, which can be generated in any form such as visual, x-ray and so on. X-
rays are the oldest source of electromagnetic radiation used for medical imaging. Medical image
enhancement methods are used like all other methods and algorithm in image processing, as
chain subsequent edits aimed at achieving a suitable result. Improving one function in the chain
is only useful if the end result is really unproved, and that does not solely depend on that
particular function; it also depends on the quality of the first image. In this paper we have
compared different types of image enhancement technique of medical image in spatial domain
and also presented a statistical analysis.
Medical Image Enhancement
Digital image processing is a technique
using which the data from an image is
digitized. With the help of computer
algorithms, various mathematical operations
are applied to the data to create an enhanced
image. In medical image enhancement we
use the input, which is the form of X-ray
film. In the input image we perform some
mathematical operation and the gray values
of the input image are mapped to another set
of gray values, so that the image is made
suitable for specific application. it is called
Image Enhancement method.
Image enhancement technique can be
divided in two broad categories spatial
domain methods, which operate directly on
pixel and frequency domain methods, which
operate directly on the Fourier, transform of
an image. In this present paper we focus our
attention on spatial domain. There are
several enhancements in spatial domain such
as-
i. Gray Scale manipulation.
ii. Histogram equalization.
iii. Image smoothing-Neighborhood
averaging and Edge preserving
smoothing.

The term spatial domain refers to the
aggregate of pixels composing an image.
These methods are procedures that operate
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-2
directly on these pixels. Spatial domain
process will be denoted by the expression
g(x>y) =T[f(x, y)] (1)
Where f(x, y) is the input image, g(x, y) is
the processed image, and T is an operator on
f, defined over some neighborhood about a
(x, y). The principle approach in defining a
neighborhood about a point (x, y) is to use a
square or rectangular sub image area
centered at (x, y), as shown in fig.1.

Fig.1.
The center of the sub image is moved from
pixel to pixel starting from the top left
corner to the bottom most right corner. The
operator T is applied at each location (x, y)
to yield the output at that location. The
process utilizes only the pixel hi the area of
the image spanned by the neighborhood.
Although other neighborhood shapes, such
as approximation to a circle, square and
rectangular arrays are by far the most
predominant because of their ease of
implementation. The simplest form of T is a
single pixel when the neighborhood is of
size 1X1.In this case, g depends only on the
value of fat(x, y), and t becomes a gray-level
(also called intensity or mapping)
transformation function of the form:
s=T(r) (2)
Where, r and s are variables denoting, the
gray level of f(x, y) and g(x, y) at any
point(x, y) respectively.
The effect of thresholding would be to
produce an image of higher contrast than
original by darkening the levels below m
(threshold) and brightening the level above
in the original image, this technique is
known as contrast stretching.
The gray scale modification method
includes Log transformation, Power law
transformation and Piecewise transformation
and Histogram equalization method include
local histogram equalization and global
histogram equalization. The third method
and many more methods are possible but we
have not discussed in this paper. Gray-level
transformation is very simple and powerful
processing method of radiographic image
enhancement. The log transformation is used
to compress dynamic range. In this
transformation function large numbers of
gray level values are mapped in to narrow
range. The general form of log
transformation is:
S=c log(1+r) (3)
Where c is a constant and it is assumed that
r>=o
Power-law transformation is used to expand
a certain range of intensity values. In these
methods the exponent of the power-law
equation is called gamma f. By varying
gamma, a family of power-law
transformation can be obtained. Power law
transformations have the basic form:
S=cr
γ
(4)
Where c and γ are positive constants.
The simplest form of Piecewise linear
function is a contrast stretching
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-3
transformation. Low-contrast image can
result from poor illumination, lack of
dynamic range in the imaging sensor, or
even wrong setting of lens aperture during
image acquisition. The idea behind contrast
stretching is to increase the dynamic range
of the gray levels in the image being
processed.
Global histogram equalization method is a
simple and computationally inexpensive
implementation, but it treats the entire
content image uniformly, regardless of the
local variation of gray levels. This may lead
to loss of visual quality of images for certain
areas and increased visibility of noise.
Discrete domain
S
k
= T(r
k
)= = (5)
Analog Domain
s= T(r) = (6)
Local histogram equalization enhances the
small region in the image features. A square
or rectangular neighborhood is defined and
the centre of this area from pixel to pixel is
moved. At each location, the histogram of
die points in the neighborhood is computed
and either a histogram equalization or
histogram specification transformation
function is obtained. Fig.3 shows various
outputs.
Medical X-ray images have narrow latitude
or dynamic range. The film’s ability to
record and view all the anatomy within the
x-ray field is limited by this narrow dynamic
range. Image enhancement is the process of
applying these techniques to facilitate the
development of a solution to a computer
imaging problem. Consequently, the
enhancement methods usually are
application specific and are often developed
empirically. Actuality of image
enhancement problem conditioned by
necessity of it’s solving in medical
diagnostics, pattern recognition,
enhancement amateur photography and etc.
Since radiation oncologists believe that final
images have a higher contrast, and from the
statistical assessment it can be concluded
that an increase of average numbers and
deviation of images value will increase the
contrast. As a result processed images via
this program will lead to have a higher
diagnostic ability.
Result and discussion
Here we have presented a statistical analysis
in table form. The variation of the mean
value and standard deviation value of input
radiographic images and enhanced
radiographic images are calculated in this
present paper. We have worked on many
radiographic images but only three images
chest, hand and leg are shown in this paper.
In global histogram equalization method
enhances the whole area of image features.
Therefore for achieving uniform distribution
of gray scale values for histogram
equalization the intensity are increased. This
raises the mean value of enhanced image.
With the increase in mean value, the average
distribution of intensity values also
increases. Consequently the standard
deviation increases.
In local histogram equalization method,
main focused area is selected regions
determined by window. The entire image
need not be enhanced. So, mean values and
standard deviation values of the enhanced
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-4
images are not affected as much as that of
global histogram method. If window size
increases then proportionately larger region
is chosen out of entire image region causing
mean values to decrease. If value of constant
k changes or increases, more uniform
distribution is seen in the histogram it
indicates that more equalization is achieved.
It is observed that in log transformation
method for different value of c, if gray value
increases then statistical values (mean,
standard deviation) of enhanced image also
increases which in turn increases the image
brightness. Similarly in power law
transformation using fixed values of gamma
and different value of c, if the gray value is
increased, the statistical values are also
increased proportionately. In power law
transformation with gamma correction
approach, if the value of gamma increases,
the value of mean of enhanced image
decreases while the value of standard
deviation increases, as result the brightness
of enhanced image is decreased.
Similar observation have been found in
piecewise contrast stretching method i.e.
enhanced image is dark for gamma=0.1 and
contrast increase as the value of gamma
increases from gamma-0.2 onwards.
Table 1, 2, 3 shows the comparison of
different method of Radiographic Image
Enhancement.
Because of flexibility of changing window
size and equalization factor k an auxiliary
region can be enhanced to the extent
required in the local histogram equalization
method. Therefore this method is also
sometimes referred to as adaptive method.
In the present enhancement approaches,
power law transformation using gamma
correction performs much better as
compared to all other methods addressed.
This conclusion is on the basis of the change
in standard deviation and mean value on the
histogram of the enhanced image.
Figure 4, 5, 6 shows plot of mean value and
standard deviation value of four methods of
Leg1, Hand1, and chest 1 image. In bar
graph the global histogram method is not
shown.
Conclusion
In this present paper we have compared
different techniques of image enhancement
in spatial domain. They have been evaluated
in terms of standard deviation, mean values
as statistical measures. As medical image
has only a finite number of gray scales, an
ideal equalization is not possible. Image
enhancement technique can be improved if
the enhancement criteria are application
dependant and can be illustrated precisely.
References
Millener AE, AubinM, Pouliot J,
ShinoharaK, Roach HIM (2004)
“Daily electronic portal imaging
from morbidly obes men undergoing
radiotherapy for localized prostate
cancer”. Int J Radiat Oncol
Bioiphys, 6.10.
Van Denberg DL, De Ridder M,
Storm GA (2004) “Imaging in
radiotherapy” European Journal of
radiology, 41-48.
Shi X.Q., Sallstrom P, and Welander
U. “A Color Coding Method for
Radiographic Image and vision
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-5
computing” Vol.20, pp:761-
767.20Q2.
Gonzalez RC and Woods R.E.
“Digital Image Processing”,
Pearson Education Singapore,
Fourth Indian Reprint, 2
nd
edition
2003.
Jain A. “ fundamentals of Digital
Image
Processing”, Englewood Cli.s
Printice Hall, 1989.
Pratt W. “Digital Image
Processing”, New York, Wiley
Interscience, 1991.
Andrews H.C., Monochrome
“Digital Image Enhancement”,
Appl., p.495, 1976.
Gonzalez RC and Woods R.E., and
Eddins SL. “Digital Image
Processing using MATLAB”.
Pearson Edition.
R.S.Gaborskl Kodak Health Imaging
Syst. Inc. Rochester, NY, USA,
…BK_ Jang, Kodak Health Imaging
Syst. Inc. Rochester, NY, USA,
“enhanceme… for Computed
Radiographic Images”.
Joung-youn Kirn; Lee-Sup Kim;
Seung-Ho Hwang. “An advanced
contrast enhancement using partially
overlapped subblock histogram
equalization” Circuits and
Systems.2000. Proceedings. ISCAS
2000. Geneva.
Matuszewski Bogdan J.. Moore
Christopher J.. “Deformable Models
in Medical Image Processing,
Advances in Image Guided
Radiotherapy”. Euspico 2007,
Pznon, Poland, Sep2007.
S.S.Y. Lau, „Global image
enhancement using local
information”. Electronics letters. Pp.
122-123, January 1994.
Jha Vandana, Thakur Kavita and
Sinha G.R. “Image Enhancement
Techniques in spatial domain”. A
case study, J. of Acta Ciecia Indica,
Vol.XXXHI p, No.2, pp:201, 2007.
JohnB. Zimmerman. et al., “An
evaluation of efectiveness of adaptive
histogram equalization for cantrast
enhancement”. IEEE Trans. On
Medical Imaging, pp.304-312.
Deember 1988.
www.wikipedia.com.
http://www.freepatentsonline.com/E
P 1397780.html.











CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-6


Figure 3: Original Image and Enhanced
images using different Methods

Original Image

Global Histogram Equalized

Adaptive Histogram Equalized

Log Transformed

Power Law Transform γ=1

Power Law Transform γ=0.9

Power Law Transform γ=0.8
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-7

Power Law Transform γ=0.7

Figure-4: Mean values and Standard
Deviation of Medical Image Enhancement
for Leg


Figure-5: Mean values and Standard
Deviation of Medical Image Enhancement
for Hand


Figure-6: Mean values and Standard
Deviation of Medical Image Enhancement
for Chest






CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-8


Table 1: Comparison of different method of medical image enhancement for chest





















Table 2: Comparison of different method of medical image enhancement for Hand

Method
Chest
Input Image Enhanced Image
Mean Std Mean Std
1.Global Histogram
Equalization
116.0488 93.1951 127.3704 74.9504
2.Adaptive histogram
equalization
116.0488 93.1951 114.8210 84.0122
3.Log transformation 116.0488 93.1951 180.8969 77.4127
4.Power law
transformation γ=1
116.0488 93.1951 116.0488 93.1951
5. γ =0.9 116.0488 93.1951 120.6711 93.2783
6. γ =0.8 116.0488 93.1951 125.9139 93.0551
7. γ =0.7 116.0488 93.1951 132.0142 92.2453

Method
Hand
Input Image Enhanced Image
Mean Std Mean Std
1.Global Histogram
Equalization
68.9736 80.6995 127.6161 74.2843
2.Adaptive histogram
equalization
68.9736 80.6995 72.7664 72.7077
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO1O5-9
















Table 3: Comparison of different method of medical image enhancement for Leg



3.Log transformation 68.9736 80.6995 157.7583 60.5003
4.Power law
transformation γ =1
68.9736 80.6995 68.9736 80.6995
5. γ =0.9 68.9736 80.6995 73.9827 81.1636
6. γ =0.8 68.9736 80.6995 80.1207 81.0990
7. γ =0.7 68.9736 80.6995 87.7081 80.3351

Method
Leg
Input Image Enhanced Image
Mean Std Mean Std
1.Global Histogram
Equalization
60.4228 70.2646 149.6279 50.3502
2.Adaptive histogram
equalization
60.4228 70.2646 72.8500 73.9090
3.Log transformation 60.4228 70.2646 170.7315 108.0530
4.Power law
transformation γ =1
60.4228 70.2646 60.4228 70.2646
5. γ =0.9 60.4228 70.2646 64.7082 73.3188
6. γ =0.8 60.4228 70.2646 69.4427 76.5475
7. γ =0.7 60.4228 70.2646 74.7634 79.9895
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-1

AN INTELLIGENT ALGORITHM FOR TASK ALLOCATION IN DISTRIBUTED
NETWORK USING STATIC APPROACH

1
Kapil Govil,
2
Dr. Avanish Kumar and
3
Rakesh Kumar Dwivedi

1,3
Teerthanker Mahaveer University, Moradabad – 244 001, (UP), INDIA
2
Bundelkhand University, Jhansi – 284 126, (UP), INDIA

1
drkapilgovil@gmail.com,
2
dravanishkumar@yahoo.com,
3
r_dwivedi2000@yahoo.com


ABSTRACT

Distributed Network is one the appropriate
network for providing the optimal solution in
real-time applications problems. The
systematic allocation of tasks needs to plays
the key role to optimize the overall processing
cost of the distributed network. Distributed
network consists of multiple
autonomous computers that communicate
through a common communication channel.
The computers interact with each other in
order to achieve a common goal. A computer
program that runs in a distributed system is
called a distributed program, and distributed
programming is the process of writing such
programs. The allocation problems in any
computer system play the key role for
deciding the performance of the system. The
allocation put the direct impact of software
resources as well as hardware resources. In
distributed network, partitioning of the
application software in to module and the
proper allocation of these modules dissimilar
processors are important factors, which
determine the efficient utilization of resources.
The static model discussed in this paper
provide an optimal solution using algorithm of
Assignment problem for assigning a set of
“m” tasks of a task to a set of “n” processors
where m > n in a computer system for
evaluation for optimal processing cost of the
system. The several sets of input data are
considered to test the efficiency of the
algorithm and it is found that the algorithm is
suitable for arbitrary number of processor in
all the cases.


Keywords: Distributed Network, Processor,
Task, Allocation, Processing Cost.

INTRODUCTION

Distributed network refers to the means by
which a single computer program runs in more
than one computer at the same time. In
particular, the different elements and objects
of a program are being run or processed using
different computer processors. In a distributed
networking, the program runs like it would in
a single computer even when it is, in fact,
using different computer processors. This
means that no single computer carries the
entire burden on system resources that running
a computer program usually implies.

One of the major research problems for
distributed networks is the allocation problem,
in which tasks are assigned to various
processors of the network, in such a way that
processing cost is to be minimized as per the
requirement. These problems may be
categorized as static (AVAN2001,
KUMA1995a, KUMA1995b, KWON2006)
and dynamic (JACQ2005, AVAN1999,
KUMA1996, KWON2006, PALM2005) types
of task allocation. Some of the other related
methods have been reported in the literature,
such as, Integer Programming (DESS1980,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-2

BRIA2003, MISR1991, MUHA2002), Load
Balancing (JACQ2005, DANI2005,
SAEE2005, SHU2009) and Modeling
(RENE2002, JAVI2000, KENT2002). Tasks
are allocated to various processors of the
distributed network in such a way that overall
processing cost of the network should be
minimized. As it is well known that the tasks
are more than the number of processors of the
network.
OBJECTIVE

In the Distributed Processing Environment
(DPE), it is the common problem to allocate
tasks where the number of tasks is more than
the number of processors. The objective of the
present research paper is to enhance the
performance of the distributed networks by
using the proper utilization of its processors
and as well as proper allocation of tasks. In
the present research paper the type of
allocation of task to the processor is static in
nature. As in this paper the performance is
measured in terms of processing cost, so we
have to minimize the processing cost to obtain
the best performance of the processors. To
overcome from the problem we have designed
an intelligent algorithm for task allocation.

TECHNIQUE

In order to evaluate the overall optimal
processing cost of a distributed network, we
have chosen the problem where a set P = {p
1
,
p
2
, p
3
, …….p
n
} of „n‟ processors and a set T =
{t
1
, t
2
, t
3
, …….t
m
} of „m‟ tasks, where m>n.
The processing cost of each task to each and
every processor is known and it is mentioned
in the Processing Cost Matrix of order m x n.
After making a matrix of same order taking in
ascending order of its sum of row and sum of
column, we apply the algorithm of assignment
problem on it. For each processor we evaluate
the overall allocation of each task; and
allocation of the task on the processor which
has the minimum processing cost. Finally we
compute total processing cost by adding total
processing cost of task, which are assigned at
specified processor.

ALGORITHM

Start Algorithm
Read the number of task in m
Read the number of processor in n
For i = 1 to m
For j = 1 to n
Read the value of
processing cost (c) in
Processor Cost Matrix
namely PCM(,)
j = j + 1
Endfor
i = i + 1
Endfor
Calculate the sum of each row and
column and store the results in
Modified Processor Task Matrix
MPTM(,)
By arranging the MPTM(,) in
ascending order of their row_sum and
column_sum, we get Arranged
Processor Task Matrix APTM(,)
i = 1
While all tasks != SELECTED
Select the biggest possible
square matrix from left upper
corner and store it into SM
i
(,)
Apply algorithm of Assignment
Problem [KANT2002] on SM
i
(,)
i = i + 1
Endwhile
Club processor-wise overall optimal
processing cost
State the results
End Algorithm

IMPLEMENTATION

In the present research paper, the distributed
network consist a set P of 4 processors {p
1
, p
2
,
p
3
, p
4
} and a set T of 10 tasks {t
1
, t
2
, t
3
, t
4
, t
5
,
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-3

t
6
, t
7
, t
8
, t
9
, t
10
}. It is shown in the figure1. The
processing cost (c) of each task to each and
every processor is known and it is mentioned
in the Processor Cost Matrix PCM(, ) of order
10 x 4.

t1 t2
t4 t5 t6
t8
t3
t9 t10
t7
t1 t2
t4 t5 t6
t8
t3
t9 t10
t7
t1 t2
t4 t5 t6
t8
t3
t9 t10
t7
t1 t2
t4 t5 t6
t8
t3
t9 t10
t7
p1 p2
p3 p4

Figure1: Processor Task Combination

4 3 2 1
10
9
8
7
6
5
4
3
2
1
p p p p
200 400 700 300
200 600 500 400
900 800 700 600
700 600 800 200
900 800 100 200
200 600 500 400
300 600 400 100
400 300 700 800
100 200 300 500
300 900 200 600
t
t
t
t
t
t
t
t
t
t
) PCM(,

Now, calculate the sum of each row and
column and store the results in Modified
Processor Task Matrix MPTM(,) of order 10 x
4.


Row_Sum
1600
1700
3000
2300
2000
1700
1400
2200
1100
2000
p p p p
200 400 700 300
200 600 500 400
900 800 700 600
700 600 800 200
900 800 100 200
200 600 500 400
300 600 400 100
400 300 700 800
100 200 300 500
300 900 200 600
Column_Sum
t
t
t
t
t
t
t
t
t
t
) MPTM(,
4200 5800 4900 4100
4 3 2 1
10
9
8
7
6
5
4
3
2
1


By arranging the MPTM(,) in ascending order
of their row_sum and column_sum, we get
Arranged Processor Task Matrix APTM(,) of
order 10 x 4.

Row_Sum
3000
2300
2200
2000
2000
1700
1700
1600
1400
1100
p p p p
800 700 900 600
600 800 700 200
300 700 400 800
800 100 900 200
900 200 300 600
600 500 200 400
600 500 200 400
400 700 200 300
600 400 300 100
200 300 100 500
Column_Sum
t
t
t
t
t
t
t
t
t
t
) APTM(,
5800 4900 4200 4100
3 2 4 1
8
7
3
6
1
9
5
10
4
2

Now, selecting the biggest possible square
matrix from left upper corner, we get Selected
Matrix SM
i
(,) of order 4 x 4.
3 2 4 1
5
10
4
2
1
p p p p
600 500 200 400
400 700 200 300
600 400 300 100
200 300 100 500
t
t
t
t
) (, SM
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-4

Now, by using algorithm of Assignment
Problem [KANT2002] on SM
i
(,), we get task
allocation as –
Processor Task
p
1
t
4

p
2
t
2

p
3
t
10

p
4
t
5


Now, we eliminate allocated tasks (i.e. t
2
, t
4
,
t
5
, t
10
) from APTM(,); and again selecting the
biggest possible square matrix from left upper
corner, we get selected matrix SM
2
(,) of order
4 x 4.
3 2 4 1
3
6
1
9
1
p p p p
300 700 400 800
800 100 900 200
900 200 300 600
600 500 200 400
t
t
t
t
) (, SM

Now, on applying the algorithm of
Assignment Problem [KANT2002] on matrix
SM
2
(,), we get task allocation as –
Processor Task
p
1
t
6

p
2
t
1

p
3
t
3

p
4
t
9


Again, we eliminate allocate tasks (i.e. t
1
, t
3
,
t
6
, t
9
) from APTM(,) and again selecting the
biggest possible square matrix from left upper
corner, we get Selected Matrix SM
3
(,) of order
2 x 2.
2 1
8
7
3
p p
9 6
7 2
t
t
) (, SM

Now, on applying the algorithm of
Assignment Problem [KANT2002] on matrix
SM
3
(,), we get task allocation as –




Processor Task
p
1
t
7

p
2
-
p
3
-
p
4
t
8


The overall processor-wise task allocation
along with optimal processing cost is
mentioned in table1.

Table 1: Processor-wise Task Allocation
Processor Assigned Task Processing Cost
p
1
t
4
*t
6
*t
7
500
p
2
t
1
*t
2
500
p
3
t
3
*t
10
700
p
4
t
5
*t
8
*t
9
1300
Optimal Processing Cost 3000

The graphical representation of the optimal
allocation is shown in figure 2.
t
4
t
6
t
7
t
1
t
2
t
5
t
8
t
9
t
3
t
10
p
1 p
2
p
3 p
4

Figure 2: Optimal Allocation

The processor-wise processing cost graph is
shown in the figure 3.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-5

Processing Cost
0
200
400
600
800
1000
1200
1400
t4*t6*t7 t1*t2 t3*t10 t5*t8*t9
p1 p2 p3 p4
Processing Cost

Figure 3: Processor-wise Processing Cost
Graph

CONCLUSION

In this research paper we have chosen the
problem, in which the number of the tasks is
more than the number of processors of the
distributed network. The model mentioned in
this paper is based on the consideration of
processing cost of the tasks to various
processors. The method is presented in pseudo
code and implemented on the several sets of
input data to test the performance and
effectiveness of the pseudo code. It is the
common requirement for any allocation
problem that the tasks have to be processed
with minimum processing cost. Here,
performance is measured in terms of
processing cost of the task that has been
processed by the processors of the network
and also these tasks have been processed
optimally. As we know that, the analysis of an
algorithm is mainly focuses on time
complexity. Time complexity is a function of
input size „n‟. It is referred to as the amount of
time required by an algorithm to run to
completion. The time complexity of the above
mentioned algorithm is O(mn). By taking
several input examples, the above algorithm
returns following results as in table 2.





Table 2: Optimal Results
No. of
processors (n)
No. of tasks
(m)
Optimal
Results
3 4 12
3 5 15
3 6 18
3 7 21
3 8 24
4 5 20
4 6 24
4 7 28
4 8 32
4 9 36
5 6 30
5 7 35
5 8 40
5 9 45
5 10 50

The graphical representation of results of table
2 are shown by figure 4, 5 and 6 as mentioned
below,
No. of Processors = 3
0
1
2
3
4
5
6
7
8
1 2 3 4
Example
N
o
.

o
f

t
a
s
k
s
0
5
10
15
20
25
m
Complexity

Figure 4: Complexity Graph

No. of Processors = 4
0
1
2
3
4
5
6
7
8
9
1 2 3 4
Example
N
o
.

o
f

t
a
s
k
s
0
5
10
15
20
25
30
35
m
Complexity

Figure 5: Complexity Graph

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-6

No. of Processors = 5
0
1
2
3
4
5
6
7
8
9
10
1 2 3 4
Example
N
o
.

o
f

t
a
s
k
s
0
5
10
15
20
25
30
35
40
45
50
m
Complexity

Figure 6: Complexity Graph

The performance of the algorithm is compared
with the algorithm suggested by Sagar et. al.
[SAGA1991]. Table 3, shows the time
complexity comparison between algorithm
[SAGA1991] and present algorithm.

Table 3: Comparison Table
n m
Time
Complexity
of algorithm
(SAGA1991)
O(m
2
n)
Time
Complexity
of present
algorithm
O(mn)
3 4 48 12
3 5 75 15
3 6 108 18
3 7 147 21
3 8 192 24
4 5 100 20
4 6 144 24
4 7 196 28
4 8 256 32
4 9 324 36
5 6 180 30
5 7 245 35
5 8 320 40
5 9 405 45
5 10 500 50

From the table 3, it is clear that present
algorithm is much better for optimal allocation
of tasks that upgrade the performance of
distributed network. Figures 7, 8 and 9 show
the pictorial representation between algorithm
SAGA1991 and present algorithm.
No. of Processors = 3
0
50
100
150
200
250
1 2 3 4 5 6
AlgorithmSAGA1991
Present algorithm

Figure 7: Comparison Graph
No. of Processors = 4
0
50
100
150
200
250
300
350
1 2 3 4 5 6 7
Algorithm SAGA1991
Present algorithm

Figure 8: Comparison Graph
No. of Processors = 5
0
100
200
300
400
500
600
1 2 3 4 5 6 7
Algorithm SAGA1991
Present algorithm

Figure 9: Comparison Graph

REFERENCES

AVAN1999 Avanish Kumar,
Optimizing for the Dynamic
Task Allocation,
Proceedings of the „III
Conference of the
International Academy of
Physical Sciences,
Allahabad; (1999), 281-
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-7

294.
AVAN2001 Avanish Kumar, An
Algorithm for Optimal
Index to Tasks Allocation
Based on Reliability and
cost, Proceedings of
„International Conference
on Mathematical Modeling,
Roorkee; (2001), 150-155.
BRIA2003 Brian Ensink, Joel Stanley,
and Vikram Adve,.
Program Control
Language: a programming
language for adaptive
distributed applications;
Elsevier Inc., Vol.
63(2003), No. 12, 1082 –
1104.
DANI2005 Daniel Grosu, and Anthony
T. Chronopoulos,
Noncooperative load
balancing in distributed
systems; Elsevier Inc., Vol.
65(2005), No. 9, 1022-
1034.
DESS1980 O.I. Dessoukiu-EI, and W.
H. Huna, Distributed
Enumeration on Network
Computers; IEEE
Transactions on Computer,
Vol. 29(1980), 818-825.
JACQ2005 Jacques Bahi, Raphael
Couturier, and Flavien
Vernier, Synchronous
distributed load balancing
on dynamic networks.
Elsevier Inc., Vol.
65(2005), No. 11, 1397-
1405.
JAVI2000 Javier Contreras, Arturo
Losi, Mario Russo, and
Felix F. Wu, DistOpt: A
Software Framework for
Modeling and Evaluating
Optimization Problem
Solutions in Distributed
Environments, Elsevier
Inc., Vol. 60(2000) No. 6,
741 – 763.
KANT2002 Kant Swaroop, P. K. Gupta
and Man Mohan,
Operations Research,
Sultan Chand & Sons,
ISBN No.: 8180540200,
2002.
KENT2002 Kent Fitzgerald, Shahram
Latifi, and Pradip K.
Srimani, Reliability
Modeling and Assessment
of the Star-Graph
Networks, IEEE
Transactions on Reliability,
Vol. 51(2002), 49-57.
KUMA1995a A. Kumar, M. P. Singh and
P. K. Yadav, An Efficient
Algorithm for Allocating
Tasks to Processors in a
Distributed System,
Proceedings of the „19th
National System
Conference, SSI,
Coimbatore, (1995), 82-87.
KUMA1995b A. Kumar, M. P. Singh, and
P. K. Yadav, A Fast
Algorithm for Allocating
Tasks in Distributed
Processing System,
Proceedings of the „30th
Annual Convention of CSI,
Hyderabad, (1995), 347-
358.
KUMA1996 A. Kumar, M. P. Singh, and
P. K. Yadav, An Efficient
Algorithm for Multi-
processor Scheduling with
Dynamic Reassignment,
Proceedings of the „6th
National seminar on
theoretical Computer
Science, Banasthally
Vidyapeeth, (1996), 105-
118.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O1-8

KWON2006 Yu-Kwong Kwok, Anthony
A. Maciejewski, Howard
Jay Siegel, Ishfaq Ahmad,
and Arif Ghafoor, A semi-
static approach to mapping
dynamic iterative tasks onto
heterogeneous computing
systems. Elsevier Inc., Vol.
66(2006) No. 1, 77-98.
MISR1991 K. B. Misra, and U.
Sharma, An Efficient
Algorithm to solve Integer
Programming Problem
arising in System
Reliability Design, IEEE
Transactions on Reliability,
Vol. 40(1991), 81-91.
MUHA2002 Muhammad K. Dhodhi,
Imtiaz Ahmad, Anwar
Yatama and Ishfaq Ahmad.
An Integrated Technique
for Task Matching and
Scheduling onto
Distributed Heterogeneous
Computing Systems
Elsevier Inc., Vol. 62(2002)
No. 9, 1338 – 1361.
PALM2005 J. Palmer, and I. Mitrani,
Optimal and heuristic
policies for dynamic server
allocation. Elsevier Inc.,
Vol. 65(2005) No. 10,
1204-1211.
RENE2002 Rene L. Bierbaum, Thomas
D. Brown, and Thomas J.
Kerschen, Model-Based
Reliability Analysis, IEEE
Transactions on Reliability,
Vol. 51(2002), 133-140.
SAEE2005 Saeed Iqbal, and Graham F.
Carey, Performance
analysis of dynamic load
balancing algorithms with
variable number of
processors. Elsevier Inc.,
Vol. 65(2005), No. 8, 934-
948.
SAGA1991 G. Sagar and A. K. Sarje,
Task Allocation Model for
Distributed System,
International Journal of
System Science, Vol.
22(1991), 1671-1678.
SHU2009 Shu Wanneng, Wang
Jiangqing, Min-Min
Chromosome, Genetic
Algorithm for Load
Balancing in Grid
Computing, International
journal of Distributed
sensor network, Vol.
5(2009). No. 1, 62-63.









CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O2-1

Improvement in Dijkstra algorithm
Ashutosh Singh
1
, Sandeep Rathour
2
Department of Electronics Engineering, H.B.T.I., Kanpur, India
1
singhash@iitk.ac.in,
2
ersandeep.hbti@gmail.com

Abstract
To improve the efficiency of a
network, many people have conducted many
studies. For finding the shortest path between
two points, Dijkstra Algorithm is a research
hotspot. The Dijkstra Algorithm has its own
merits and demerits when finding the shortest
path between two points. The two main
disadvantage of Dijkstra Algorithm can be
summarized as (i) Storage structure (ii)
Searching area. One more process by which
Dijkstra Algorithm can be improved. It can be
improved by reducing the no. of repeated
operation and can find the shortest path and
path length from the start point to all the
other remaining nodes by a concept of
Feature matrix.

Keywords: Dijkstra Algorithm , Storage
structure, Searching area, Shortest Path.

1. Introduction
In the graph theory, Dijkstra
Algorithm is the most classical algorithm for
finding the shortest path in the graph. But the
time complexity of the Dijkstra Algorithm is
higher and also it takes larger space for
storage. To overcome these shortcomings of
Dijkstra Algorithm many scholars have done
many researches and their result have its own
merits and demerits. In this paper there are
some improvement in classical Dijkstra
Algorithm to reduce the time complexity of
the Dijkstra Algorithm and also reduce the
searching area of the algorithm. And also for
improvement in Dijkstra Algorithm, if the no.
of repeated operation for searching the
shortest path is reduced, then the efficiency of
the Dijkstra Algorithm is also get improved.


2. Storage structure[1]
The storage structure of Dijkstra
Algorithm is large i.e. it takes large space for
storing it.
Let take a example of 10 nodes. For the
classical Dijkstra Algorithm it requires
10x10[100 values] for its storage. The figure
1 shows the example for 10 nodes. Since it
require 10x10 matrix for storage, the matrix
shown in figure 2 have following properties.
(a) if the start point and the end point are
same,
i.e. i=j, then a
ij
=0
(b) a
ij
=a
ji

(c) if there is no any direct path between
two nodes , then it has ∞(large value) value.

Example
10
9
6
8
4 2
7
1
5
3
2
3 4
5
3
1
2 3
3
3
3
4
5
2

Figure.1 nodes arrangement(10 nodes)

For reducing the storage structure of the
above network shown in figure.1, in the
matrix shown in figure.2, the places where the
values are 0(zero) and ∞(infinity) can be
removed, only non-∞ elements are stored in
the simple linked list format. The improved
structure is shown in figure.3. in this form
each node is having only two elements, first
one is the column value and second one is its
weight.

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O2-2

0 2


∞ ∞
∞ 3


2 0 4



∞ 3 2

∞ 4 0 3


∞ ∞
∞ 5

∞ 3 0 5 2




∞ ∞
∞ 5 0 4




∞ ∞
∞ 2 4 0 3

∞ ∞


∞ ∞
∞ 3 0 3

1
3 3 ∞ ∞

∞ 3 0


∞ 2

3



∞ 0 5

∞ 5



1

5 0

Figure.2 adjancy matrix(10x10)

The storage format of the above matrix get
reduced and given as-
2 2
3 4 1 2
4 ^ 3
9 ^ 2
2 4
8 3
8 ^ 3
3 3
6 ^ 4 4 5
5 ^ 4 4 2
5 5
6 3
2 3 1 3
4 3
10 ^ 1
2 2
7 ^ 4
8 4
9 ^ 3 6 4
G1
G2
G3
G4
G5
G6
G8
G7
G9
G10 7 1
10 ^ 5
9 ^ 5
figure.3 improved storage strucuture

From the above figure it can be told that this
structure takes less space in comparison to the
classical Dijkstra algorithm.

3. Complexity Analysis
Let for the list array, the space
complexity is given as O(T). where T is the
edgentain one or more tables to no. of the
directed graph.
In the worst case T=n
2
, then the space
complexity will be O(n
2
).



4. Searching Area
Since the searching area of a classical
Dijkstra Algorithm is large, there are some
process by which the searching area of the
classical Dijkstra algorithm can be reduced.
Since the shortest path between two points is
a straight line. The direction from the start
point to the is generally strike of the shortest
path when plan the route of the real road
network. The shortest path between two
points is generally the both side of connection
line of the start point to the destination point
and usually it is near to the connection line. If
there are only one edge between the start
point and the destination point, the edge itself
is the shortest path.

4.1 Searching Area is rectangle
The author[1] shows in figure 4 to
find the shortest path from the source S to the
destination D, the theoretical searching area
of classical Dijkstra algorithm is a circle
whose center is source S and radius is R. the
restricted searching area algorithm, the
theortical searching area is small rectangle. It
may be considered that there may exist a
reserve path, each side of the small rectangle
can be increased to threshold T
2
, and it make
a large rectangle. So the restricted searching
area is formed by the two lines L
1
and L
2

which cuts the larger rectangle, so the final
searching area is shown by dark line in figure
4.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O2-3


S
L2
L1
D
T1
T2
R2

Figure.4 rectangular searching area


In the figure, S is the start point,
D is the destination point,
T
1
& T
2
are threshold.
R is the distance between source and
destination.
(i) for the classical Dijkstra Algorithm the
searching area is a circle of radius R , then the
will be
A
1
=π*R
2
(ii) for the restricted searching area algorithm
the searching area is given by
A
2
=2T
1
[R+ T
2
]
Where T
1
/T
2
=
(iii) since the searching node is the directly
proportional to the searching area, so the time
complexity can be given as O(A
1
2
) and
O(A
2
2
).
Commonly two threshold T
1
& T
2
are
relatively small constant.
So the ratio of the time complexities can be
expressed as –
Time complexity ratio=
= [ ]
2

From the above equation it can be compared
that the searching area algorithm is smaller
than the classical Dijkstra algorithm. If there
is increase in the scale of road network and
the distance between the neighbours, the time
complexity difference is more significant. So
this improved algorithm restricting the
searching area reasonably reduce the scale of
searching, so it minimizes the complexity.
And also it improves the efficiency of the
system. The improved algorithm itself restrict
the searching area, unfortunately it may be
the algo. Sometimes can not search the actual
shortest path and it can only be the an
approximate shortest path, it may be a
defective shortest path algo and also it reduce
the accuracy.

4.2 Searching area is an ellipse
When searching area is an ellipse
shown in figure 5, the following parameter
are used for finding the searching area. In
case of ellipse,
SD=R=2a [major axis of ellipse],
Let the eccentricity e=0.7 [<1],
b [semi minor axis] can be obtained.
In this case the searching area is shown in
figure 5 by a thick line, and can be obtained
as,
A
2
= T
1
(4a+πT
2
)
Now the time complexity ratio can be given
as
Time complexity ratio=
= [ ]
2

S
D
L2
L1
T1
T2
T2
T2
T1
SD=2a(major axis)
Eccentricity(e)=.7

figure 5 when searching area is an ellipse
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O2-4

from the above equation, it can be compared
that the ratio of time complexities of, when
searching area is an ellipse is less in
comparison to the ratio of time complexities
of , when searching area is rectangular.
Elliptical searching area gives better result.

5. Feature matrix[2]
Another way for the improvement of
classical dijkstra algorithm. Since if we to
find the shortest path between two points, for
this purpose there are a no. of operations by
which it can be done. But if in any way it can
reduce these no of operations, then the
efficiency of the classical dijkstra algorithm
will get increases. For this purpose there is a
concept of feature matrix. From this feature
matrix we can draw the shortest path tree and
can get the shortest path and shortest path
length from source node to all other
destination node.
For understand the concept of feature matrix
let’s take a example of 6 node which are
connected as shown in figure.6.

1
5
2
4
3
6
12
7 2
2
3
4
1
2
1

Figure.6 nodes arrangement










The abutment matrix for the example is
given as A=
0 2 1 ∞ 7 12
2 0 ∞ 2 ∞ ∞
1 ∞ 0 1 2 ∞
∞ 2 1 0 3 ∞
7 ∞ 2 3 0 4
12 ∞ ∞ ∞ 4 0

There are following steps which from which
feature matrix can be obtained.
1. source S={v1}, D={0,2,1,∞,7,12}
2. firstly find the shortest distance from the
source i.e. D[3]=1, so S={v1,v3}
3. now the nodes connected via 3 to the 1 ,
distance can be obtain by
D[3]+A[3][4]=2<D[4]=∞
& D[3]+A[3][5]=3<D[5]=7
So that modification will D[4]=2 ,D[5]=3 and
the D matrix will become D={0,2,1,2,3,12}
4. iterate operation as follows,
The second time we get that,
S={v1,v3,v2,v4}
& D={0,2,1,2,3,12}
The third time we get S={v1,v3,v2,v4,v5}
& D={0,2,1,2,3,7}
The fourth time S={v1,v3,v2,v4,v5,v6}
& D={0,2,1,2,3,7}
5. compare the following D and A matrix we
will get the following matrix called feature
matrix(F).

0 0 0 0 0 0
2 0 0 0 0 0
1 0 0 0 0 0
0 0 1 0 0 0
0 0 2 0 0 0
0 0 0 0 4 0

CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011

COPO3O2-5

6.From the feature matrix it can be seen that
v1 is the precursor node of the node v2 and
v3. v3 is the precursor node of the node v4
and v5, and v5 is the precursor node of the
node v6.
7. now shortest path tree can be obtained like
shown in figure.7.

1
5
3
4
2
6
1
4
2
1
2

Figure.7 shortest path tree

From the shortest path tree we can find the
shortest path from the source to the all other
nodes and also can find the path length.



6. Conclusion
Finally after studying the above
algorithm, it can concluded that the restricted
searching area algorithm has lower time
complexity than classical dijkstra algorithm
and also when searching area is an ellipse has
lower time complexity when searching area is
a rectangular.
By using the concept of feature matrix, it can
be found the shortest path from source node
to any other node by reducing the no. of
searching operation which increases its
efficiency.

7. References:
[1]. Dong Kai fan, Ping Shi, “Improvement of dijkstra
algorithm and its application in route planning”,
Shandong University of technology, China, IEEE,
International conference on FSKD, 2010, pp 1901-
1904.
[2]. Ji-Xian Xiao, Fang-Ling Lu, “An improvement of
the shortest Path algorithm based on Dijkstra
algorithm”, IEEE, ICCAE, 2010, pp. 383–385.








CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27 2011
COPO303-1
EXECUTED WRIST MOVEMENT CLASSIFICATION FOR
BRAIN COMPUTER INTERFACE
Mariam Siddiqui, Ayman Nidal Khalidi, Omar Farooq,Yusuf U Khan

Department of Electronics Engineering, Department of Electrical Engineering
Zakir Hussain College of Engineering & Technology, Aligarh Muslim University, Aligarh (India)




Abstract—The aim of our study is to
investigate best features for the
classification of pronation and extension
of wrist movements and further use the
best possible classifier to achieve accurate
classification of movements. 10-20
International Montage was used for the
recording of EEG data on 16 channels.
Data was processed, filtered and
windowed. Histogram plots were used for
selection of features. Standard statistical
features were employed in the
investigation and the result suggests that
they provide discriminatory conditions.
While the scope and depth of the study
was limited, the results do suggest that
variance, mean, skewness and kurtosis
can be used for designing the classifier for
both the movements. Different Standard
classifiers are finally designed and trained
to distinguish between different motor
movements.

Keywords:BCI, EEG, Pronation, Extension, Variance,
Skewness, Kurtosis, Mean
I. INTRODUCTION
A brain-computer interface (BCI) is a
system which translates a subject’s
intentions into a control signal for a device,
e.g., a computer application, a wheelchair or
a neuroprosthesis[1]. BCI has given us hope
where an alternate medium can be created
for people with severe motor disability so
that they can somehow work over their
disability to communicate with the external
world in a non-biological and easier way.
Encouraged by this understanding , a
growing interest has developed in recording,
detecting and analyzing brain (neural)
signals to investigate, explore and
understand human motor control systems
with the aim of attempting to build
interfaces that use real– time brain signals to
generate commands to control and/or
communicate with the environment [1].
We mainly focus on noninvasive,
electroencephalogram- (EEG) based BCI
systems which can be used as tools of
communication for the disabled or for
healthy subjects who might be interested in
exploring a new path of human-machine
interfacing. When measured non-invasively,
brain activity is acquired by scalp-recorded
EEG from a subject that tries to convey its
intention by reacting to well-defined tasks or
instructions [1]. Research in developing
successful brain–computer interfaces (BCIs)
endeavors to improve the speed and
accuracy of communication of these
interfaces. This is achieved by improving the
EEG signal processing techniques and
selection of features that are translated into
commands to control and communicate with
augmentative and assistive devices [3].
In this paper we propose and experimentally
verify different features that can be used for
training the classifier for wrist movement in
two directions. A histogram-based analysis
was applied to the EEG signals from
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27
2011
COPO303-2
different channels, to determine combination
of features and channels that yielded the best
class separability. Standard classifiers were
then designed and trained to achieve the best
possible classification accuracy
II. BACKGROUND RESEARCH
This domain is still quite fresh and
acqiuring different shapes ,since it emerged
40 years back in 1970.At that time, the
research subjects were initially animals(rats
and monkeys)but then it took a dramatic
shift in 1990s when the first successful
working implant in humans occured .Now
not only damaged eyesight, hearing,
impaired movements’ etc. can be restored
but researchers could now conceivably
attempt to produce BCI that augment human
functions rather than simply restoring
them.[1]
In order to meet the requirements of the
growing technology expansion,some sort of
standardization was required not only for the
guidance of future researchers but for the
validation and checking of new
developments with other systems as
well,thus a general purpose system was
developed called BCI2000 which made
analysis of brain siganl recording easy by
defining the output formats and operating
protocols to facilitate the researchers in
developing any sort of application.This
made it easier to extract specific features of
brain activity and translate them into device
control signals. [4]

.
III. METHODOLOGY

Fig1: Block diagram for feature extraction of wrist movement
A.Experimental Setup and Data Acquisition
For this study, data was recorded from three
healthy subjects, who signed consent before
their participation in the study. EEG was
recorded at the biomedical signal processing
lab, department of electronic engineering,
AMU.16 channels of EEG data were
recorded using a Brain Tech clarity
TM
system with the electrodes positioned
according to the standard 10-20 system.
Sampling rate was 256 Hz. The subject was
seated on an armchair with eyes closed. The
subject was instructed to avoid speaking and
to ensure least body movement. The
subjected was instructed to move the wrist
from neutral position on an audio prompt
from the experimenter. The series of trials
recorded consisted of following sequences:
First of all the subject was asked to relax for
some time then followed by pronation, after
reaching the extreme position the subject was
asked to hold for some time which was
followed by extension.

Fig2 Sequence followed during
experimental recording

Times of recording of different movements
were different.
B.Data Processing
EEG data was notch filtered to remove 50Hz
frequency from each channel present in the
data due ac power supply of the EEG
machine. Harmonics of 50Hz frequency
were also removed from the data. IIR second
order notch filter with the quality factor (or
Q factor) of 3.55 was used to remove the
desired frequency.
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27
2011
COPO303-3
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
-30
-25
-20
-15
-10
-5
0
5
Normalized Frequency ( rad/sample)
M
a
g
n
i
t
u
d
e

(
d
B
)
Magnitude Response (dB)


Filter #1

Fig3: Magnitude response of a notch filter designed at 50Hz (or
.39∏ radian per sample) frequency

Power spectral density of filtered and
unfiltered EEG signal of all the 16 channels
stacked over one another is shown below
0 20 40 60 80 100 120 140
-200
0
200
400
600
800
1000
1200
frequency
a
m
p
l
i
t
u
d
e

Fig4: PSD plot of 16 channel filtered and unfiltered EEG signal


EEG by nature is non stationary signal. So it
was fragmented into frames so that it can be
assumed stationary for small segment. EEG
data is divided into frames of 1s duration i.e.
frame size of 256 samples /s.

C.Feature extraction
The recorded data was performed for both
pronation and extension wrist movement
related activity and neutral state of right
hand. Histogram of each channel was plotted
and compared for each of the three
movements and the best possible plots which
showed maximum demarcation were used to
extract possible features. Further,from16
channel filtered and segmented data for each
of motor activity and non-motor activity’s
basic statistical features [2] mean, variance,
skewness and kurtosis were extracted and
then averaged over all possible channels
.These features could further be used to
classify between the given movements and
between movements and neutral state.

D. CLASSIFIER DESIGN:
Motor movements are finally differentiated
by using different classifiers on the recorded
data

Table1: percentage values of classification for motor
positions

Classifier Extension Pronation
SVM 77 80
Quad 90 88.8
Linear 80 77.7
Diaglinear 80 66.6

From the observations in Table1 we see that
Quad classifier gives us the best possible
results with high classification percentage
accuracy for both extension and pronation
movements.


IV. RESULT AND DISCUSSION
Histogram plots for 16 channels for different
movements pave way for demarcating
different movements. From Fig 5 which
shows the histogram plot for electrode F3 for
extension shows that the data has long tail,
hence kurtosis could be one of the features.
Further, we can see in the same figure that
the data is asymmetric about the mean
position which shows that it skewed towards
right. Moreover, if we compare Fig 5 & 6,
which is plot for pronation shows that both
different dynamic range and have different
mean values. Likewise, compare Fig 5, 6,
and 7 they differ in their dynamic ranges. Fig
7 which shows the histogram plot of neutral
state doesn’t have long tail and it doesn’t
have much variation about the mean position
thus statistical features can be used to easily
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27
2011
COPO303-4
differentiate between motor and non-motor
activity.
Thus from the above discussion we conclude
that mean, variance, kurtosis and skewness
can prove to be good features for
classification of wrist movement.



-120 -100 -80 -60 -40 -20 0 20 40
0
500
1000
1500
2000
2500
range of values
f
r
e
q
u
e
n
c
y

o
f

o
c
c
u
r
a
n
c
e
F3-Extension

Fig5: Histogram plot of frontal electrode (F3) for
extension.

-60 -40 -20 0 20 40 60
0
500
1000
1500
2000
2500
range of values
f
r
e
q
u
e
n
c
y

o
f

o
c
c
u
r
a
n
c
e
F3-Pronation

Fig6: Histogram plot of frontal electrode (F3) for
pronation.
-40 -30 -20 -10 0 10 20 30 40
0
500
1000
1500
2000
2500
3000
3500
4000
Range of values
f
r
e
q
u
e
n
c
y

o
f

o
c
c
u
r
a
n
c
e
F3-neutral state
Fig7: Histogram plot of frontal electrode (F3) for Neutral
state.







Table2: Statistical features values for different
positions

From the result tabulated in Table2
Extension has larger variance, mean, kurtosis
and skewness values while pronation has
smaller values. Also for both the motor
activity values of features are far away from
the non-motor activity. Thus both the
movements can be easily demarcated from
each other.

From Table1, we justify our findings by
designing different classifiers and training
them for the two class movements to get the
best possible accuracy from the 16 channel
recorded sample values






features
Neutral
position
Extension Pronation

Variance 120.16

586.85 299.06
Mean -0.07 0.33 0.24
kurtosis 4.77 6.44 5.75
skewness 0.13 0.42 0.28
CONFERENCE ON “SIGNAL PROCESSING AND REAL TIME OPERATING SYSTEM (SPRTOS)” MARCH 26-27
2011
COPO303-5
V. CONCLUSION
EEG data class separation was investigated
for two wrist movement classes, pronation
and extension using a 16 channel setup. Four
features were extracted from the histogram
plots of signal for both the movement and
neutral state. Variance, mean, skewness and
kurtosis are the features which easily
demarcates between the two movements. It
was a three class problem since three data
sets were used. Different classifiers were
designed and trained based on the recorded
data sets on the basis of which we classified
between motor movement


VI. ACKNOWLEDGEMENT
The authors are indebted to the UGC.This
work is a part of the funded major research
project CF.No 32-14/2006(SR)

REFERENCES

1. Wolpaw, J.R., Birbaumer, N.,
McFarland, D.J., Pfurtscheller, G., Vaughan,
T.M.: Brain–computer interfaces for
communication and control. Clinical
Neurophys. 113 (2002) 767–791
2. Vaughan, T.M.: Brain-computer
interface technology: a review of the second
international meeting IEEE Trans. on Neural
Systems and Rehab. Eng. 11(2) (2003) 94–
109
3. Syed M.Siddique, laraib Hassan
Siddique: EEG based Brain computer
Interface:Journal of software.4(6)(2009)550-
555
4. F. Sepulveda, M. MeckesCluster
:.Separation Index Suggests Usefulness of
Non-Motor EEG Channels in Detecting
Wrist Movement Direction Intention:
Proceedings of the 2004 IEEE Conference
on Cybernetics and Intelligent Systems
Color Images Enhancement Based on Piecewise
Linear Mapping
Anubhav kumar
1
, Awanish Kr Kaushik
1
, R.L.Yadava
1
, Divya Saxena
2

1
Department of Electronics & Communication Engineering
Galgotia’s College of Engineering & Technology, Gr.Noida, India
2
Department of Mathematics
Vishveshwarya Institute of Engineering and Technology , G.B.Nagar,India
rajput.anubhav@gmail.com




Abstract- Image enhancement is a basic process for all vision and
image processing function in various areas like computer vision,
biomedical image analysis, forensic video/image analysis, remote
sensing and fault detection image enhancement technology plays
a very important role in image processing. By enhancing some
information and restraining other information selectively, it can
improve image visual effect. For a digital image enhancement
process, the digital image is represented in the YCbCr color
space. In this paper, an enhancement arithmetic is used for color
images enhancement based on piecewise linear mapping is
discussed. The method can improve the image sharpness and
make the image characteristics more evident. The enhancement
result is useful for further analysis. Experimental results tested
from a data set of images have demonstrated that the proposed
method is effective and practical. Some parameters like as
RFSIM (Image Quality Assessment) are analyzed for both
existing and proposed method to determine the success and
limitation of our method.

Keywords - Color image enhancement, RGB color space, YCbCr
color space, Piecewise linear mapping, RFSIM.


I.INTRODUCTION

Image enhancement processes consist of a collection of
techniques that seek to improve the visual appearance of an
image or to convert the image to a form better suited for
analysis by a human or machine. Nowadays there is a rapid
increase in the application of color video media. This has
resulted in a growing interest in color image enhancement
techniques.

Some other common technique to enhance the contrast of
images is to perform histogram equalization and
Homomorphic methods. The advantage of histogram
equalization technique is that it works very well for grayscale
images, however, when histogram equalization is used to
enhance color images, it may cause shift in the color scale
resulting in artifacts and an imbalance in image color. The
homomorphic filtering is used to correct non uniform
illumination and to enhance contrasts in the image. It’s a
frequency filtering, preferred to others techniques because it
corrects non uniform lighting and sharpens the edges at the
same time [2]. Another method for color image enhancement
is Bright stretching, also used auto scaling method which is a
common linear mapping function to enhance the brightness
and contrast level of an image [10].

Color image processing and enhancement is a more
complicated process than its counterpart of black-and-white
images [4] due to the involvement of multiple color channels
and the need to preserve the color information content [5]
while enhancing the contrast. Novel techniques that address
these problems are in great demand. For instance, in [6], Chen
et al. proposed to enhance the image quality by making use of
local contrast information and fusing the morphologically
enhanced image with the original.

Similarly Fairweather [8] have used techniques such as
contrast stretching and Markov Random field. They applied
bimodal histogram model to the images in order to enhance
the underwater image. Yoav [9] have used a Physics-based
model. They developed scene recovery algorithm in order to
clear underwater images/scenes through polarizing filter. This
approach addresses the issue of backscatter rather than blur.

In this paper color images enhancement based on piecewise
linear mapping has been proposed .In Section II, an illustration
of theoretical foundations of piecewise linear mapping
function are presented. The proposed enhancement algorithm
is developed in Section III. Experiments conducted using a
variety of color images are described in Section IV and results
are discussed. A conclusion is drawn in Section V.


II. THEORETICAL FOUNDATIONS

Piecewise linear Mapping Function –

The enhancement is done by first transforming the intensity
values using a piecewise linear mapping function for the
intensity component. The piecewise linear mapping function
consists of three line segments as indicated in Figure 1, where
vmin and vmax denote the minimum and maximum intensity
levels in the original image, respectively. This type of
mapping function permits proper allocation of the dynamic
range to different ranges of intensity levels. The actual values
of vlower and vupper determine the dynamic range allocation
for lower, intermediate and higher ranges of intensity levels.



Fig. 1 The piecewise linear mapping function for intensity
Component.

The first moment of the intensity histogram serves as an
indicator about the exposure of the image. The first moment is
closer to vmin for an under exposed image, and closer to vmax
for an over exposed image. Thus based on the first moment of
the histogram, two intensity levels, vlower and vupper are
determined such that the closer the first moment is to vmzn,
the larger the vupper is, and the smaller the dower is, and vice
versa.


III. PROPOSED ALGORTHEM

The color image enhancement based on piecewise linear
mapping function is specified as follows.

Step 1 – In the first step, firstly read the input image and
reshape the input color image (m*n size) in to (m*n)*1
column vector using array. After this change the Sort array
elements in ascending or descending order in the image has
been done.

Step 2 – In the second step, Employ ceil and array function
and find out Vmin and Vmax.Let the full dynamic range for the
display device be [V
MIN
, V
MAX
], where V
MIN
and V
MAX

correspond to the minimum and maximum displayable
intensities, respectively.

Step 3 - Using the luminance and chrominance mapping
functions, the luminance and chrominance values of the
original image are transformed to create the optimally
enhanced color image in NTSC YCbCr space directly.

Fig. 2 Flow chart of Proposed Algorithm

Step 4 - In the last contrast increases by adjusting image
intensity values. It suppresses eventually outliers’ pixels to
improve contrast stretching. It then stretches contrast to use
the whole range of intensity channel and if necessary it
saturates some low or high values. Contrast stretching
attempts to improve the contrast in an image by stretching’ the
range of intensity values it contains to a desired range of
values, e.g. the full range of pixel values that the image
concerned allows.

gሺݔ, ݕሻ = ൮
ሺࢌሺ௫,௬ሻି୫୧୬ ሻ
ሺ௠௔௫ି୫୧୬ ሻ
݂݅ 0 < ࢌሺݔ, ݕሻ < 1
0 ݂݅ 0 > ࢌሺݔ, ݕሻ
1 ݂݅ 1 < ࢌሺݔ, ݕሻ
൲ (1)

where, g( x,y ) represents the images, on the left side it
represents the output image, while f(x,y) represents input
image. In this equation the min and max are the minimum
intensity value and the minimum intensity value in the current
image.

Enhanced Image = g( x,y ) * No. of intensity levels (2)

Here No. of intensity levels shows the total number of
intensity values that can be assigned to a pixel. For example,
normally in the gray-level images, the lowest possible
intensity is 0, and the highest intensity value is 255. Thus "no.
of intensity levels" is equal to 255.



IV. RESULT AND ANALYSIS

On our personal laptop of Pentium 4 the experiment has been
carried out at 2.4 GHz. The test set for this evaluation
experiment randomly selected from the internet with 256*256
and 512*512 sizes. Figures(3a,4a,5a,6a ) show the original
color images, figure (3b,4b,5b,6b) show the enhanced image
by using Murtaza method [1], and the figure (3c,4c,5c,6c)
show the enhanced images with the help of our proposed .



(a ) (b)