UNCLASSIFIED AD NUMBER LIMITATION CHANGES

Transcription

1 TO: UNCLASSIFIED AD NUMBER ADB3758 LIMITATION CHANGES Approved for puli release; distriution is unlimited. FROM: Distriution: Further dissemination only as direted y President, Naval Postgraduate Shool, Attn: Code 61, Monterey, CA , JUN 011, or higher DoD authority. NPS ltr dtd 14 May 014 AUTHORITY THIS PAGE IS UNCLASSIFIED

2 NAVAL POSTGRADUATE SCHOOL MONTEREY, CALIFORNIA THESIS PERFORMANCE OF COMPLEX SPREADING MIMO SYSTEMS WITH INTERFERENCE y Efstathios Mintzias June 011 Thesis Co-Advisors: Tri Ha Ri Romero Further dissemination only as direted y (Naval Postgraduate Shool) (June 011) or higher DoD authority Approved for puli release; distriution is unlimited

3 THIS PAGE INTENTIONALLY LEFT BLANK

4 DUDLEY KNOX LIBRARY. May 13, 014 SUBJECT: Change in distriution statement for Performane of Complex Spreading MIMO Systems with Interferene June Referene: Mintzias, Efstathios. Performane of Complex Spreading MIMO Systems with Interferene. Monterey, CA: Naval Postgraduate Shool, Department of Eletrial and Computer Engineering, June 011. UNCLASSIFIED [Further dissemination only as direted y Naval Postgraduate Shool June 011 or higher DoD authority].. Upon onsultation with NPS faulty, the Shool has determined that this thesis may e released to the puli, its distriution is unlimited, effetive May 13, 014. University Lirarian Naval Postgraduate Shool h t t p : / / w w w. n p s. e d u

5 THIS PAGE INTENTIONALLY LEFT BLANK

6 REPORT DOCUMENTATION PAGE Form Approved OMB No Puli reporting urden for this olletion of information is estimated to average 1 hour per response, inluding the time for reviewing instrution, searhing existing data soures, gathering and maintaining the data needed, and ompleting and reviewing the olletion of information. Send omments regarding this urden estimate or any other aspet of this olletion of information, inluding suggestions for reduing this urden, to Washington headquarters Servies, Diretorate for Information Operations and Reports, 115 Jefferson Davis Highway, Suite 104, Arlington, VA 0-430, and to the Offie of Management and Budget, Paperwork Redution Projet ( ) Washington DC AGENCY USE ONLY (Leave lank). REPORT DATE June REPORT TYPE AND DATES COVERED Master s Thesis 4. TITLE AND SUBTITLE Performane of Complex Spreading MIMO Systems 5. FUNDING NUMBERS With Interferene 6. AUTHOR(S) Efstathios Mintzias 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) Naval Postgraduate Shool Monterey, CA SPONSORING /MONITORING AGENCY NAME(S) AND ADDRESS(ES) N/A 8. PERFORMING ORGANIZATION REPORT NUMBER 10. SPONSORING/MONITORING AGENCY REPORT NUMBER 11. SUPPLEMENTARY NOTES The views expressed in this thesis are those of the author and do not reflet the offiial poliy or position of the Department of Defense or the U.S. Government. IRB Protool numer N/A. 1a. DISTRIBUTION / AVAILABILITY STATEMENT Further dissemination only as direted y (Naval Postgraduate Shool) (June 011) or higher DoD authority Approved for puli release; distriution is unlimited 1. DISTRIBUTION CODE A 13. ABSTRACT (maximum 00 words) The ojetive of this thesis is to investigate the performanes of digital ommuniation systems that employ omplex spreading modulation shemes in a Rayleigh fading hannel. First, we examine inary modulation systems with multiple-input, single-output (MISO) onfigurations. Later, we study MISO systems for omplex spreading modulation systems in general. We demonstrate the performanes of speifi and widely used modulation shemes suh as quadrature phase-shift keying (QPSK), 16-quadrature amplitude modulation (16-QAM) and 64-quadrature amplitude modulation (64-QAM). We also investigate the performanes of the previous systems for multiple-input, multiple-output (MIMO) onfigurations for various ominations of transmit and reeive antennas. In all systems, we apply maximal ratio omining (MRC) in order to otain the maximum signal-to-noise ratio for MIMO systems. Finally, for all the different systems and onfigurations, we evaluate their performanes for a Rayleigh fading hannel and in the presene of different types of jamming (arrage noise jamming, pulsed-jamming and tone jamming). 14. SUBJECT TERMS Complex Spreading, MISO, MIMO, Rayleigh Fading, Interferene, Barrage Noise Jamming, Pulsed Jamming, Tone Jamming, Maximum Ratio Comining, MRC. 17. SECURITY CLASSIFICATION OF REPORT Unlassified 18. SECURITY CLASSIFICATION OF THIS PAGE Unlassified 19. SECURITY CLASSIFICATION OF ABSTRACT Unlassified 15. NUMBER OF PAGES PRICE CODE 0. LIMITATION OF ABSTRACT NSN Standard Form 98 (Rev. -89) Presried y ANSI Std UU i

7 THIS PAGE INTENTIONALLY LEFT BLANK ii

8 Further dissemination only as direted y (Naval Postgraduate Shool) (June 011) or higher DoD authority Approved for puli release; distriution is unlimited PERFORMANCE OF COMPLEX SPREADING MIMO SYSTEMS WITH INTERFERENCE Efstathios Mintzias Lieutenant Junior Grade, Helleni Navy B.S., Helleni Naval Aademy, 004 Sumitted in partial fulfillment of the requirements for the degrees of ELECTRICAL ENGINEER and MASTER OF SCIENCE IN ELECTRICAL ENGINEERING from the NAVAL POSTGRADUATE SCHOOL June 011 Author: Efstathios Mintzias Approved y: Tri Ha Thesis Co-Advisor Ri Romero Thesis Co-Advisor Ralph C. Roertson Chair, Department of Eletrial and Computer Engineering iii

9 THIS PAGE INTENTIONALLY LEFT BLANK iv

10 ABSTRACT The ojetive of this thesis is to investigate the performanes of digital ommuniation systems that employ omplex spreading modulation shemes in a Rayleigh fading hannel. First, we examine inary modulation systems with multiple-input, single-output (MISO) onfigurations. Later, we study MISO systems for omplex spreading modulation systems in general. We demonstrate the performanes of speifi and widely used modulation shemes suh as quadrature phase-shift keying (QPSK), 16-quadrature amplitude modulation (16-QAM) and 64-quadrature amplitude modulation (64-QAM). We also investigate the performanes of the previous systems for multiple-input, multiple-output (MIMO) onfigurations for various ominations of transmit and reeive antennas. In all systems, we apply maximal ratio omining (MRC) in order to otain the maximum signal-to-noise ratio for MIMO systems. Finally, for all the different systems and onfigurations, we evaluate their performanes for a Rayleigh fading hannel and in the presene of different types of jamming (arrage noise jamming, pulsed-jamming and tone jamming). v

11 THIS PAGE INTENTIONALLY LEFT BLANK vi

12 TABLE OF CONTENTS I. INTRODUCTION...1 A. OVERVIEW...1 B. LITERATURE REVIEW... C. THESIS OBJECTIVE... D. THESIS OUTLINE...3 II. BACKGROUND...5 A. DIRECT SEQUENCE SPREAD SPECTRUM MODULATION...5 B. FADING CHANNELS...7 C. MULTIPLE INPUT-MULTIPLE OUTPUT SYSTEMS...9 D. MAXIMAL-RATIO COMBINING- RAYLEIGH FADING CHANNEL Maximal Ratio Comining Rayleigh Fading Channel...1 a. PSK QPSK MQAM...13 E. PRESENCE OF JAMMING...14 III. PERFORMANCE ANALYSIS OF DS-PSK MISO...15 A. SYSTEM DESCRIPTION...15 B. BROADBAND JAMMING Diversity L= Diversity L= Diversity L= Diversity L=4... C. PULSED JAMMING...3 D. TONE JAMMING...8 IV. PERFORMANCE ANALYSIS OF IQ COMPLEX SPREADING MISO SYSTEM...33 A. SYSTEM DESCRIPTION...33 B. BROADBAND JAMMING QPSK QAM QAM...47 C. PULSED JAMMING QPSK QAM QAM...56 D. TONE JAMMING QPSK QAM QAM...65 vii

13 V. PERFORMANCE ANALYSIS OF IQ COMPLEX SPREADING MIMO SYSTEM...69 A. SYSTEM DESCRIPTION...69 B. MIMO SYSTEM FULL DIVERSITY FOR I-Q MODULATION MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X C. BROADBAND JAMMING QPSK...7 a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X QAM...77 a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X QAM...81 a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X D. PULSED JAMMING QPSK...86 a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X QAM...91 a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X QAM...95 a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X E. TONE JAMMING QPSK a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X QAM a. MIMO X, X3 X MIMO 3X, 3X3 3X MIMO 4X, 4X3 4X QAM a. MIMO X, X3 X MIMO 3X, 3X3 3X viii

14 . MIMO 4X, 4X3 4X VI. CONCLUSIONS AND FUTURE WORK A. CONCLUSIONS B. FUTURE RESEARCH AREAS LIST OF REFERENCES INITIAL DISTRIBUTION LIST...11 ix

15 THIS PAGE INTENTIONALLY LEFT BLANK x

16 LIST OF FIGURES Figure 1. Bandwidths of the narrowand (lue) and the spread (red) signal. The lak urve represents the noise level. (From [13])...6 Figure. The data waveform (lue) and the PN, or hipping, waveform (red). (From [13])...6 Figure 3. Linear feedak shift register. (From [11, p. 434])...7 Figure 4. n Autoorrelation funtion of an m-sequene, N = 1. (From [11, p. 435])...7 Figure 5. Multipath intensity profile of a fading hannel. (From [11, pp ])...8 Figure 6. Configuration of a MISO system. (After [11, p. 600])...9 Figure 7. Configuration of a MIMO system...10 Figure 8. DSSS PSK Reeiver...15 Figure 9. BER of DS PSK system for roadand jamming and diversity L= Figure 10. BER of DS PSK system for roadand jamming and diversity L=....1 Figure 11. BER of DS PSK system for roadand jamming and diversity L=3.... Figure 1. BER of DS PSK system for roadand jamming and diversity L= Figure 13. BER of DS PSK system for pulsed jamming and diversity L= Figure 14. BER of DS PSK system for pulsed jamming and diversity L=....6 Figure 15. BER of DS PSK system for pulsed jamming and diversity L=3:...7 Figure 16. BER of DS PSK system for pulsed jamming and diversity L= Figure 17. BER of DS PSK system for tone jamming and diversity L= Figure 18. BER of DS PSK system for tone jamming and diversity L= Figure 19. BER of DS PSK system for tone jamming and diversity L= Figure 0. BER of DS PSK system for tone jamming and diversity L= Figure 1. IQ omplex spreading reeiver Figure. BER of DS QPSK system for roadand jamming and diversity L= Figure 3. BER of DS QPSK system for roadand jamming and diversity L= Figure 4. BER of DS QPSK system for roadand jamming and diversity L= Figure 5. BER of DS QPSK system for roadand jamming and diversity L= Figure 6. BER of DS 16-QAM system for roadand jamming and diversity L= Figure 7. BER of DS 16-QAM system for roadand jamming and diversity L=...45 Figure 8. BER of DS 16-QAM system for roadand jamming and diversity L= Figure 9. BER of DS 16-QAM system for roadand jamming and diversity L= Figure 30. BER of DS 64-QAM system for roadand jamming and diversity L= Figure 31. BER of DS 64-QAM system for roadand jamming and diversity L=...48 Figure 3. BER of DS 64-QAM system for roadand jamming and diversity L= Figure 33. BER of DS 64-QAM system for roadand jamming and diversity L= Figure 34. BER of DS QPSK system for pulsed jamming and diversity L= Figure 35. BER of DS QPSK system for pulsed jamming and diversity L= Figure 36. BER of DS QPSK system for pulsed jamming and diversity L= Figure 37. BER of DS QPSK system for pulsed jamming and diversity L= Figure 38. BER of DS 16QAM system for pulsed jamming and diversity L= Figure 39. BER of DS 16QAM system for pulsed jamming and diversity L= xi

17 Figure 40. BER of DS 16QAM system for pulsed jamming and diversity L= Figure 41. BER of DS 16QAM system for pulsed jamming and diversity L= Figure 4. BER of DS 64QAM system for pulsed jamming and diversity L= Figure 43. BER of DS 64QAM system for pulsed jamming and diversity L= Figure 44. BER of DS 64QAM system for pulsed jamming and diversity L= Figure 45. BER of DS 64QAM system for pulsed jamming and diversity L= Figure 46. BER of DS QPSK system for tone jamming and diversity L= Figure 47. BER of DS QPSK system for tone jamming and diversity L=....6 Figure 48. BER of DS QPSK system for tone jamming and diversity L= Figure 49. BER of DS QPSK system for tone jamming and diversity L= Figure 50. BER of DS 16QAM system for tone jamming and diversity L= Figure 51. BER of DS 16QAM system for tone jamming and diversity L= Figure 5. BER of DS 16QAM system for tone jamming and diversity L= Figure 53. BER of DS 16QAM system for tone jamming and diversity L= Figure 54. BER of DS 64QAM system for tone jamming and diversity L= Figure 55. BER of DS 64QAM system for tone jamming and diversity L= Figure 56. BER of DS 64QAM system for tone jamming and diversity L= Figure 57. BER of DS 64QAM system for tone jamming and diversity L= Figure 58. Configuration of a MIMO system...69 Figure 59. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 60. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 61. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 6. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 63. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 64. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 65. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 66. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 67. BER of DS QPSK MIMO for roadand jamming and diversity L= Figure 68. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 69. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 70. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 71. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 7. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 73. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 74. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 75. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 76. BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 77. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 78. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 79. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 80. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 81. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 8. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 83. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 84. BER of DS 64QAM MIMO for roadand jamming and diversity L.= xii

18 Figure 85. BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 86. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 87. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 88. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 89. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 90. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 91. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 9. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 93. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 94. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 95. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 96. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 97. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 98. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 99. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 100. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 101. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 10. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 103. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 104. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 105. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 106. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 107. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 108. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 109. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 110. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 111. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 11. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= Figure 113. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 114. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 115. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 116. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 117. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 118. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 119. BER of DS QPSK MIMO for tone jamming and total diversity L= Figure 10. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 11. BER of DS QPSK MIMO for tone jamming and diversity L= Figure 1. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 13. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 14. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 15. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 16. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 17. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 18. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 19. BER of DS 16QAM MIMO for tone jamming and diversity L= xiii

19 Figure 130. BER of DS 16QAM MIMO for tone jamming and diversity L= Figure 131. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 13. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 133. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 134. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 135. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 136. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 137. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 138. BER of DS 64QAM MIMO for tone jamming and diversity L= Figure 139. BER of DS 64QAM MIMO for tone jamming and diversity L= xiv

20 LIST OF ACRONYMS AND ABBREVIATIONS AWGN BER DS DSSS FEC FHSS GPS I-Q LFSR MIMO MISO MQAM MRC OSTBC PN PSD PSK QPSK SISO SIMO SNR SJR SJNR Additive White Gaussian Noise Bit Error Rate Diret Sequene Diret Sequene Spread Spetrum Forward Error Corretion Frequeny Hopping Spread Spetrum Gloal Positioning System In-Phase/Quadrature Linear Feedak Shift Register Multiple Input Multiple Output Multiple Input Single Output M-ary Quadrature Amplitude Modulation Maximal Ratio Comining Orthogonal Spae Time Blok Code Pseudo Noise Power Spetral Density Phase Shift Keying Quadrature Phase Shift Keying Single Input Single Output Single Input Multiple Output Signal to Noise Ratio Signal to Jamming Ratio Signal to Jamming plus Noise Ratio xv

21 SS TCM W-CDMA Spread Spetrum Trellis Coded Modulation Wideand Code Division Multiple Aess xvi

22 EXECUTIVE SUMMARY Usually, when we refer to a typial ommuniation system, we think of a system with one transmit antenna and one reeive antenna. Reently, a very important emerging tehnology for ommuniations systems is the use of antenna diversity. This terminology is used for systems with multiple transmit and reeive antennas, also referred to as multiple input, multiple output (MIMO) systems. MIMO systems appear quite often in modern wireless ommuniations appliations. This tehnology is used to improve the overall performane of a ommuniation system. One very important harateristi of MIMO systems is that they provide etter performane in fading environments. The use of MIMO systems in fading hannels is now uiquitous. Another widespread tehnology that originated from military appliations is spread spetrum (SS) ommuniations. By using periodi pseudo-noise (PN) sequenes, we spread the spetrum of a narrowand signal to ahieve etter performane in the presene of interferene signals. In our researh, we analyze and evaluate a variety of ommuniation systems. All the systems that we evaluate use a speifi type of SS, namely, diret sequene spread spetrum (DSSS). We onsider Rayleigh fading hannel and three different types of interferene: roadand (arrage) noise, pulsed-noise and tone. In addition, we use maximal ratio omining (MRC), whih is a omining tehnique that ensures maximum signal to-noise ratio (SNR) when only additive white Gaussian noise (AWGN) is present. Another important harateristi that we examine is the numer of transmit and reeive antennas. Firstly, we examine multiple input, single output (MISO) systems, and then we evaluate MIMO systems. The analysis applies to in-phase/quadrature (I-Q) omplex spreading modulation, ut we first fous on DSSS phase-shift keying (DS-PSK), whih is a inary modulation sheme. Then, we study DSSS quadrature phase-shift xvii

23 keying (DS-QPSK), DSSS 16-quadrature amplitude modulation (DS-16QAM) and DSSS 64-quadrature amplitude modulation (DS-64QAM), whih are some of the widespread I- Q modulation shemes. We analyze the preeding systems in order to determine the most effetive system omination of different diversities for the three types of jamming. The omparison assumes fixed signal andwidth and fixed transmit power. We indiate the roustness of a speifi onfiguration in tone jamming, and we see that for most ases roadand jamming is more effetive than tone jamming. We also evaluate the worst-ase ρ in order to perform effetive pulsed jamming. Another aspet that we analyze is the orthogonal spae-time lok odes (OSTBC), whih we use for different diversities. We demonstrate the improved performane that we otain via an Alamouti OSTBC. Finally, we analyze MIMO systems for different transmit and reeive diversities. We point out the MIMO onfiguration with inreased reeive diversity that ahieves the est performane to improve overall performane without imposing any power penalty on the system. xviii

24 ACKNOWLEDGMENTS I dediate this work to my wife, Kalliopi, for her ontinuous loving support and patiene, and to my daughter, Athanasia, for the happiness that she rings in my life. I would like to deeply thank my advisor, Professor Tri Ha, for the invaluale guidane and support that he provided me during the researh and for his unique patiene. I would also like to express my appreiation to Professor Ri Romero for evaluating this work and giving me an additional insight. xix

25 THIS PAGE INTENTIONALLY LEFT BLANK xx

26 I. INTRODUCTION A. OVERVIEW The use of spread spetrum (SS) ommuniations in many appliations is widespread. This tehnique originated from military appliations and is now ommonly used in a variety of ivilian appliations suh as ellular systems, gloal positioning systems (GPS) and more. One of the main advantages of this tehnique is the improvement of performane in the presene of interferene signals [1], [], [3]. This is a vital element for military appliations, where the presene of hostile jamming signals is likely and an severely degrade the performane of ommuniation systems. For ivilian appliations, the great advantage SS offers is the elimination of o-hannel interferene [1], whih does not originate from hostile signals ut from other users in the system. This is very ommon in ellular systems like IS-95, wideand ode-division multiple aess (W-CDMA) and CDMA 000, where the signal of eah user is interferene to other users in the network. The appliation of spread spetrum helps in reduing the effet of interferene and also provides the users a higher level of privay [1], [], [3]. Another emerging tehnology is the use of antenna diversity. Systems with multiple transmit and reeive antennas, alled multiple input-multiple output (MIMO), have eome popular in wireless ommuniations. The advantages of this tehnology are numerous. The effetive range of the ommuniation link and the data throughput are signifiantly inreased without additional transmit power or andwidth and the performane in fading hannels is improved [4], [5]. These important harateristis of MIMO make it the hoie for modern wireless ommuniation standards suh as IEEE 80.11, 4G and WiMAX [4], [6], [7]. 1

27 B. LITERATURE REVIEW In 3G standards WCDMA and CDMA000, the omplex spreading modulation tehnology is employed. In [8], Rihardson presents the use of omplex spreading in WCDMA downlink for hannelization for single user. He also introdues omplex spreading for multiple users and presents the equivalent onstellation diagrams. In [9], a study of adaptive quadrature amplitude modulation (QAM) with omplex spreading for high-speed moile multimedia ommuniations with non-oherent detetion is presented. The study reports an evaluation analysis and simulation for oth flat fading and frequeny-seletive fading hannels and a omparison etween omplex spreading and an oversampling tehnique. The result is that omplex spreading of adaptive QAM does not perform effiiently for frequeny-seletive fading. In [10], a study of omplex spreading for WCDMA uplink is presented, where the analysis is foused on the elimination of the nonlinear orthogonal interferene in power amplifiers y introduing a sheme of omplex sramling with phase estimation. An analysis of a omplex spreading in-phase/quadrature (I-Q) orthogonal overing system is presented in [11]. The system refers to multiple hannels and applies pilot tone-aided demodulation and Doppler traking in fading environment without the presene of interferene for single transmit-single reeive antenna systems. C. THESIS OBJECTIVE In our researh, we evaluate omplex spreading I-Q modulation systems for a fading environment and in the presene of various interferenes. Moreover, we apply transmit and reeive diversity to the aove systems, whih is a ontriution of this work. We analyze various spread spetrum ommuniation systems. In this work, the systems that we evaluate employ diret sequene spread spetrum (DSSS). We study phase-shift keying (PSK), whih is a inary modulation sheme, quadrature phase-shift keying (QPSK), 16-quadrature amplitude modulation (16QAM), and 64-quadrature amplitude modulation (64QAM), whih are I-Q modulation systems.

28 The analysis is performed assuming a Rayleigh fading hannel and various interferene environments, namely, roadand noise, pulsed-noise and tone jamming. These different types of interferene are presented in [1]. We also examine multiple input, single output (MISO) systems and MIMO systems. The performanes of the various systems are analyzed for different modulation shemes and different types of jamming and for oth MISO and MIMO systems. By evaluating the results of the researh, we an determine the most effetive omination of different diversities with the three types of jamming. The omparison is ased on fixed signal andwidth and fixed transmit power. D. THESIS OUTLINE The thesis is organized as follows. Chapter I is the introdution that outlines the importane of the researh. In Chapter II, the neessary akground is developed in order to understand the analysis in the susequent hapters. In Chapter III, we introdue and evaluate MISO-PSK systems. In Chapter IV, we evaluate MISO-IQ systems, and in Chapter V, we evaluate MIMO-IQ systems. Finally, ased on the results of the analyses of the previous hapters, we present our onlusions in Chapter VI. 3

29 THIS PAGE INTENTIONALLY LEFT BLANK 4

30 II. BACKGROUND The intention of this hapter is to provide asi akground knowledge and onepts to the reader in order to understand the analyses presented in the sueeding hapters. A. DIRECT SEQUENCE SPREAD SPECTRUM MODULATION A spread spetrum modulated signal is a signal that has a muh larger andwidth than its narrowand ounterpart. This type of modulation originated from military appliations ut nowadays is widely used in many ivilian appliations. One reason that makes spread spetrum popular is its aility to redue the effetiveness of a jamming signal. The effetiveness of jamming is redued sine the jamming signal must spread its power over the larger andwidth of the spread spetrum signal. This results in a redution of the magnitude of the power spetral density (PSD) of the jamming signal. Thus, the effetiveness of jamming is redued. There are two types of SS modulation tehniques, DSSS and the frequeny-hopping spread spetrum (FHSS). In our researh, we study only DSSS systems. We assume a data sequene (0s and 1s) with a it rate of R = 1/ T, where T is the it time. We also assume a periodi pseudo-noise (PN) sequene of N hips (0s and 1s) with a hip rate of R = 1/ T = NR, where the sequene period is the same as the it time. Assuming that the data and PN sequenes are synhronized, we an omine the two sequenes; if we repeat eah data it N times and modulo- add it to the N hips of the PN sequene, then we generate the DSSS signal that is used to modulate the arrier signal. Beause the hip rate R is N times the it rate, the DSSS signal andwidth is N times the andwidth of the narrowand signal. The PN sequene period may also e seleted to e an integer multiple of the it duration. The numer N = T / T is defined as the spread fator of the DSSS signal. The effet of spreading the signal andwidth is seen in Figure 1. The hipping (PN) and data waveforms are seen in Figure. 5

31 Figure 1. Bandwidths of the narrowand (lue) and the spread (red) signal. The lak urve represents the noise level. (From [13]) Figure. The data waveform (lue) and the PN, or hipping, waveform (red). (From [13]) A very important element of DSSS is the PN sequene. The most important and ommonly used type of PN sequene is the maximal length sequene ( m -sequene), whih an e generated y linear feedak shift registers (LFSR). If the numer of shift n registers is n, then the PN sequene has a period of N = 1 hips. An m-sequene an e generated y a primitive polynomial h(x): where h { 0,1} i n h( x) = 1 + hx+ hx h x + x (.1) n 1 1 n 1. We show the onfiguration of an LFSR in Figure 3. A very important point to notie is that an m-sequene has 1 n ones and 1 n 1 zeros. We oserve in Figure 4 the autoorrelation funtion of a periodi m-sequene when a unit amplitude square wave of duration T is employed as the pulse shape of a hip. For our researh, we onsider only m-sequenes. 6

32 h 1 h h 3... h n 1 i 1 i i 3... i n 1 i n x 3 n 1 x x x x Figure 3. Linear feedak shift register. (From [11, p. 434]) n n Figure 4. Autoorrelation funtion of an m-sequene, N = 1. (From [11, p. 435]) B. FADING CHANNELS A very effetive hannel model that simulates a realisti hannel is the multipath fading hannel. This type of hannel is enountered in most moile wireless ommuniations appliations. In a multipath fading hannel, the transmitted signal arrives at the reeiver via multiple paths. These multipaths arise via signal refletions from ground, uildings, and various strutures in general. They also arise from signal diffration via ending around uilding orners or other edged points and signal sattering from a variety of ojets suh as vehiles, trees, et. Eah signal path results in a randomly delayed, attenuated, and phase-shifted opy of the transmitted signal. These multipath opies omine at the reeiver to give rise to a reeived signal whose attenuation an e modeled as a Rayleigh fading proess (for no line-of-sight path), a Riian fading proess (for one line-of-sight path), or a Nakagami fading proess. In addition, eause the arrival times of the 7

33 multipath opies are random, espeially in a moile environment, the multipath opies in some ases overlap the next it or symol and ause inter-symol interferene. For the time-invariant hannel, the omplex envelope and the frequeny response of the hannel are otained as [11, pp ]: h( τ) = hiδτ ( τi), (.) i jπ fτ ( ) i, i ι H f = he (.3) respetively. The multipath delay spread of a fading hannel is defined as the maximum of the differene in propagation times etween the first path with power P 1 and the last signifiant path of power P l suh that the ratio P 1 / P l is over a given threshold γ [11, pp ]. A typial multipath intensity profile of a fading hannel is shown in Figure 5. A hannel is defined as flat fading when the multipath delay spread is less than the symol time [11, pp ]. Therefore, for a flat fading hannel, we otain the omplex hannel tap as j h= he θ, where h is the attenuation oeffiient of the signal and θ is the phase shift that the fading hannel introdues. In this researh, we assume that the omplex hannel tap is availale via perfet hannel estimation. Magnitude Figure 5. Multipath intensity profile of a fading hannel. (From [11, pp ]) Time 8

34 C. MULTIPLE INPUT-MULTIPLE OUTPUT SYSTEMS The simplest system that we examine is a system with one transmit and one reeive antenna, also alled single input-single output (SISO). In order to ahieve etter performane, espeially in fading hannels, one method is to implement antenna diversity. This an e done y using multiple transmit antennas, referred to as MISO, or multiple reeive antennas, referred to as single input multiple output, referred to as SIMO, or y using multiple transmit and reeive antennas, referred to as MIMO. In this work, we study and ompare MISO and MIMO onfigurations. The system onfiguration of a MISO system is shown in Figure 6. Transmit antenna diversity provides spatial repetition of a transmitted symol via different antennas. The idea is similar to time diversity, where a transmitted symol is repeated multiple times. Time diversity is a type of repetition ode in whih the symol is transmitted multiple times. These hannel-tap weighted opies are availale at the input of the ominer. Spatial repetition via different transmit antennas alone annot provide separate opies of the symol at the reeiver eause the single-antenna reeiver reeives the sum of these opies without the aility to separate them for omining. A omination of transmit antenna diversity and time diversity should provide the reeiver suffiient statistis to separate the opies in order to omine them. We an ahieve this y applying orthogonal spae-time lok odes (OSTBC). Finally, we assume that the transmit antennas have suffiient inter-element spaing in order to have unorrelated paths. L antennas Transmitter Reeiver Figure 6. Configuration of a MISO system. (After [11, p. 600]) 9

35 If we use a MIMO system, we an ahieve full diversity enefit. We represent transmit diversity y using spae-time oding with a omplex L s L t ode matrix G, where we otain L t -fold diversity via L t transmit antennas for m omplex symols transmitting over L s symol times and a ode rate R= m/ Ls. On the reeive side, we otain diversity with L r reeive antennas. With the proper use of maximal ratio omining (MRC), we an ahieve diversity of order L= LL t r [11, pp ], [11, pp ]. The MIMO onfiguration is shown in Figure 7. L r Antennas Transmitter L t Antennas Reeiver Figure 7. Configuration of a MIMO system. D. MAXIMAL-RATIO COMBINING- RAYLEIGH FADING CHANNEL 1. Maximal Ratio Comining The tehnique presented in this setion is known as MRC. For an L-fold diversity system, we assume fixed throughput and fixed transmitted power in order to otain a proper omparison. Thus, the energy of a transmitted opy is 1/L of the symol energy. Therefore, an optimum oherent ominer in the reeiver must exeute a omining operation in order to otain a signal-to-noise ratio that is the sum of signal-to-noise ratios of the L reeived opies. More speifially, this means that the ominer must rotate the phases of the deision sample of L opies to align their phases (o-phasing) and weight eah opy with their respetive su-hannel tap efore summing (weighting). Therefore, strong su-hannels are taken into aount more than weak su-hannels. Consider the 10

36 pre-omining samples of L opies of an aritrary transmitted symol s i at the output of the mathed filter. Eah opy has a orresponding omplex su-hannel tap h l. We have Y = hs + N, l= 1,,..., L, (.4) l l i l where Y l is the reeived omplex signal from l -th su hannel and N l is the omplex additive white Gaussian noise (AWGN) of l -th su hannel. The omplex Gaussian random variale vetors suh that N has a variane σ [11, pp ]. We an express (.4) in l Y = hs i + N, (.5) where = [ Y1 Y... Y ] t, [ 1... ] t and [ 1... ] t L = h h hl = N N NL Y h N. We assume perfet hannel estimation. Thus, the suffiient statisti for oherent demodulation is ( h / ) h Y, where h is the onjugate transpose of vetor h and h is the norm of vetor h. We have the following MRC deision variale h h X = ( hsi + N) = h si + N. (.6) h h The omplex Gaussian noise( h / h ) N has variane σ. The instantaneous MRC output signal-to-noise ratio SNR 0,i given a symol s i is 0, i = si / σ = hl si /σ = l l li, SNR h SNR, (.7) where li = hl si σ SNR, / [11, pp ] is the instantaneous signal-to-noise ratio of the pre-omining th l opy of symol s i. Thus, MRC ahieves the maximum output signal-to-noise ratio, whih is the sum of L input SNR of the multiple opies of the reeived signal. Assuming perfet hannel estimation and using the mapping s i s i for omplex variale to two-dimensional vetor for I-Q modulation signals, we get the onditional pair-wise error proaility etween two vetors hsand i hs j as [11, pp ] 11

37 h si s j Pr ( hsi hs j) = Q. (.8) σ The onditional it error proaility for the Gray-oded signal set is given y the following approximation [11, pp ] where the minimum Eulidean distane is P N h d h, (.9) log M σ T n min ( ) Q d = min s s, σ = N / 0 is the noise variane, M is the total numer of symols of the modulation sheme, and numer of symols at the minimum Eulidean distane d min min i, j i j N n is average or the average numer of nearest neighors. Sine we know the statistis of the Rayleigh fading hannel, we an alulate the it error proaility as N h dmin n P = E P( h ) E Q. (.10) log M σ We also define the modulation-dependent fator [11, pp ] as energy. 1 dmin = α SNR. (.11) σ The diversity symol energy E s is given y E s = Es / L, where E S is the symol. Rayleigh Fading Channel When the su-hannel tap magnitudes normalized mean square value E ( i ) h i are Rayleigh distriuted with h = 1, the omining hannel tap magnitude L h = hl has a hi-square density funtion with L degrees of freedom: [11, pp. l= ] The orresponding it error proaility is given y [11, pp ] 1 f h L 1 y y e ( y) =. ( L 1)! (.1)

38 P L L 1 l N 1 1 n µ L + l 1+ µ, log M l= 0 l (.13) where [11, pp ] µ = α SNR. 1+ α SNR (.14) We otain the following it error proaility expressions y hanging the various parameters for the modulation shemes of interest. a. PSK The minimum Eulidean distane is d min = E, where is the diversity it energy [11, pp ]. We also know that N n = 1and M =. Thus, the it error proaility for PSK is P L L 1 l 1 µ L 1+ l 1+ µ =. l= 0 l (.15). QPSK The minimum Eulidean distane is d min = E s, where E s is the diversity symol energy [11, pp ]. We also know that N n = and M = 4. Thus, the it error proaility for QPSK is the same as that of PSK: P L L 1 l 1 µ L 1+ l 1+ µ =. l= 0 l (.16). MQAM The minimum Eulidean distane is dmin = 6 E / ( M 1) [11, pp ]. We also know that N = 4 4/ M. Thus, the it error proaility for M-ary quadrature amplitude modulation (MQAM) is given y n s P L L 1 l 4 4/ M 1 µ L 1+ l 1+ µ =. log M l= 0 l (.17) 13

39 E. PRESENCE OF JAMMING In this thesis, all the systems that we onsider are evaluated for a Rayleigh fading hannel and three types of jamming: roadand noise, pulsed-noise and tone. We present these types of jamming in detail in the following hapters to evaluate system performane in a hostile environment. In addition, we evaluate the effetiveness of the MIMO diversity and DSSS modulation in these systems. 14

40 III. PERFORMANCE ANALYSIS OF DS-PSK MISO A. SYSTEM DESCRIPTION In this hapter, we analyze the performane of a DSSS PSK MISO system. We examine this speifi system under three types of jamming: roadand jamming, pulsed jamming and tone jamming. The system is shown in Figure 6. It has L transmit antennas and one reeive antenna. At the reeiver, in order to otain a omining deision variale whose SNR is the sum of SNR of the L reeived opies of the transmit signal, we use MRC, whih is an optimum oherent omining tehnique as mentioned in the previous hapter. The system experienes Rayleigh fading. The reeiver is shown in Figure 8. r() t + N() t + J() t T 0 Threshold Detetor t ( )os( π ft + θ κ ) T Figure 8. DSSS PSK Reeiver In Figure 8, we denote rt () as the total reeived signal (the sum of the signal from L transmit antennas), J(t) is the jamming signal and N(t) is the AWGN. The waveform (t) is generated y the PN sequene, f is the arrier frequeny of the signal, T is the it time and θ k is the aritrary phase of the arriving signal. We assume perfet hannel estimation. The power spetral density (PSD) of AWGN equals N /. 0 15

41 B. BROADBAND JAMMING In this setion, we assume that roadand jamming is present. It is a signal similar to AWGN. The PSD of the jamming signal is J '/ 0. The reeived signal from k- th antenna is r () t = h Ad ()os( t π f t + θ ) (3.1) k k k k for it < t < ( i + 1) T, where h k is the hannel tap for k-th antenna, A is the amplitude of the signal, d k is the information it, and θ k is the phase of the hannel tap. Assuming perfet hannel estimation means that we know h k and signal, inluding noise and jamming signals, is θ k. Thus, the total reeived x () t = h Ad ()os( t π f t + θ ) + n() t + j() t. (3.) k k k k Therefore, the signal after the mixer eomes x t h A d t ft t nt ft k '() = k k ()os( π + θk) + () ()os( π + θk) T T + t () jt ()os( π ft + θk ) T (3.3) hk Ad k xk '() t = (1+ os(4π ft + θk)) + t () nt ()os( π ft + θk) T T. (3.4) + t () jt ()os( π ft + θk ) T The high frequeny terms are negleted in our analysis sine they are rejeted y the mathed filter [11, p. 451]. After the integrator, the signal eomes where and T xk ''( t) = hk Ad k + Nk + Jk, (3.5) T N = () t n ()os( t π f t + θ ) dt (3.6) k k T 0 T J = () t j ()os( t π f t + θ ) dt. (3.7) k k T 0 16

42 We alulate the variane of AWGN and jamming signals to e T T σn = N ( ( ) ( ) ( ) ( )os( )os( )) k k Nk = E n t n τ t τ π ft + θκ π fτ + θκ dt dτ T (3.8) 0 0 and T T Nk κ T 0 0 [ ( ( ) ( )) ( ( ) ( )os( )os( ))] σ = E n t n τ E t τ π f t + θ π f τ + θκ dt dτ. (3.9) We know that the auto ovariane of noise is [11, p. 98] N0 δ( t τ) Entn ( ( ) ( τ )) =. (3.10) Thus, the variane of AWGN is T T N0δ( t τ) Nk κ T 0 0 σ = E( ( t) ( τ)os( π f t + θ )os( π f τ + θκ)) dt dτ (3.11) Τ N0 Nk k T 0 σ = E( ( t)os ( π f t + θ )) dt (3.1) Τ N0 Nk T 0 1 os(4π ft + θk) σ = E( + ) dt (3.13) N T N σ = =. (3.14) 0 0 Nk T Similarly, we have T T σj = J ( ( ) ( ) ( ) ( )os( )os( )) k k Jk = E j t j τ t τ π ft + θκ π fτ + θκ dt dτ T (3.15) 0 0 T T Jk κ T 0 0 [ ( ( ) ( )) ( ( ) ( )os( )os( ))] σ = E j t j τ E t τ π f t + θ π f τ + θκ dt dτ. (3.16) The jamming signal is a noise-like signal y definition and, therefore, we have T T J0 ' δ( t τ) Jk κ T 0 0 σ = E( ( t) ( τ)os( π f t + θ )os( π f τ + θκ)) dt dτ (3.17) Τ Τ 0 0 Jk k T T 0 0 J ' J ' 1 os(4π ft + θk) σ = E( ( t)os ( π f t + θ )) dt = E( + ) dt (3.18) J ' T J ' σ = =. (3.19) 0 0 Jk T 17

43 In order to perform a proper omparison of the performanes of a DS system and a non-ds system under arrage noise jamming, we assume that the overall jammer power is the same for oth ases. We onsider that the equivalent noise andwidth of the onventional signal is 1/ T and the equivalent noise andwidth of the spread spetrum signal is 1/ T [11, p. 449], where T is the hip time. We also denote, for the onventional system, the PSD of the jamming signal as J 0. For the spread spetrum system, we denote the PSD of the jamming signal as J 0 '. Thus, the power of the jammer for the onventional system is PJ = ( J0 / )( / T) and for the SS system is J' 0 P = ( J '/ )( / T). Then, we have PJ = PJ' J0 / T = J0'/ T, and finally, we otain where N = T / T is the spread fator. J = JT = J (3.0) ' T N The total variane of noise and jamming signal is N ' 0 + J0 Tk Nk Jk σ = σ + σ =. (3.1) Sine we utilize a Gray oded signal and MRC for a Rayleigh fading hannel, we an use equation (.13). Thus, the proaility of error is given y L L 1 l N 1 1 n µ L + l 1+ µ P = log M l= 0 l, (3.) where for PSK modulation sheme, we have N n = 1 and logm = 1. We also define asjnr µ =, (3.3) 1+ asjnr and α-signal-to-jamming plus noise ratio ( asjnr ) is asjnr 1 d min = σ T, (3.4) where dmin = ε, σ T is the total variane of jamming signal plus noise, and ε = E / L. Thus, for DS-PSK, y using (3.1) and (3.4) we otain 18

44 Therefore, we otain 1 d min 1 4 ε ε asjnr = = = (3.5) σ ( N0 J0') T + 4 N0 + J0' E 1 asjnr = = LN 1 1 ( J 0 0') E E. (3.6) + L + N0 J0' 1 1. (3.7) E E asjnr L N0 J0' µ = = 1 1 Finally, y using (3.0) and (3.7), we otain the equation 1 1. (3.8) E 1 E asjnr L N0 N J 0 µ = = 1 1 In Figures 9-1, we present the it error rates (BER) versus E / N 0 for various diversities (L transmit antennas) and signal-to-jamming ratios ( SJR E / J ) =. We use equations (3.8) and (3.) to alulate the BER. We also present the OSTBC odes that we use for eah diversity order. All the plots are for a spread fator N=

45 1. Diversity L=1 In this ase, we use one transmit antenna, thus we do not have an OSTBC Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 Figure 9. BER of DS PSK system for roadand jamming and diversity L=1.. Diversity L= In this ase, we use an Alamouti OSTB ode, whih has a ode rate of R=1, and its matrix G is given y [14, pp ] s1 s G = * * s s. (3.9) 1 In order to do a proper omparison of the performanes of the various systems examined, we onsider fixed transmitted power. Beause of the two transmit antennas, the symol energy is divided y two. In addition, we have two symols transmitted over two symol times (ode rate R=1). Thus, the symol energy eomes E BER plot is shown in Figure 10. ' s = E /. The s 0

46 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 Figure 10. BER of DS PSK system for roadand jamming and diversity L=. 3. Diversity L=3 In this ase, we use an OSTBC ode, three transmit antennas and rate R = 3/4. The matrix G of the ode is [14, pp ] G = s * * * 3 1 * s3 s s s s s s s. (3.30) We onsider fixed transmitted power and eause of the transmit antennas, the symol energy is divided y three. In addition, we have three symols transmitted over four symol times (ode rate R=3/4). Thus, the symol energy eomes E ' s = E /3. Therefore, in order to maintain the same throughput, we have to inrease the symol rate y multiplying it y a fator of 4/3. Thus, the symol energy is degraded y 3/4, and we otain: E = E = ( E / 3)(3 / 4) = E / 4. The BER plot is shown in Figure 11. '' ' s s s s s

47 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 Figure 11. BER of DS PSK system for roadand jamming and diversity L=3. 4. Diversity L=4 In this ase, we use an OSTBC ode, four transmit antennas and rate R=3/4. The matrix G of the ode is [14, pp ] G = s 0 s s * * s s s 1 3 * * * * s3 s 0 s1 s s s. (3.31) We onsider fixed transmitted power, and eause of the transmit antennas diversity, the symol energy is divided y four. In addition, we have three symols transmitted over four symol times (ode rate R=3/4). Thus, the symol energy eomes E ' s = E /4. Therefore, in order to maintain the same throughput, we have to s

48 inrease the symol rate y multiplying it y a fator of 4/3. Thus, the symol energy is degraded y 3/4, and we otain shown in Figure 1. E = E = ( E / 4)(3 / 4) = (3 /16) E. The BER plot is '' ' s s s s 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 Figure 1. BER of DS PSK system for roadand jamming and diversity L=4. C. PULSED JAMMING In this setion, the jamming signal is pulsed jamming. The jammer is on for ρ perent of the time. Therefore, ρ is defined to e the duty yle of the jammer, and its values are etween zero and one. Thus, when the jammer is off, the system performs with AWGN only and when the jammer is on, the system is affeted y oth AWGN and the jamming signal. From the total proaility theorem, we otain [11, p. 450] P = ρp + (1 ρ) P, (3.3) J+ N N 3

49 where P J+ N is the BER with jamming and AWGN, and N P is the BER with AWGN only. We onsider the PSD of jamming signal as J 0 ''/. The it error proailities are [11, pp ] where and where If we onsider P J+ N L L 1 l 1 µ L 1+ l 1+ µ = l= 0 l, (3.33) E E asjnr L N0 J0'' µ = = 1 1 P N (3.34) L L 1 l 1 µ ' L 1+ l 1 + µ ' = l= 0 l, (3.35) 1 1 µ ' = = asnr L N (3.36) 1 1 E E '', (3.37) N0 J0 then the AWGN an e negleted, [11, p. 450], [15, p. 753]. We otain where P ρp, (3.38) J P J is the BER for jamming only. Therefore, we an use where P J L L 1 l 1 µ '' L 1 l 1 + µ '' = l= 0 l, (3.39) '' 1 1 µ = = asjr L J Similar to the ase of arrage noise jamming, in order to perform a proper omparison of the performanes of a DS system and a non-ds system under pulsed noise 4 0 '' + 1. (3.40)

50 jamming, we assume that the overall jammer power is the same for oth ases. We note that the equivalent noise andwidth of the onventional signal is 1/ T and the equivalent noise andwidth of the spread spetrum signal is 1/ T [11, p. 449], where T is the hip time. We also denote, for the onventional system, the PSD of the jamming signal as J 0. For the spread spetrum system under pulsed noise, the PSD of the jamming signal as J 0 '' when the jammer is on. Thus, the power of the jammer for the onventional system is P J T J = ( 0 / )( / ) and for the SS system is J' 0 have PJ = PJ' J0 / ( ρt) = J0''/ T, and J '' J 0 ρn By omining the (3.40) and (3.41), we otain 1 1 µ '' = = L asjr Nρ J0 P = ρ( J ''/ )( / T ). Then, we 0 =. (3.41) 1. (3.4) + 1 By using equations (3.4) and (3.39), we present the BER performanes versus E / J 0, for various diversities (L transmit antennas) and various values of duty yle ρ in Figures We use the same OSTBC that we desried in the setion on roadand jamming for the orresponding antenna diversities. All the plots are for spread fator N=64. 5

51 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 13. BER of DS PSK system for pulsed jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 14. BER of DS PSK system for pulsed jamming and diversity L=. 6

52 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 15. BER of DS PSK system for pulsed jamming and diversity L=3: 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 16. BER of DS PSK system for pulsed jamming and diversity L=4. 7

53 D. TONE JAMMING In this setion, the jamming signal is tone jamming. The system onfiguration is as shown in Figure 8, and the jamming signal is given y jt ( ) = A os( π ft+ θ ), (3.43) J J where A J is the amplitude and θ J is the aritrary phase of the jamming signal. Thus, the total reeived signal from k-th antenna is After the mixer, we otain x() t = h A dt ()os( π ft+ θ ) + nt () + jt (). (3.44) k k k k x t h A d t ft nt t ft k '() = k k ()os( π + θk) + () ()os( π + θk) T T + t ( )os( π ft + θk) AJ os( π ft + θj) T hk Ad k xk '() t = [ 1+ os(4π ft + θk) ] + nt () t ()os( π ft + θk) T T 1 + AJ t ( ) os(4 ft + k + J) + os( k J) T [ π θ θ θ θ ] (3.45) (3.46) hk Ad k x'() k t = [ 1+ os(4π ft + θk) ] + nt () t ()os( π ft + θk) T T. (3.47) AJ t () + [ os(4 π ft + θk + θj ) + os( θk θj ) ] T By negleting the high frequeny terms after the integrator [11, p. 451], we have T xk ''( t) = hk Ad k + N+ J, (3.48) where and T N = n () t ()os( t π ft + θk ) dt (3.49) T 0 T T A J t ( ) AJ os( θk θj) θk θ J 0 T T 0. (3.50) J = os( ) dt = ( t) dt If we let (t) e derived from a maximal sequene, we otain 8

54 J AJ os( θk θj) T =, (3.51) T where θk θj = θ and ( T / T) = N (spread fator). From (3.14), we know that σ = N / 0. We alulate the variane of the jamming signal to e Nk Therefore, * AJ T AJ T J J J E os E os T = = = T θ (3.5) AJ T σ J =. 4N (3.53) N0 AJ T σt = σn + σj = +. (3.54) 4N Based on (.15), we reall that where P L L 1 l 1 µ L 1+ l 1+ µ = l= 0 l, (3.55) and µ = asjnr 1+ asjnr (3.56) asjnr min = 1 d, (3.57) σ T where dmin = ε and ε = E / L. Thus, for DS-PSK, we otain asjnr 1 d 1 4 ε min = = = σ T N A 4 0 AJ T J + N0 + 4N E asjnr = = We know that the it energy is [11, p. 18] ε T N 1 A J T E AJ T 0 + L N N L + N0 N E 1 (3.58). (3.59) 9

55 E 1 = AT, (3.60) where A is the amplitude of the information signal. By using (3.60) and y replaing the E at the seond term of the denominator of (3.59), we otain asjnr Thus, from (3.56) and (3.61) we have = 1 1 E A 1 L + N0 AJ N 1 µ = = E A 1 asjnr L N0 AJ N. (3.61) 1 1. (3.6) By using (3.6) and (3.55), we otain the BER versus E / N 0 shown in Figures 17-0 for various diversities (L is the numer of transmit antennas) and information signal-to-jamming signal amplitude ratio( A / A ). We also present the OSTBC odes used for eah diversity. All the plots are for spread fator N=64. J 30

56 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 17. BER of DS PSK system for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 18. BER of DS PSK system for tone jamming and diversity L=. 31

57 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 19. BER of DS PSK system for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 0. BER of DS PSK system for tone jamming and diversity L=4. 3

58 IV. PERFORMANCE ANALYSIS OF IQ COMPLEX SPREADING MISO SYSTEM A. SYSTEM DESCRIPTION In this hapter, we analyze the performane of a general I-Q omplex spreading MISO system. We examine the system for various modulation shemes under three types of jamming: roadand, pulsed and tone. The system is shown in Figure 6. It has L transmit antennas and one reeive antenna. We also use MRC. The system experienes Rayleigh fading. The reeiver is shown in Figure 1. I( t) os( π ft + θ κ ) T s Σ Ts 0 Q( t) sin( π ft + θ κ ) T s Minimum Eulidian Distane Detetor I( t) sin( π ft + θ κ ) T s Σ Ts 0 Q( t) os( π ft + θ κ ) T s Figure 1. IQ omplex spreading reeiver. 33

59 In Figure 1, we denote I () t and Q () t the PN waveforms of the I and Q hannels, respetively, and T s as the symol time. We assume perfet hannel estimation, and the PSD of AWGN is equal to N /. 0 B. BROADBAND JAMMING In this setion, the jamming signal is assumed to e roadand jamming. The PSD of the jamming signal is J '/, and the symol energy 0 E s is A is the amplitude of the information signal. E s =, where ( ATs)/ In our system, we define the I-it as d and the Q-it as ˆd. The I-it is spread y I () t, and the Q-it is spread y () t. In addition, the Q-it is spread y I () t, and the Q I-it is spread y () t. Therefore, we otain I and Q aseand signals as Q and where The omplex aseand signal is s () t = d () t d ˆ () t (4.1) LI, I Q s () t = d ˆ () t + d () t. (4.) LQ, I Q s () t = s () t + js () t =d(t), (4.3) L LI, LQ, d = d + jdˆ is the omplex symol and (t) = () t + j () t is the omplex PN I Q funtion. The reeived I-Q signal from k-th antenna is ( ˆ A ) ( ˆ ) A rk() t = hk di() t dq() t os( πft + θ) hk di() t + dq() t sin( πft + θ) (4.4) + nt () + jt () for lt t ( l + 1) T. We alulate the I-hannel deision variale after the two mixers S and the adder to get S rk '() t = r() ()os( ) ()sin( ) I k t I t π ft + θ Q t π ft + θ T (4.5) S 34

60 A r '() ()os( ) ˆ ()os( ) ˆ k t = h ()sin( ) I k di t π ft + θ dq t π ft + θ di t π ft + θ TS 4 (4.6) dq( t)sin( π ft + θ) I( t)os( π ft + θ) Q( t)sin( π ft + θ) hk A rk '() t = d()os( ) () ()sin( )os( ) I I t π ft + θ di tq t π ft + θ π ft + θ T S d ˆ t t ft+ + d ˆ t ft+ ft+ I( ) Q( )os ( π θ) Q( )sin( π θ)os( π θ) d ˆ ( t)sin( π ft+ θ)os( π ft+ θ) + d ˆ ( t ) ( t)sin ( π ft+ θ) I I Q ' ' I () Q( t)sin( π ft + θ)os( π ft + θ) + dq( t)sin ( π ft + θ) + NI + JI d t h A r t = d ft+ d t t ft+ ft+ { k k '( ) os ( π ) ( ) ( )sin( )os( ) I θ I Q π θ π θ TS d ˆ ( t ) ( t) os ( π ft+ θ) sin ( π ft+ θ) + dsin ( π ft+ θ) + N + J ' ' I Q I I { h A r t = d ft+ + ft+ d t t ft+ k k '( ) os ( ) sin ( ) ( ) ( )sin(4 ) I I Q T π θ π θ π θ S d ˆ () t () t os( π f t + θ) 1 + N + J ' ' I Q I I + N + J ' I ' I S } } (4.7) (4.8) (4.9) hk A r '() () ()sin(4 ) ˆ k t = d d () ()os(4 ) I I t Q t π ft + θ di t Q t π ft + θ T. (4.10) The high frequeny terms are negleted sine they are rejeted y the mathed filter [11, p. 98]; therefore, the I-hannel deision variale redues to h A r '( t) = d + N + J. (4.11) k ' ' ki I I TS Similarly, we alulate the Q-hannel deision variale after the two mixers and the adder to get rk '() t = r() ()sin( ) ()os( ) Q k t I t π ft + θ + Q t π ft + θ T (4.1) S A r '() ()os( ) ˆ ()os( ) ˆ k t = h ()sin( ) Q k di t π ft + θ dq t π ft + θ di t π ft + θ TS 4 (4.13) d ( t)sin( π ft+ θ) ( t)sin( π ft+ θ) + ( t)os( π ft+ θ) + N + J ' ' Q I Q Q Q 35

61 h A r t = d t ft+ ft+ + d t t ft+ k k '() ()sin( )os( ) () ()os( ) Q I I Q T π θ π θ π θ S d ˆ t t ft+ ft+ d ˆ t ft+ I( ) Q( )sin( π θ)os( π θ) Q( )os ( π θ) d ˆ ( t)sin ( π ft+ θ) d ˆ ( t ) ( t)sin( π ft+ θ)os( π ft+ θ) I I Q ' ' I () Q( t)sin ( π ft + θ) dq( t)sin( π ft + θ)os( π ft + θ) + NQ + JQ d t r t = h A { d t t π ft+ θ π ft+ θ k k '() () () os( ) sin ( ) Q I Q T S ˆ ˆ di( t ) Q( t)sin( π ft + θ)os( π ft + θ) d os ( π ft + θ) + sin ( π ft + θ) + N + J ' ' Q Q hk A ˆ kq { I Q I Q T S r '() t = d () t () t os( π f t + θ) 1 d () t ()sin(4 t π f t + θ) dˆ + N + J ' Q ' Q ' Q hk A r '() ˆ ˆ k t = d + d () ()sin(4 ) () ()os(4 ) Q I t Q t π ft + θ di t Q t π ft + θ T + N + J ' Q S The high frequeny terms are negleted in our analysis sine they are rejeted y the mathed filter [11, p. 451]; therefore, the Q-hannel deision variale redues to h A r '( t) = dˆ + N + J. (4.18) k ' ' kq Q Q TS After the integrators, we otain the I-hannel deision variale '' TS hk A '' '' rk () t = d + N I I + JI (4.19) and the Q-hannel deision variale '' TS hk A ˆ '' '' rk () t = d + N Q Q + JQ. (4.0) Next, we alulate the varianes of AWGN and the jamming signals. Firstly, we alulate the variane of '' N I. We know that ' N I equals ' nt () NI = I( t)os( π ft θ) Q( t)sin( π ft θ) T + +. (4.1) S Thus, after the integrator, we otain 36 } (4.14) } (4.15) (4.16) (4.17)

62 { () ()os( π θ) ()sin( π θ) } 1 N = N dt = n t t f t + t f t + dt T '' S T ' S I 0 I 0 I Q T S Then, the variane of. (4.) '' N I is TS TS 1 { Ι } { '' σ '' =Ε N =Ε nt () I()os( t π ft θ) Q()sin( t π ft θ) Ν Ι T + + S 0 0 (4.3) } } n( τ) I( τ)os( π fτ + θ) Q( τ)sin( π fτ + θ) dtdτ T T 0 0 { S S 1 σ '' = E{ ntn () ( τ) } I()os( t π ft + θ) Q()sin( t π ft + θ) Ν Ι T (4.4) S I( τ)os( π fτ + θ) Q( τ)sin( π fτ + θ) dtdτ We know that the auto-ovariane of noise is [11, p. 98] Thus, we have TS TS 1 N0 '' Ν Ι TS 0 0 N0 δ( t τ) Entn ( ( ) ( τ )) =. (4.5) σ = δ( t τ) I( t)os( π ft + θ) Q( t)sin( π ft + θ) (4.6) I( τ)os( π fτ + θ) Q( τ)sin( π fτ + θ) dtdτ T 1 N S 0 σ '' = I( t)os( π ft + θ) Q( t)sin( π ft + θ) dt Ν Ι T (4.7) S 0 T N S 0 '' Ν I I Q Ι 4T S 0 σ = ( t)os ( π ft+ θ) ( t ) ( t)sin( π ft+ θ)os( π ft+ θ) (4.8) t f t dt + Q( )sin ( π + θ) T N S 0 '' Ν Ι 4TS 0 σ = 1 I() t Q()sin(4 t π ft + θ) dt. (4.9) The integral of the high frequeny term is equal to zero. Therefore, we otain N0 σ = '' Ν Ι 4. (4.30) '' ' Now, we alulate the variane of N Q. We know that NQ equals ' nt () NQ = I( t)sin( π ft + θ) + Q( t)os( π ft + θ) T. (4.31) Thus, after the integrator, we otain S 37

63 { () ()sin( π θ) ()os( π θ) } 1 N = N dt = n t t f t + + t f t + dt T '' S T ' S Q 0 Q 0 I Q T S The variane of (4.3) '' N Q is TS TS 1 { } { '' σ '' =Ε NQ =Ε () I()sin( ) Q()os( ) N nt t π ft θ t π ft θ Q T S 0 0 (4.33) } } n( τ) I( τ)sin( π fτ + θ) + Q( τ)os( π fτ + θ) dtdτ T T 0 0 { S S 1 σ '' = E{ ntn () ( τ) } I()sin( t π ft + θ) + Q()os( t π ft + θ) ΝQ T (4.34) S I( τ)sin( π fτ + θ) + Q( τ)os( π fτ + θ) dtdτ TS TS 1 N0 σ '' = δ( t τ) I( t)sin( π ft θ) Q( t)os( π ft θ) ΝQ T (4.35) S 0 0 I( τ)sin( π fτ + θ) + Q( τ)os( π fτ + θ) dtdτ TS 1 N 0 σ '' = I( t)sin( π ft + θ) + Q( t)os( π ft + θ) dt ΝQ T (4.36) S 0 TS N0 '' Ν I I Q Q 4T S 0 σ = ( t)sin ( π ft+ θ) + ( t ) ( t)sin( π ft+ θ)os( π ft+ θ) (4.37) t f t dt + Q( )os ( π + θ) TS N0 '' ΝQ 4TS 0 σ = 1 + I() t Q()sin(4 t π ft + θ) dt. (4.38) The integral of the high frequeny term is equal to zero. Therefore, we otain N0 σ Ν '' =. (4.39) Q 4 Next we alulate the varianes jamming signals. We know y definition that they are oth noise-like signals. Firstly, we alulate the variane of '' J I. We know that ' J I is given y ' jt () JI = I( t)os( π ft θ) Q( t)sin( π ft θ) T + +. (4.40) S Thus, after the integrator, we otain T '' S ' 1 TS J I = J { () ()os( ) ()sin( )} 0 Idt = j t 0 I t π ft + θ Q t π ft + θ dt T (4.41) S 38

64 The variane of '' J I is TS TS 1 { Ι } { '' σ '' =Ε J =Ε () I()os( ) Q()sin( ) J jt t π ft θ t π ft θ Ι T + + S 0 0 } } j( τ) I( τ)os( π fτ + θ) Q( τ)sin( π fτ + θ) dtdτ T T 0 0 { (4.4) S S 1 σ '' = E{ jt () j( τ) } I()os( t π ft + θ) Q()sin( t π ft + θ) JΙ T (4.43) S I( τ)os( π fτ + θ) Q( τ)sin( π fτ + θ) dtdτ TS TS 1 J0 ' σ '' = δ( t τ) I( t)os( π ft θ) Q( t)sin( π ft θ) JΙ T + + (4.44) S 0 0 I( τ)os( π fτ + θ) Q( τ)sin( π fτ + θ) dtdτ T 1 J 0 ' S σ '' = I( t)os( π ft + θ) Q( t)sin( π ft + θ) dt JΙ T (4.45) S 0 T J0 ' S '' J I I Q Ι 4T S 0 σ = ( t)os ( π ft+ θ) ( t ) ( t)sin( π ft+ θ)os( π ft+ θ) (4.46) t f t dt + Q( )sin ( π + θ) T J0 ' S '' JΙ 4TS 0 σ = 1 I() t Q()sin(4 t π ft + θ) dt. (4.47) The integral of the high frequeny term is equal to zero. Therefore, we otain Now we alulate the variane of σ J ' 0 '' JΙ =. (4.48) 4 '' J Q. We know that JQ ' is given y ' jt () JQ = I( t)sin( π ft + θ) + Q( t)os( π ft + θ) T. (4.49) Thus, after the integrator, we otain S { () ()sin( π θ) ()os( π θ) } 1 J = J dt = j t t f t + + t f t + dt T '' S T ' S Q 0 Q 0 I Q T S The variane of.(4.50) '' J Q is TS TS 1 { } { '' σ '' =Ε JQ =Ε () I()sin( ) Q()os( ) J jt t π ft θ t π ft θ Q T S 0 0 (4.51) } } j( τ) I( τ)sin( π fτ + θ) + Q( τ)os( π fτ + θ) dtdτ 39

65 T T 0 0 { S S 1 σ '' = E{ jt () j( τ) } I()sin( t π ft + θ) + Q()os( t π ft + θ) JQ T (4.5) S I( τ)sin( π fτ + θ) + Q( τ)os( π fτ + θ) dtdτ TS TS 1 J0 ' σ '' = δ( t τ) I( t)sin( π ft θ) Q( t)os( π ft θ) JQ T (4.53) S 0 0 I( τ)sin( π fτ + θ) + Q( τ)os( π fτ + θ) dtdτ TS 1 J 0 ' σ '' = I( t)sin( π ft + θ) + Q( t)os( π ft + θ) dt JQ T (4.54) S 0 TS J0 ' '' J I I Q Q 4T S 0 σ = ( t)sin ( π ft+ θ) + ( t ) ( t)sin( π ft+ θ)os( π ft+ θ) (4.55) t f t dt + Q( )os ( π + θ) TS J0 ' '' JQ 4TS 0 σ = 1 + I() t Q()sin(4 t π ft + θ) dt. (4.56) The integral of the high frequeny term is equal to zero. Therefore, we otain J ' 0 '' J Q Thus, the varianes of I-hannel and Q-hannel variales are the detetor is σ =. (4.57) 4 N ' 0 + J0 σi = σq =. (4.58) 4 Due to the summation of the varianes of I and Q hannels, the total variane in N ' 0 + J0 T σ =. (4.59) In order to perform proper omparison and evaluation etween a non-ds and a DS system under roadand jamming, we assume that the overall jammer power is the same for oth. If we denote the PSD of the non-dsss system as J / 0 and the PSD of a DSSS system as J '/ 0, then from (3.0) we otain J ' J 0 N Sine we utilize a Gray oded signal in a Rayleigh fading hannel and MRC, we use (.10). Thus, the proaility of error is given y (.13): 0 =. (4.60) 40

66 From (.11) we have L L 1 l N 1 1 n µ L + l 1+ µ P = log M l= 0 l. (4.61) asjnr min = 41 1 d. (4.6) σ T In the following su-setions, we evaluate the performanes of the system for three different I-Q modulation shemes. We apply QPSK, 16-QAM and 64-QAM. We alulate the BER versus the E / N for various diversities (L transmit antennas) and 0 signal-to-jamming ratios ( SJR E / J ) =. We use the OSTBCs that we presented in 0 Chapter III, Setion B for the orresponding diversities. All the plots are for spread fator N= QPSK For QPSK, we have N n =, logm =, dmin = ε S and ε S = ES / L [11, pp ], where ε S is the diversity symol energy. From (4.59) and (4.6), we otain Therefore, we otain 1 d min 1 εs εs asjnr = = = (4.63) σ ( N0 J0') T + 4 ( N0 + J0') ES E 1 asjnr = = =. (4.64) 1 1 LN ( 0 + J0') LN ( 0 + J0') E E L + N0 J0' Finally, we use (4.60) to otain 1 1. (4.65) E E asjnr L N0 J0' µ = = (4.66) E 1 E asjnr L N0 N J 0 µ = = 1 1

67 Thus, as expeted, we oserve that the QPSK performane is idential to PSK. We otain Figures -5 for the different diversities and use the same OSTBCs that we used for the orresponding diversities for PSK Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 Figure. BER of DS QPSK system for roadand jamming and diversity L=1. 4

68 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 Figure 3. BER of DS QPSK system for roadand jamming and diversity L= Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 Figure 4. BER of DS QPSK system for roadand jamming and diversity L=3. 43

69 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure /N 0 BER of DS QPSK system for roadand jamming and diversity L= QAM For 16-QAM, the BER is given y (.17) and M=16, N n =3, log M=4. dmin = 6 ε S /15 and ε = E / L [11, pp ]. We otain from (4.6) and (4.59) S S 1 d min 1 6 εs εs asjnr = = = (4.67) σ ( N0 J0') T ( N0 + J0') ES E asjnr = = =. (4.68) LN ( 0 + J0) 5 LN ( 0 + J0) E E 5L + N0 J 0 From (4.60), we otain 1 1. (4.69) E 1 E asjnr L N0 N J 0 µ = = 1 1 We use the same OSTBCs for the orresponding diversities as PSK. The BER plots for the different diversities are shown in Figures

70 10 0 Figure 6. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16-QAM system for roadand jamming and diversity L= Figure 7. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16-QAM system for roadand jamming and diversity L=.. 45

71 10 0 Figure 8. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16-QAM system for roadand jamming and diversity L= Figure 9. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16-QAM system for roadand jamming and diversity L=4. 46

72 3. 64-QAM For 64-QAM, the BER is given from (.17) and M=64, N n = 7/, logm = 6, dmin = 6 ε S / 63 and ε S = ES / L [11, pp ]. From (4.6) and (4.59), we get 1 d min 1 6 εs εs asjnr = = = (4.70) σ ( N0 J0') T ( N0 + J0') ES E 1 asjnr = = =. (4.71) LN ( 0 + J0') 7 LN ( 0 + J0') E E 7L + N0 J0' Therefore, using (4.60), we otain 1 1. (4.7) E 1 E asjnr 7L N0 N J 0 µ = = 1 1 We use the same OSTBCs for the orresponding diversities as were used for PSK. The BER plots for the different diversities are shown in Figures

73 10 0 Figure 30. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 64-QAM system for roadand jamming and diversity L= Figure 31. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 64-QAM system for roadand jamming and diversity L=. 48

74 10 0 Figure 3. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 64-QAM system for roadand jamming and diversity L= Figure 33. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 64-QAM system for roadand jamming and diversity L=4. 49

75 C. PULSED JAMMING In this setion, the jamming signal is pulsed jamming. The jammer is on for ρ perent of the time. Using the total proaility theorem [11, p. 450], we otain where P J+ N P = ρp + (1 ρ) P, (4.73) J+ N N is the BER with jamming and AWGN, and P N is the BER with AWGN only. Based on results of Setion III-C and using (3.41), we define the PSD of the jamming signal as J 0 ''/, where J '' = J / ( ρn). (4.74) 0 0 We define J 0 as the PSD of the jamming signal in the ase of roadand jamming for the non-spread spetrum equivalent system, noting that the jammer is assumed to have the same power in oth ases in order to have a proper omparison. For QPSK, we have where and P J+ N L L 1 l 1 µ L 1+ l 1+ µ = l= 0 l, (4.75) 1 1, (4.76) E E asjnr L N0 J0'' µ = = 1 1 where P N L L 1 l 1 µ ' L 1+ l 1 + µ ' = l= 0 l, (4.77) For MQAM, we have ' 1 1 µ = = asnr L N (4.78) where P J+ N L L 1 l 4 4/ M 1 µ L 1+ l 1+ µ = log M l= 0 l, (4.79) 50

76 If we onsider 1 1. (4.80) ( M 1) E E asjnr L log M N0 J0'' µ = = E E '', (4.81) N0 J0 then the AWGN an e negleted [11, p. 450], [15, p. 753], and we otain where P J is the BER for jamming only. For QPSK, P ρp, (4.8) J where P J L L 1 l 1 µ '' L 1 l 1 + µ '' = l= 0 l, (4.83) Then, using (4.74), we otain 1 1 µ '' = = asjr L J 0 '' + 1. (4.84) where For MQAM, P J+ N 1 1 µ '' = = L asjr Nρ J (4.85) L L 1 l 4 4 / M 1 µ '' L 1+ l 1 + µ '' = log M l= 0 l, (4.86) 1 1 µ '' = = 1 1 ( M 1) + 1 asjr 3log L M J 0 Again using the relationship (4.74), we otain '' + 1. (4.87) 51

77 1 1 µ '' = = ( M 1) L asjr 3log MNρ J (4.88) We demonstrate the performane of the system for three different I-Q modulation shemes. We use (4.88), (4.86) and (4.8) for MQAM and (4.83), (4.85) and (4.8) for QPSK. We utilize QPSK, 16-QAM and 64-QAM. We alulate the BER versus E / J 0 for various diversities (L transmit antennas) and various values of duty yle ρ and otain Figures We use the same OSTBCs that were used in Chapter III, Setion B. All the plots are for a spread fator N= QPSK 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = Figure /J 0 BER of DS QPSK system for pulsed jamming and diversity L=1. 5

78 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 35. BER of DS QPSK system for pulsed jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 36. BER of DS QPSK system for pulsed jamming and diversity L=3. 53

79 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 37. BER of DS QPSK system for pulsed jamming and diversity L=4.. 16QAM 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = Figure /J 0 BER of DS 16QAM system for pulsed jamming and diversity L=1. 54

80 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = Figure /J 0 BER of DS 16QAM system for pulsed jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 40. BER of DS 16QAM system for pulsed jamming and diversity L=3. 55

81 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 41. BER of DS 16QAM system for pulsed jamming and diversity L= QAM 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = Figure /J 0 BER of DS 64QAM system for pulsed jamming and diversity L=1. 56

82 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = Figure /J 0 BER of DS 64QAM system for pulsed jamming and diversity L= Figure 44. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS 64QAM system for pulsed jamming and diversity L=3. 57

83 10 0 Figure 45. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS 64QAM system for pulsed jamming and diversity L=4. D. TONE JAMMING In this setion, the jamming signal is assumed to e tone jamming. The system onfiguration is shown in Figure 1, and the jamming signal is where jt ( ) = A os( π ft+ θ ), (4.89) J J AJ is the amplitude and θj is the aritrary phase of the jamming signal. Thus, the total reeived signal from the k-th antenna is desried y (4.4). After the mixers, the adders and the integrators, we otain the same deision variales as in (4.19) and (4.0). Based on the results of Setion B, Chapter IV, the varianes of I-hannel and Q-hannel '' '' noise random variales of AWGN are N /4 I = NQ = N. 0 For I-hannel deision variale, after the mixers, the jamming signal is given y A J = os( π ft+ θ ) ( t)os( π ft+ θ) ( t)sin( π ft+ θ) (4.90) ' J I J I Q TS ' AJ JI = [ I( t)os( π ft + θ)os( π ft + θj) Q( t)sin( π ft + θ)os( π ft + θj) T (4.91) S 58

84 ' AJ 1 JI = I( t) [ os(4 π ft + θ + θj) + os( θ θj) ] T S (4.9) 1 Q( t) [ sin(4 π ft + θ + θj) + sin( θ θj) ] The high frequeny terms are negleted [11, p. 451], and we let θ θ = θ'. Then, we otain After the integrator, we have ' AJ JI = I()os t θ' Q()sin t θ' T. (4.93) S T S '' AJ I = I θ Q 0 T S J ()os t ' ()sin t θ' dt (4.94) TS TS '' A J J I = os θ' I() t dt sin θ' Q() t dt TS 0 0. (4.95) We onsider the PN sequenes I () t and Q () t maximal sequenes and N = Ts / T. Thus, we otain We alulate the variane of the jamming signal to e Similarly, we derive '' AJ JI = [ os θ' T sin θ' T] (4.96) T S '' AT J J I = [ os θ' sin θ' ]. (4.97) T S '' ''* AT J σ '' = JI JI = E ( os θ' sin θ' ) J I TS (4.98) AT J σ '' = { os ' sin 'os ' sin ' J E θ θ θ + θ } I 8T (4.99) σ and the jamming signal is s AT 1 1 AT σ '' = J + = I 8T 8N JQ JI J J J s 59 J. (4.100) = σ = ( AT) / (8 N). Therefore, the total variane of AWGN N0 AJ T σt = σn + σ I N + σ Q J + σ I J = +. (4.101) Q 4N

85 For QPSK, ased on (.16), we reall that where P L L 1 l 1 µ L 1+ l 1+ µ = l= 0 l, (4.10) and µ = asjnr, (4.103) 1+ asjnr asjnr 1 d min = where dmin = ε s and ε = E / L. Thus, we otain s s 1 d min 1 εs = = = σ T N0 AJ T, (4.104) σ T εs asjnr (4.105) AJ T 4 + N0 + 4N N Es 1 asjnr = =. (4.106) 1 A J T L N E AJ T 0 + N L + N0 N E We know that the it energy equals [11, p. 18] E 1 = AT (4.107) where A is the amplitude of the information signal. Thus, y using equation (4.107) and replaing the Therefore, we have E in the seond term of the denominator of equation (4.106), we otain asjnr = 1 1 E A 1 L + N0 AJ N 1 µ = = E A 1 asjnr L N0 AJ N Similarly, for MQAM we have. (4.108) 1 1. (4.109)

86 where P J+ N L L 1 l 4 4/ M 1 µ L 1 l 1+ µ = log M l= 0 l (4.110) and 1 dmin 3log M asjnr = =, (4.111) 1 4σT E A 1 LM ( 1) + N0 AJ N 1 1. (4.11) µ = = ( M 1) E A 1 asjnr L 3log M N0 AJ N We illustrate the performane of the system for three different I-Q modulation shemes. We utilize QPSK, 16-QAM and 64-QAM. By using (4.11), (4.110), (4.109) and (4.10), we alulate the BER versus the information signal-to-jamming signal amplitude ratio( A / A ) for various diversities (L transmit antennas) and otain Figures J We use the same OSTBCs that were used in Chapter III, Setion B. All the plots are for spread fator N= QPSK 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 46. BER of DS QPSK system for tone jamming and diversity L=1. 61

87 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 47. BER of DS QPSK system for tone jamming and diversity L= /N 0 Figure 48. No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db BER of DS QPSK system for tone jamming and diversity L=3. 6

88 10 0 Figure 49. No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 BER of DS QPSK system for tone jamming and diversity L=4.. 16QAM 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db Figure /N 0 BER of DS 16QAM system for tone jamming and diversity L=1. 63

89 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db Figure /N 0 BER of DS 16QAM system for tone jamming and diversity L= Figure 5. No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 BER of DS 16QAM system for tone jamming and diversity L=3. 64

90 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 53. BER of DS 16QAM system for tone jamming and diversity L= QAM 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db Figure /N 0 BER of DS 64QAM system for tone jamming and diversity L=1. 65

91 10 0 Figure 55. No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 BER of DS 64QAM system for tone jamming and diversity L= Figure 56. No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 BER of DS 64QAM system for tone jamming and diversity L=3. 66

92 10 0 Figure 57. No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 BER of DS 64QAM system for tone jamming and diversity L=4. In this hapter, we studied MISO systems, whih use I-Q modulation shemes. More speifially, we utilized QPSK, 16QAM and 64QAM. In the next hapter, we introdue MIMO systems for the same modulation shemes. We evaluate their performane and apply exploitation of full diversity of the systems for three types of jamming: roadand, pulsed-noise and tone. 67

93 THIS PAGE INTENTIONALLY LEFT BLANK 68

94 V. PERFORMANCE ANALYSIS OF IQ COMPLEX SPREADING MIMO SYSTEM A. SYSTEM DESCRIPTION In this hapter, we analyze the onfiguration of a MIMO system and the enefits of its use. We also evaluate the performane of a general I-Q omplex spreading MIMO system. We examine the system for various modulation shemes under three types of jamming: roadand, pulsed-noise and tone. We also use MRC. The system is assumed to experiene Rayleigh fading and we assume perfet hannel estimation. The system onfiguration is shown in Figure 58. The asi advantage for the use of a MIMO system is that we an exploit an inreased diversity from oth transmit and reeive antennas. We implement transmit diversity via OSTBC with a omplex Ls Lt ode matrix G whih provides L t -fold diversity via L t transmit antennas for m omplex symols transmitting over L s symol times. Therefore, the ode rate is m/ L s. The reeiver diversity omes from L r reeive antennas. We use MRC in order to exploit full diversity of the desried system [11, pp ], [11, pp ]. The final diversity is L= Lt Lr. L r Antennas Transmitter L t Antennas Reeiver Figure 58. Configuration of a MIMO system. 69

95 B. MIMO SYSTEM FULL DIVERSITY FOR I-Q MODULATION 1. MIMO X, X3 X4 Firstly, we onsider a two-t x (transmitter), two-t r (reeiver) antenna system for general I-Q modulation. We apply an Alamouti ode, whih has a ode rate of R=1. Thus, we have L =, L =, L =, and the ode matrix is [14, pp ] t s r s1 s G = * * s s. (5.1) 1 As mentioned efore, in order to do a proper omparison of the performanes of the various systems examined, we onsider fixed transmitted power. Beause of the transmit antennas diversity, the symol energy is divided y two. In addition, we have two symols transmitted over two symol times (ode rate R=1). Thus, the symol energy eomes E ' s = E /. By using proper MRC and assuming perfet hannel s estimation, we ahieve a full diversity of L= L L = 4. Assuming that we use the same Alamouti OSTBC, we an show that for MIMO systems of T x -3T r (X3) and T x - t r 4T r (X4), the symol energy eomes E ' s = E /, and the diversities are L= L L = 6 s t r and L= L L = 8, respetively. t r. MIMO 3X, 3X3 3X4 We now onsider a 3 T x T r antenna system for general I-Q modulation. We apply an OSTBC with ode rate R=3/4. Therefore, we have L = 3, L = 4, L =, and the ode matrix is [14, pp ] G = s 0 * 1 0 s s s 1 3 * * 3 1 * s3 s 0 s s s t s r. (5.) We onsider fixed transmitted power and eause of transmit diversity, the symol energy is divided y three. In addition, we have three symols transmitted over four symol times (ode rate R = 3/4). Thus, the symol energy is E = E /3. In order ' s s 70

96 to maintain the same throughput, we have to inrease the symol rate 4/3. Thus, the symol energy is degraded y 3/4, and we otain E '' = E ' = ( E / 3)(3 / 4) = E / 4. By s s s s using proper MRC and assuming perfet hannel estimation, we ahieve a diversity of L= L L = 6. Assuming that we use the same OSTBC, we an show that for MIMO t r systems of 3T x -3T r (3X3) and 3T x -4T r (3X4), the symol energy is diversities are L= L L = 9 and L= L L = 1, respetively. t r t r E ' s = E /4, and the s 3. MIMO 4X, 4X3 4X4 Finally, we onsider a 4T x -T r antenna system for general I-Q modulation. We apply an OSTBC with ode rate R=3/4. Therefore, we have L = 4, L = 4, L =, and the ode matrix is [14, pp ] G = s 0 s s * * s s s 1 3 * * * * s3 s 0 s1 s s s t s r. (5.3) We onsider fixed transmitted power and, eause of transmit diversity, the symol energy is divided y four. In addition, we have three symols transmitted over four symol times (ode rate R = 3/4). Thus, the symol energy is E ' s = E /4,. Therefore, in order to maintain the same throughput, we have to inrease the symol rate y 4/3. Thus, the symol energy is degraded y 3/4, and E = E = ( E / 4) (3 / 4) = (3 /16) E. By using proper MRC and assuming perfet '' ' s s s s hannel estimation, we ahieve a diversity of L= L L = 8. Assuming that we use the same OSTBC, we an similarly prove that for MIMO systems of 4T x -3T r (4X3) and t r s 4T x -4T r (4X4), the symol energy is E ' s = 3 E /16, and the diversities are s L= L L = 1 and L= L L = 16, respetively. t r t r 71

97 10 0 Without Jamming C. BROADBAND JAMMING Based on the analysis of Setion B of this hapter and the results and the performane analysis for I-Q modulation in the previous hapter, we use the same expressions for the BER using the full diversity otained y the equivalent MIMO onfiguration. The results are shown in Figures We use spread fator N= QPSK a. MIMO X, X3 X4 For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 59. BER of DS QPSK MIMO for roadand jamming and diversity L=4. 7

98 10 0 Without Jamming 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 60. BER of DS QPSK MIMO for roadand jamming and diversity L=6. For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 61. BER of DS QPSK MIMO for roadand jamming and diversity L=8. 73

99 10 0 Without Jamming 10 0 Without Jamming. MIMO 3X, 3X3 3X4 For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 6. BER of DS QPSK MIMO for roadand jamming and diversity L=6. For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 63. BER of DS QPSK MIMO for roadand jamming and diversity L=9. 74

100 10 0 Without Jamming 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 64. BER of DS QPSK MIMO for roadand jamming and diversity L=1.. MIMO 4X, 4X3 4X4 For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 65. BER of DS QPSK MIMO for roadand jamming and diversity L=8. 75

101 10 0 Without Jamming 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 66. BER of DS QPSK MIMO for roadand jamming and diversity L=1. For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB Figure 67. BER of DS QPSK MIMO for roadand jamming and diversity L=16. 76

102 . 16-QAM a. MIMO X, X3 X Figure 68. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 69. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L=6. 77

103 10 0 Figure 70. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L=8.. MIMO 3X, 3X3 3X Figure 71. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L=6. 78

104 10 0 Figure 7. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 73. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L=1. 79

105 . MIMO 4X, 4X3 4X Figure 74. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L= Figure 75. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L=1. 80

106 10 0 Figure 76. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /N 0 BER of DS 16QAM MIMO for roadand jamming and diversity L= QAM a. MIMO X, X3 X Figure 77. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 BER of DS 64QAM MIMO for roadand jamming and diversity L=4. 81

107 10 0 Figure 78. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 79. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 BER of DS 64QAM MIMO for roadand jamming and diversity L=8. 8

108 . MIMO 3X, 3X3 3X Figure 80. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 BER of DS 64QAM MIMO for roadand jamming and diversity L= Figure 81. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 BER of DS 64QAM MIMO for roadand jamming and diversity L=9. 83

109 10 0 Figure 8. Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 BER of DS 64QAM MIMO for roadand jamming and diversity L=1.. MIMO 4X, 4X3 4X Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 Figure 83. BER of DS 64QAM MIMO for roadand jamming and diversity L=8. 84

110 10 0 Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 Figure 84. BER of DS 64QAM MIMO for roadand jamming and diversity L.= Without Jamming For E/Jo=5dB For E/Jo=10dB For E/Jo=15dB /J 0 Figure 85. BER of DS 64QAM MIMO for roadand jamming and diversity L=16. 85

111 D. PULSED JAMMING Based on the analysis of Setion B of this hapter and the results and the performane analysis of I-Q modulation in the previous hapter, we use the same expressions for BER using the full diversity otained y the equivalent MIMO onfiguration. We present the results in Figures We use spread fator N= QPSK a. MIMO X, X3 X Figure 86. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS QPSK MIMO for pulsed-noise jamming and diversity L=4. 86

112 10 0 Figure 87. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= Figure 88. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS QPSK MIMO for pulsed-noise jamming and diversity L=8. 87

113 . MIMO 3X, 3X3 3X for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 89. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 90. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L=9. 88

114 10 0 Figure 91. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= MIMO 4X, 4X3 4X4 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 9. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L=8. 89

115 10 0 Figure 93. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS QPSK MIMO for pulsed-noise jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 94. BER of DS QPSK MIMO for pulsed-noise jamming and diversity L=16. 90

116 . 16-QAM a. MIMO X, X3 X Figure 95. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= Figure 96. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L=6. 91

117 10 0 Figure 97. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L=8.. MIMO 3X, 3X3 3X Figure 98. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L=6. 9

118 10 0 Figure 99. for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 100. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L=1. 93

119 . MIMO 4X, 4X3 4X for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 101. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 10. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L=1. 94

120 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 103. BER of DS 16QAM MIMO for pulsed-noise jamming and diversity L= QAM a. MIMO X, X3 X for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 104. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L=4. 95

121 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 105. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 106. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L=8. 96

122 . MIMO 3X, 3X3 3X for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 107. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 108. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L=9. 97

123 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 109. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L=1.. MIMO 4X, 4X3 4X for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 110. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L=8. 98

124 10 0 for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 111. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L= for ρ = 1 for ρ = 0.5 for ρ = 0. for ρ = 0.1 for ρ = 0.05 for ρ = 0.01 for ρ = /J 0 Figure 11. BER of DS 64QAM MIMO for pulsed-noise jamming and diversity L=16. 99

125 E. TONE JAMMING Based on the analysis of Setion B of this hapter and the results and the performane analysis of I-Q modulation in the previous hapter, we use the same expressions for the BER using the full diversity otained y the equivalent MIMO onfiguration. We present the results in Figures We use spread fator N= QPSK a. MIMO X, X3 X No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 113. BER of DS QPSK MIMO for tone jamming and diversity L=4. 100

126 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 114. BER of DS QPSK MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 115. BER of DS QPSK MIMO for tone jamming and diversity L=8. 101

127 . MIMO 3X, 3X3 3X No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 116. BER of DS QPSK MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 117. BER of DS QPSK MIMO for tone jamming and diversity L=9. 10

128 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 118. BER of DS QPSK MIMO for tone jamming and diversity L=1.. MIMO 4X, 4X3 4X No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 119. BER of DS QPSK MIMO for tone jamming and total diversity L=8. 103

129 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 10. BER of DS QPSK MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 11. BER of DS QPSK MIMO for tone jamming and diversity L=

130 . 16-QAM 10 0 a. MIMO X, X3 X4 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 1. BER of DS 16QAM MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 13. BER of DS 16QAM MIMO for tone jamming and diversity L=6. 105

131 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 14. BER of DS 16QAM MIMO for tone jamming and diversity L= MIMO 3X, 3X3 3X4 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 15. BER of DS 16QAM MIMO for tone jamming and diversity L=6. 106

132 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 16. BER of DS 16QAM MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 17. BER of DS 16QAM MIMO for tone jamming and diversity L=1. 107

133 . MIMO 4X, 4X3 4X No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 18. BER of DS 16QAM MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 19. BER of DS 16QAM MIMO for tone jamming and diversity L=1. 108

134 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 130. BER of DS 16QAM MIMO for tone jamming and diversity L=

135 3. 64-QAM 10 0 a. MIMO X, X3 X4 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 131. BER of DS 64QAM MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 13. BER of DS 64QAM MIMO for tone jamming and diversity L=6. 110

136 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 133. BER of DS 64QAM MIMO for tone jamming and diversity L= MIMO 3X, 3X3 3X4 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 134. BER of DS 64QAM MIMO for tone jamming and diversity L=6. 111

137 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 135. BER of DS 64QAM MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 136. BER of DS 64QAM MIMO for tone jamming and diversity L=1. 11

138 . MIMO 4X, 4X3 4X No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 137. BER of DS 64QAM MIMO for tone jamming and diversity L= No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 138. BER of DS 64QAM MIMO for tone jamming and diversity L=1. 113

139 10 0 No Jamming /A = 0 db /A = 3 db /A = 7dB /A = 10 db /N 0 Figure 139. BER of DS 64QAM MIMO for tone jamming and diversity L=

140 VI. CONCLUSIONS AND FUTURE WORK A. CONCLUSIONS This researh foused on the performane analysis of I-Q omplex spreading ommuniation systems in a Rayleigh fading hannel in the presene of three types of interferene (roadand, pulsed-noise and tone) for a variety of transmit and reeive antenna diversity. Firstly, we studied MISO systems starting with a inary modulation sheme (DS- PSK) for transmit antenna diversity L=1 to 4. For roadand jamming, the performane improved as the diversity inreased. For pulsed jamming, for L=, the est performane orresponded to the worst ase ρ, ut as L inreased from two to four, the performane degraded. We also notied that even though the performane for the worst-ase ρ dereased, for large ρ, the performane improved. Finally, in the ase of tone jamming (0 db and 3 db of AJ / A) and in the ase of no jamming, as L inreased the performane improved. For AJ / A=7dB, performane remained onstant. For AJ / degraded with respet to E / N A=10 db, it We also evaluated MISO systems that employed I-Q omplex spreading. As expeted, the performanes of DS-PSK and DS-QPSK were idential. We also notied that the performane of DS-QPSK was superior to the other modulation shemes (DS- 16QAM and DS-64QAM) for all ominations of jamming and diversity. This was the expeted result sine QPSK is more power effiient than 16QAM, whih in turn is more power effiient than 64QAM. In arrage noise jamming ase, 16QAM performane inreased as L inreased for the no jamming, 15 db, and 10 db values of E / J 0. For E / J 0 =5 db, the improvement was only for L=. For L=3 and L=4, the performanes (for 5 db) remained almost the same. 64QAM performane improved for diversity L=; then degraded as L inreased. In the ase of pulsed-noise jamming, 16QAM and 64QAM had the same performane. For L=, the BER performane that orresponds to the worst-ase ρ remained the same, ut

141 for L=3 and L=4, the performanes degraded with respet to E / J 0. For other values of ρ, more speifially the large values, the performane inreased tremendously. Finally, for the ase of tone jamming, 16QAM and 64QAM performanes improved for L= and as L inreased. The performanes experiened degradation with respet to E / N 0. We notied that the system performanes were affeted y the OSTBCs. In most ases, there was improvement of performane orresponding to diversity L=. It is eause, for that diversity, we used an Alamouti OSTBC whih has ode rate R=1. For L = 3 and L = 4, the ode rate is R = 3/4. This reflets the performane impat of using rates smaller than unity. In order to maintain the same throughput and the same transmit power, we were fored to inrease the it rate. Thus, the it energy was redued. The it rate redution and the use of ode rates smaller than unity are the reasons why, for most ases, the performanes did not improve for higher transmit antenna diversities (L=3 and L=4). We also evaluated the performanes of the orresponding MIMO systems. For all the modulation shemes and all the ominations of diversity, we notied that as the reeive antenna diversity inreased, the performane inreased eause there is no power penalty for reeive antenna diversity [11, pp ], [14, pp ]. The other important point is that for all MIMO onfigurations, the est overall performane was ahieved y X4 diversity; this was due to the use of the unity rate ( R = 1) Alamouti ode. Finally, we an make some interesting oservations aout the effetiveness of eah jamming tehnique. By oserving the BER plots orresponding to the ase of tone jamming, we see that it was relatively ineffetive. This an e explained y (4.109) and (4.11), where the spread fator N is raised to the power of two, whih redued the effet of the jamming. For omparison purposes, we mention that for the ases of roadand jamming, in (4.66), (4.69) and (4.7), and pulsed-noise jamming, in (4.88) and (4.85), the orresponding terms are multiplied y a fator of N. Considering all the performanes studied, the most effetive type of interferene was pulsed-noise jamming. It was very effetive against QPSK, 16 and 64QAM for 116

142 transmit diversity greater than one. Another important oservation aout the effetiveness of pulsed-noise jamming is the following: if the hostile interferer has knowledge of the worst-ase ρ, then the total diversity L may not e effetive. This is eause even if L inreases, the worst-ase ρ BER does not improve. However, if the hostile interferer does not have this information and uses aritrary ρ, then the total diversity (numer of transmit and reeive antennas) is a huge advantage suh that the effet of jamming an e eliminated, espeially for large values of ρ; the MIMO system for large values of ρ experienes tremendous performane improvement. B. FUTURE RESEARCH AREAS The area of MIMO systems in fading hannels and in the presene of interferene is a sujet of researh in the future for military and ivilian appliations. One sujet of future researh is to onsider other types of fading hannels suh as Rie or Nakagami. Different modulation shemes and onfigurations an e applied suh as W-CDMA or trellis oded modulation (TCM). Another area of researh would e the appliation of various forward error orretion (FEC) tehniques to the MIMO systems studied in this researh work. 117

143 THIS PAGE INTENTIONALLY LEFT BLANK 118

144 LIST OF REFERENCES [1] A. Viteri, Spread spetrum ommuniations-myths and realities, IEEE Communiations Magazine, vol. 17, p. 11, May [] D. T. Magill, F. D. Natali and G. P. Edwards, Spread-spetrum tehnology for ommerial appliations, Proeedings of the IEEE, vol. 8, p. 57, April [3] R. Dixon, Why spread spetrum?, Communiations Soiety, vol. 13, p. 1, July [4] Y. J. Choi, N. H. Lee and S. Bahk, IEEE Performane enhanement y MIMO spatial multiplexing, Personal, Indoor and Moile Radio Communiations, 005. IEEE 16 th International Symposium, vol. 1, p. 87, Septemer 005. [5] B. A. Bjerke and J. G. Proakis, Multiple-antenna diversity tehniques for transmission over fading hannels, Pro. IEEE WCNC 98, pp , Otoer [6] A. Belhouji, C. Deroze, D. Carsenat, M. Mouhamadou, S. Reynaud and T. Monediere, A MIMO WiMax-OFDM ased system measurements in real environments, Antennas and Propagation, 009. EuCAP, 3 rd European Conferene, p. 1106, Marh 009. [7] L. Qinghua et al. MIMO tehniques in WiMAX and LTE: a feature overview, IEEE Communiations Magazine, vol. 48, p. 86, May 010. [8] A. Rihardson, WCDMA Design Handook. Camridge: Camridge University Press, 005. [9] C. Yeonho and S. Kihong, Adaptive QAM modulation with omplex spreading for high-speed moile multimedia ommuniations, IEEE Vehiular Tehnology Conferene, 000, vol. 1, p. 384, 000. [10] J. Xiangdong, Y. Jing and L. Xiong, Study on the harateristi of WCDMA uplink omplex spreading and modulation, Wireless Communiations, Networking and Moile Computing, WiCom th International Conferene, p. 1, Septemer 009. [11] T. T. Ha, Theory and Design of Digital Communiation Systems, Camridge: Camridge University Press, 011. [1] R. L. Peterson, R. E. Ziemer and D. E. Borth, Introdution to Spread Spetrum Communiations. New Jersey: Prentie-Hall In,

145 [13] National Instruments Developer Zone, Understanding Spread Spetrum for Communiations, April [14] E. G. Larsson and P. Stoia, Spae-Time Blok Coding for Wireless Communiations. Camridge: Camridge University Press, 003. [15] J. G. Proakis, Digital Communiations, 4 th edition. New York: MGraw-Hill, In,

146 INITIAL DISTRIBUTION LIST 1. Defense Tehnial Information Center Ft. Belvoir, Virginia. Dudley Knox Lirary Naval Postgraduate Shool Monterey, California 3. Chairman Department of Eletrial and Computer Engineering Naval Postgraduate Shool Monterey, California 4. Professor Tri Ha Department of Eletrial and Computer Engineering Naval Postgraduate Shool Monterey, California 5. Professor Ri Romero Department of Eletrial and Computer Engineering Naval Postgraduate Shool Monterey, California 6. Emassy of Greee Offie of Naval Attahé Washington, Distrit of Columia 7. LTJG Mintzias Efstathios Helleni Navy General Staff Athens, Hellas (Greee) 11

147 DEPARTMENT OF THE NAVY NAVAL POSTGRADUATE SCHOOL DUDLEY KNOX LIBRARY 411 DYER ROAD. ROOM 110 MONTEREY, CALIFORNIA T900 NPS (130) 14 May 14 From: University Lirarian, Naval Postgraduate Shool To: Defense Tehnial Information Center (DTIC-OQ) Suj: CHANGE IN DISTRIBUTION STATEMENT FOR ADB Request a distriution statement hange for: ADB3758: Mintzias, Efstathios. Performane of Complex Spreading MIMO Systems With Interferene. Monterey, CA: Naval Postgraduate Shool, Department of Eletrial and Computer Engineering, June 011. UNCLASSIFIED [Further dissemination only as direted y Naval Postgraduate Shool June 011 or higher DoD authority.]. Upon onsultation with NPS faulty, the Shool has determined that, effetive May 13, 014, the distriution limitations on this report have een removed and distriution has een roadened to puli release. 3. POC for this request is George Gonalves. Lirarian, Restrited Resoures and Servies, , DSN (gmgonal@nps.edu). &lild~.r ELEANOR S. UHL~ University Lirarian