DESIGN AND ANALYSIS OF MIMO SYSTEM FOR UWB COMMUNICATION

Transcription

1 DESIGN AND ANAYSIS OF IO SYSTE FOR UWB COUNICATION iir N. oanty, onalisa Bol, axmi Prasad isra 3, Sanjat Kumar isra 4 ITER, Siksa O Anusandan University, Bubaneswar, Odisa, 75030, India Seemanta Engineering College, Jarpokaria, ayurbanj, Odisa, India ABSTRACT ultiple transmit and receive antennas are used IO system. Te system creates parallel IO subcannels to transmit independent streams of data under te appropriate cannel conditions. Similarly, Ultrawideband (UWB) communication as attracted great interest for various applications in recent days. Spatially multiplexed (S) multiple-input multiple-output (IO) systems gains te spectral efficiency as well as ig data rates witout consuming additional power, bandwidt or time slots. In tis paper, we extend te concept of IO to UWB systems. Te correlated cannel for suc purpose is considered and te performance as been analyzed for spatial multiplexing S-UWB-IO system wic is required for estimation. Te system performance substantially degrades in te presence of ig values of spatial correlation. To avoid te degradation of suc system, it as been designed for virtual UWB-IO Time Reversal (TR) system, so tat it is not affected by te transmit correlation. Anoter novel metod to reduce te effect of correlation as been cosen by taking te Eigen value of te cannel matrix for te computation of te system performance. Te result sows its performance. KEYWORDS IO, UWB, BER, Cannel Capacity, Spatial Correlation, TR, Eigen Value. INTRODUCTION Wireless propagation cannels ave been investigated for more tan two decades and a large number of cannel models are designed by many researcers. Ultrawideband (UWB) communication system as become most promising for ig data rate as well as sort-range communication systems. Terefore, it as attracted great interests from bot academic and industrial aspects recently. Because of te restrictions on te transmit power, UWB communications are best suited for sort-range communications []. Increasing demand for iger wireless system capacity as catalyzed several transmission tecniques, among wic multiple-input/multiple-output (IO) tecnology is popular one. Extending IO tecnology to te UWB regime, a large gain in te cannel capacity, robustness and coverage radius is noticed in UWB indoor communications systems []. Tese systems are equipped wit multiple antennas, at bot te transmitter and receiver in order to improve communication performance, in contrast to conventional communication systems wit only one antenna on te transmitter and one antenna on te receiver. DOI : 0.5/ijwmn

2 IO tecnologies overcome te deficiencies of te traditional metods troug te use of spatial diversity. Data can be transmitted over transmit antennas to N receive antennas supported by te receiver terminal. Suc systems are used in wireless communication for enancement of capacity and bit error rate (BER). It offers significant increases in data trougput and link range witout additional bandwidt or transmit power. Tese caracteristics are essential for te coming generation of Telecommunications systems. Rayleig fading as been considered as te propagation cannel for verification. Diversity gain and spatial multiplexing (S) are te two main advantages of IO systems tat are used to study te effect of increase in bit rate wit increasing te number of transmitter and receiver antennas. In IO system, we primarily need to take into account te spatial correlation. Te effect of spatial correlation as to be minimized to obtain better system performance. In [3], te time-reversed cannel impulse response (CIR) is implemented as a filter at te transmitter side. It is well known tat te IO- TR-UWB system can acieve transmit diversity, but it suffers from bot transmit and receive antenna correlations. Te single-input multiple-output TR-UWB (SIO-TR-UWB) or virtual IO-TR-UWB does not face te transmit antenna correlation because it as only one transmit antenna.. REATED ITERATURE An overview of reported measurements and modelling of te UWB indoor wireless cannel is presented in [4]. Different UWB cannel sounding tecniques are discussed and approaces for te modelling of te UWB cannel are reviewed. A considerable work as been performed in [5-7] to caracterize communication cannels for general wireless applications. As IO systems operate at an unprecedented level of complexity to exploit te cannel space-time resources, a new level of understanding of te cannel spacetime caracteristics is required to assess te potential performance of practical multi-antenna links. Empirical investigation of spatial correlation in UWB indoor cannels as been presented in [8]. It was observed tat te coerence distance falls wit cannel bandwidt in end-fire arrays but not in broadside arrays. Te complex correlation decays less rapidly wit distance in broadside arrays tan in end fire arrays, especially under line-of-sigt. Strong dependence of spatial correlation and coerence distance on te cannel centre frequency was observed. Spatial multiplexing single-input multiple-output (S-SIO) UWB communication system using te TR tecnique as been proposed. Te system wit only one transmit antenna, using a spatial multiplexing sceme, can transmit several independent data streams to acieve a very ig data rate. TR can mitigate not only te ISI but te SI caused by multiplexing several data streams simultaneously as well [9]. Antenna selection sceme for IO UWB communication system wit TR is investigated in [0]. Time reversal tecnique as advantage in igly scattering environments to acieve signal focusing troug transmitter-side processing tat enables te use of simple receivers. Te autors ave also demonstrated UWB time reversal system arcitecture taking into account some practical constraints []. 0

3 3. ETHODOOGY A tractable correlated IO UWB cannel model is essential wen developing multiple-antenna UWB systems in order to accurately predict teir performance. Te system is designed based on Alamouti code. Figure sow for a X antenna system. In Alamouti encoding sceme, during any given transmission period two signals are transmitted simultaneously from two transmit antennas. Fig.. Encoder for Alamouti scemes At time t, antenna transmits s, and simultaneously, antenna transmits s0. At time t + T, were T is te symbol duration, signal transmission is switced, wit s0* transmitted by antenna and s* simultaneously, transmitted by antenna. We present te measurements of a IO system under line-of-sigt conditions. 3. Spatial Correlation Toug te space-time focusing feature is one of te benefits, spatial multiplexing as a major role in IO systems. Witout te expansion of bandwidt, ig data rate can be acieved by using spatial multiplexing sceme wit multiple transmit and receive antennas. Te spatial correlation in te multipat cannel is a critical factor in te performance of a IO system and is evaluated. It is mainly caused by inadequate antenna spacing in bot transmitting and receiving side. It causes correlation between te received signals, wic degrades te signal quality, capacity and bit error rate (BER) performance. Capacity increases and BER performance also increases as signal correlation decreases. Te fading correlation between te array elements sould be sufficiently low for a IO system to offer any performance enancement. Te IO cannel matrix is assumed to be independent of eac oter. Oter assumptions tat can be made for suc model analysis is as follows: Te correlation among te receive antennas is independent of te correlation between te transmit antennas. Te effect of antenna coupling is neglected, and we focus only on te spatial correlation. Te transmit and receive correlation matrices are fixed. Te correlation as been included to te IO UWB cannel model by introducing fixed transmit and receive correlation matrices following te well-known Kronecker model [-4], / / H = R H R () rx W tx were, H = cannel matrix of independent cannel realization. w R = transmit correlation matrix wit dimension x. / tx / R = receive correlation matrix wit dimension N x N. rx 03

4 H = correlated cannel 3. etods for Computing Spatial Correlation. Gatering a large amount of data in te target propagation environment. For tat it becomes necessary to estimate a large number of correlation coefficients in an x N IO system. Suppose, tere are N spatial sub-cannels, and correlating eac pair of tem would give rise to (N ) correlation values. A disadvantage of tis approac, beside te fact tat it may be very time-consuming, is tat it may be necessary to estimate a large number of correlation coefficients.. Using fixed correlation matrices as: i. Transmit correlation matrix ii. Receive correlation matrix Te transmit correlation matrix is given as: R tx = O 3 () Similarly, receive correlation matrix is given as: R rx = N N O N 3 N N (3) Te advantage of te fixed correlation model is its simplicity and its immediate application to te existing IEEE a standard. Te fixed correlation matrices R tx and R rx appropriate for a particular environment can be determined by selecting te numerical values of (transmit correlation coefficient) and (receive correlation coefficient), suc tat a close matc is obtained to te BER results acieved wen conducting te system simulation using te measured indoor cannel. Te correlation coefficient values ranges from 0 to. We present BER results for various IO UWB systems for various values of te cannel correlation coefficient. 3.3 Proposed etod for Reduction of Correlation Effect. Virtual IO-UWB-Time Reversal Tecnique System Te time reversal (TR) tecnique, wic is originated from under-water acoustics and ultrasonic, now as been used in many applications suc as localization, imaging and green wireless communications. TR also as sown its potential in dealing wit te ISI problems in UWB. In a TR system, te time-reversed cannel response (CIR) is implemented as a filter at te transmitter side [5]. Wit te elp of te TR pre-filter, te system wit only one transmit antenna can deliver several independent data streams at te same time. We ave taken te spatial correlation into account in, 04

5 were a constant spatial correlation model for IO UWB as been applied. Te performance of system over correlated line of- sigt (OS) cannels is investigated wit te same correlation model. Tis scenario is referred to as cannel model (C) in te IEEE a standard. Te BER results on te adopted correlated cannel model wit an appropriate value of correlation coefficient are sown closely matcing wit tose on te measured indoor cannel. It is well known tat te IO-TR-UWB system can acieve transmit diversity, but it suffers from penalty caused by bot transmit and receive antenna correlations. eanwile, te singleinput multiple-output TR-UWB (SIO-TR-UWB) or virtual IO-TR-UWB does not face te transmit antenna correlation because it as only one transmit antenna. It is sown te virtual IO outperforms te true IO system in terms of te BER performance. Te cannel impulse response (CIR) between te transmit antenna j and te receive antenna i is, [5], i, j i. j i, j = α δ ( t τ ) (4) were, α is te amplitude τ is te delay and te value is considered for it as IEEE a standard from te Table-, H i. j ( t ) =, N,( t),, N, O,, ( ) N, t (5) Te TR pre-filter matrix of te IO system is given by: H i, j ( t ) =,,, N,( t) N, O ( ) N, t were, ( t ) = ( t ) ( t ) (7),,, (6) Te matrix of te equivalent cannel is a square matrix wit te entries in te main diagonal being te summation of te autocorrelation of te original CIRs and oter entries being te summation of te cross-correlation of te original CIRs between te transmit and receive antennas. Te TR matrix H i. j ( t ) is used instead of H W (t) in equation (), to calculate te BER performance of te IO UWB system. Te following parameters ave been considered for evaluation of UWB cannel model. It as been tested according to te IEEE standard a UWB model. Table. Cannel model Parameter of IEEE a standard. Parameters Cannel odel Frequency Cannel odulationτ Specific Values considered C for ine of Sigt communication 3 GHz Rayleig fading cannel QPSK 5.05 ns 05

6 It is sown te virtual IO outperforms te true IO system in term of te BER performance. Anoter metod to reduce te effect of correlation as been cosen by taking te Eigen value of te cannel matrix for te computation of te system performance.. Eigen Values of te Correlation atrix Te sub-cannel correlation, power gains of supported eigen modes, and branc power ratios are analyzed. Te mutual information capacity is found to scale almost linearly wit te IO array size, wit very low variance. Eigen value of te correlation matrix is considered, tat can be expressed as te relation: det( λ I A) = 0 were, λ = te eigen value of A and A = square matrix (8) 4.RESUTS AND DISCUSSION In Fig., sows te performance of cannel capacity in te wireless cannel. Te result is measured according to te various SNR values. Te capacity of te IO cannel as been simulated for number of transmitter and receiver antennas, suc as x, 3 x 3, 4 x 4, and 8 x 8 IO systems. It is observed tat capacity gradually increases wit te number of antennas. Fig. 3, i.e., BER versus SNR, wit different number of transmitting and receiving antennas( x, 3 x 3, 4 x 4, and 8 x 8). It is seen tat te BER performance increases gradually wit te increase in te number of antennas x IO 3 x 3 IO 4 x 4 IO 8 x 8 IO Capacity of different and antennas Capacity (bits/s/hz) SNR Fig.. Performance of capacity wit respect to SNR 06

7 BER of different and antennas x IO 3 x 3 IO 4 x 4 IO 8 x 8 IO Bit Error Rate SNR(dB) Fig. 3 Performance of BER Fig. 4 sows te capacity results for =N=, tat is ( x ) for te systems operating in te C wit correlation coefficients and to be 0, 0.3 and 0.9 in te measured UWB OS cannel. Te capacity for different correlation factors is tested and it as been found tat te capacity decreases wit increase in te correlation factors and also decreases wit increase in te number of correlated antenna elements. BER results for =N=, tat is ( x ) for te systems operating in te C wit correlation coefficients and to be 0, 0.3 and 0.9 and witout applying time reversal in te measured UWB-OS cannel is sown in Fig corr 0.3 corr 0.9 corr Cannel Capacity wit correlation Capacity (bits/s/hz) SNR (db) Fig.4. Capacity performance wit correlation in x systems BER performance wit correlation 0 corr 0.3 corr 0.9 corr Bit Error Rate SNR (db) Fig. 5. BER performance wit correlation 07

8 It as been observed tat te BER performance decreases wit increase in te value of te correlation coefficients BER performance wit correlation and wit application of time reversal 0 corr 0.3 corr 0.9 corr Bit Error Rate SNR (db) Fig. 6. BER performance wit correlation and time reversal tecnique Te BER results for =, N=, tat is ( x ) for te systems operating in te C wit correlation coefficients and to be 0, 0.3 and 0.9 is sown in Fig. 6, by applying virtual IO time reversal in te measured UWB -OS cannel. Also it as been sown tat te BER performance is better tan te one witout time reversal. 0 0 BER wit correlation and taking Eigen Values Witout Eigen value Wit Eigen value 0 - BER SNR (db) Fig. 7. BER wit Eigen values In Fig. 7, BER performance of x IO-UWB systems in te OS indoor C wit correlation coefficients = = 0.4 is sown. Te comparison of te BER sows its efficacy. It is considered wit correlation and te Eigen value of te correlation matrix in sown. Here, it as been observed tat, te BER performance is even better for eigen values of te correlation matrix if considered. 5.CONCUSION ultiple antenna communications tecnologies offer significant advantages over single antenna systems. We ave proposed different metods to reduce te impact of correlation in te IO- UWB cannel and presented a compreensive comparative analysis wit a distance independent spatial correlation model. Comparison between te BER performance of a system wit correlation using virtual IO time reversal tecnique and witout using time reversal tecnique is evaluated. BER performance of te system wit eigen values and witout eigen values of te 08

9 correlation matrix respectively are also evaluated. Capacity result for x systems wit different correlation coefficient values are sown as te proof. Tese advantages include extended range, improved reliability in fading environments and iger data trougputs. REFERENCES [] Andreas F. olisc., Jeffrey R. Foerster.: Cannel odels For Ultrawideband Personal Area Networks,Vol. 0. IEEE Wireless Communications (003) 4-. [] Wasim Q. alik., David J. Edwards.: easured IO Capacity and Diversity Gain Wit Spatial and Polar Arrays in Ultrawideband Cannels,Vol. 55. IEEE Transactions On Communications (007) [3] T. K. Nguyen., H. Nguyen, F. Zeng., T. Kaiser.: Spatial correlation in S-IO-UWB systems using a pre-equalizer and pre-rake filter, Proceedings of te IEEE International Conference on Ultra-Wideband (ICUWB) (00) [4] Zoubir Iraauten., Homayoun Nikookar., Gerard J.. Janssen.: An Overview Of Ultra Wide Band Indoor Cannel easurements And odeling, Vol. 4. IEEE icrowave And Wireless Components etters (004) [5] iir Narayan oanty., Sika isra.: Design of C based Wireless System using Wavelet Packet Network & its PAPR Analysis, IEEE Conference, ICCPCT-03 (03) [6] iir Narayan oanty., axmi Prasad isra., Saumendra Kumar oanty.: Design of IO Space-Time Code for Hig Data Rate Wireless Communication, Vol. 3, No.. IJCSE, (0) [7] icael A. Jensen., Jon W. Wallace.: IO Wireless Cannel odeling and Experimental Caracterization, Space-Time Processing for IO Communications, Jon Wiley & Sons, td (005) -39. [8] Wasim Q. alik.: Spatial Correlation in Ultrawideband Cannels, Vol. 7. IEEE Transactions On Wireless Communications (008) [9] Hieu Nguyen., Zao Zao, Feng Zeng., Tomas Kaiser.: Preequalizer Design for Spatial ultiplexing SIO-UWB TR Systems,Vol.59. IEEE Transactions On Veicular Tecnology (00) [0] Hieu Nguyen., Feng Zeng., Tomas Kaiser.: Antenna Selection for Time Reversal IO UWB Systems, IEEE Transactions on Signal Processing (009) -5. [] Nan Guo., Brian. Sadler., Robert C. Qiu.: Reduced-Complexity UWB Time-Reversal Tecniques and Experimental Results,Vol. 6. IEEE Transactions on Wireless Communications (007) -6. [] T. Kaiser., F. Zeng., E. Dimitrov.: An Overview of Ultrawide-band Systems wit IO,Vol. 97. Proceedings of te IEEE (009) [3] A. Paulraj., R. Nabar., D. Gore.: Introduction to Space-Time Wireless Communications, Cambridge University Press (003). [4] S.. oyka.: Cannel Capacity of IO Arcitecture Using te Exponential Correlation atrix, Vol. 5. IEEE Communications etters (00) [5] Hieu Nguyen., Van Duc Nguyen., Trung Kien Nguyen., Kiattisak aicalernnukul., Feng Zeng., Tomas Kaiser.: On te Performance of te Time Reversal S-IO-UWB System on Correlated Cannels, Hindawi Publising Corporation International Journal of Antennas and Propagation (0)