Multiple Antenna Techniques in WiMAX Systems

Transcription

1 Elektrotehniški vetnik 75(1):24-30, 2008 Electrotechnical Review: Ljubljana, Slovenija Multiple Antenna Technique in WiMAX Sytem T. Celcer 1, S. Plevel 1,2, T. Javornik 1, C. Fortuna 1 and G. Kandu 1 1 Jožef Stefan Intitute, Ljubljana, Slovenia 2 Telima, d.o.o., Trzin-Ljubljana, Slovenia tine.celcer@ij.i Abtract. Multiple antenna ytem can offer ignificant improvement in ytem performance due to their ability to exploit multipath ignal propagation and take advantage of it random nature in order to achieve diverity or patial multiplexing gain. A uch, they are very uitable for implementation in WiMAX (Worldwide Interoperability for Microwave Acce) ytem to improve it performance in harh urban and indoor environment. WiMAX phyical layer i baed on Orthogonal Frequency Diviion Multiplexing (OFDM), which i very robut to multipath propagation and enable traightforward uage of MIMO technique. In thi paper multiple antenna technique and their ue in Fixed and Mobile WiMAX ytem are dicued. We preent tet reult, collected with field meaurement during a Fixed WiMAX ytem deployment, and decribe reult obtained by applying a channel imulator whoe parameter are tuned according to field meaurement data. A performance comparion between STC (Space-Time Coding) and non-stc operation mode i analyzed for different coding and modulation cheme pecified in the tandard. Keyword: multiple antenna, WiMAX, pace-time coding, patial multiplexing, OFDM Uporaba ve'antenkih itemov v omrežjih WiMAX Povzetek. Ve.antenki itemi omogo.ajo znatno izboljšanje u.inkovitoti brezži.nega komunikacijkega itema, aj izkoriš.ajo latnot, kot je širjenje ignala po ve. poteh, v vojo korit, tako da izrabljajo naklju.not oziroma nekoreliranot teh poti za protorko lo.evanje o.ano polanih ignalov. Kot taki o ti itemi zelo primerni za implementacijo v itemih WiMAX, ki na fizi.ni ravni uporabljajo tehnologijo OFDM, odporno proti širjenju ignala po ve. poteh. V pripevku ta podana pregled tehnik v ve.- antenkih itemih in njihova uporaba v fiknih in mobilnih itemih WiMAX. Predtavljeni o tudi rezultati meritev na terenu in laboratorijkih meritev. V laboratorijkih meritvah mo uporabili imulator radijkega kanala parametri radijkega kanala, pridobljenimi z meritvami na terenu. Izvedli mo primerjavo u.inkovitoti dveh na.inov delovanja, in icer z uporabo in brez uporabe protorko-.aovnega kodiranja. Primerjavo mo opravili za razli.ne kodno-modulacijke heme, ki o predpiane v tandardu. Klju'ne beede: ve.antenki itemi, WiMAX, protorko-.aovno kodiranje, protorki multiplek, OFDM (ortogonalno frekven.no multiplekiranje) 1 Introduction The main attribute that are deired in a modern communication ytem are high pectral efficiency and high data rate, along with high quality of ervice (QoS) - meaning low outage probability, low bit error rate, etc. - and wide coverage. However, a wirele channel preent a very hotile and difficult environment for proviion of uch attribute. There are variou Received 21 November, 2007 Accepted 4 February 2008 drawback, uch a ignal attenuation due to path lo, limited bandwidth, co-channel interference (CCI) due to the preence of other uer and, mot importantly, evere fluctuation in ignal level, referred to a fading [1]. Fading i a reult of multipath propagation and the Doppler pread which i caued by the mobility of the uer a well a variation in the environment. Two main olution to the above problem are typically propoed by modern tandard. The firt i adaptive coding and modulation (ACM), which i baed on the concept that the coding and modulation cheme adapt dynamically to the channel condition. The econd i the ue of multiple antenna at the tranmitter and receiver. While ACM can cope with low fading, multiple antenna technique can alo combat fat fading. Optimal performance can be obtained by taking advantage of both olution. Multiple input multiple output (MIMO) ytem exploit multipath propagation and random ignal fading to increae the ytem performance without extra bandwidth and power cot. There are four different benefit offered by MIMO ytem: diverity gain, patial multiplexing gain, array gain, and interference reduction. The ue of MIMO technique i alo included in Fixed and Mobile WiMAX (Worldwide Interoperability for Microwave Acce) ytem profile. Several multiple antenna option are upported, enabling the exploitation of all the above tated benefit. The paper i organized a follow: firt we preent a hort overview of multiple antenna technique and their main benefit. In Section 3, the IEEE tandard i

2 Multiple Antenna Technique in WiMAX Sytem 25 dicued briefly and in Section 4 the ue of multiple antenna in WiMAX ytem, a decribed in the tandard, i explored. In Section 5, reult obtained from field and laboratory meaurement are compared with thoe from theory, a dicued in previou ection. Concluion are preented in Section 6. 2 Overview of multiple antenna technique MIMO ytem are wirele ytem equipped with multiple antenna at the receiver and tranmitter. The great interet in MIMO ytem i due to their ability to increae ytem capacity or reliability without any increae in tranmitting power or bandwidth. MIMO technology make ue of the patial dimenion by taking advantage of multipath propagation channel characteritic. Spectral efficiency can be increaed by the imultaneou tranmiion of data over different antenna uing random fading a the mean of ignal eparation. On the other hand increae in ytem reliability i achieved, with the inertion of redundancy, by tranmitting multiple copie of the ame ignal over different propagation path, thu achieving patial diverity gain. Suppoe we have a MIMO ytem with M tranmit and N receive antenna, then there are M N ubchannel between the tranmitter and the receiver. Auming frequency non-elective or flat fading, the received ignal can be expreed a: y = Hx + n, (1) where H i the N M dimenional channel matrix with complex coefficient h ij that repreent the channel repone between the j-th tranmit and i-th receive antenna, y and x are the received and tranmitted vector repectively, and n i the noie vector. In Single Input Single Output (SISO) ytem, capacity grow logarithmically with ignal-to-noie ratio (SNR). In [2], Telatar ha hown that the capacity of MIMO ytem, compared to SISO ytem, grow linearly with the number of independent ubchannel, which equal the rank of channel matrix H and can be at mot min(m,n). The increae in capacity, compared to that of the SISO ytem, i referred to a patial multiplexing gain. Alternatively, a diverity gain of M N can be achieved by auming M N random fading propagation path [3]. Diverity technique are baed on the aumption that there i low probability of all path being in a deep fade. Hence, diverity gain decreae the fluctuation in received ignal power, which mitigate fading effect. Diverity gain d tell u how fat the decoding error probability P e decay with the increae of SNR: P e SNR) d ( (2) Beide patial multiplexing and diverity gain, other ignificant benefit, uch a array gain or interference cancellation/avoidance, can be achieved. Array gain i the increae in SNR due to coherent combining of ignal at the receiver and can be achieved, even in a highly correlated channel, either at the tranmitter or receiver ide, providing that the channel repone i known at the repective ide. Interference cancellation i important in cellular multiuer ytem, ince the preence of ubcriber uing the ame frequency band caue CCI. The interference can be mitigated by patially eparating ignal to/from different uer. Application of uch a method at the tranmitter i called beamforming. It i ued at the bae tation (BS) ide and enable better frequency reue and thu increae the overall capacity of the cellular ytem. However, it i not poible to exploit thee benefit imultaneouly to the full, due to conflicting demand on the patial degree of freedom. A certain trade-off between poible gain mut be taken into account, baed on the ytem requirement. The maximum trade-off between the two principal gain (patial diverity and multiplexing) i a piecewie linear function [4], howing that, if maximum patial multiplexing gain i to be achieved, no diverity gain can be exploited, and vice vera. 2.1 Sytem exploiting patial diverity Receive diverity The optimal receive diverity technique in a Rayleigh channel i Maximal Ratio Combining (MRC), where the output i the weighted um of all received ignal component. The weight are choen baed on the SNR of the repective ignal component, o that the average SNR of the combined ignal i maximized. MRC yield the maximum diverity order N, auming N receive antenna [1]. Other receive diverity technique are Equal Gain Combining (EGC), where all ignal branche are weighted with the ame factor, and Selection Combining (SC), where only a ubet of antenna are ued for diverity combining. Space-time coding (STC) The dual technique known a tranmit diverity can be ued at the tranmitter ide if the channel tate i known at the tranmitter, which i not uually the cae. Hence, another coding technique that require no channel tate information at the tranmitter i applied in order to achieve diverity gain. It exploit pace and time diverity and i referred to a pace-time coding (STC). STC code are divided in two group, namely Space-Time Trelli Code (STTC) and Space-Time Block Code (STBC) [3]. STTC perform ymbol mapping via a trelli diagram, o that a modified Viterbi

3 26 Celcer, Plevel, Javornik, Fortuna, Kandu decoding algorithm can be ued at the receiver. Thee code achieve maximum diverity gain a well a coding gain, however, the decoding proce i very complex when dealing with greater number of tranmit antenna and higher modulation level, becaue thee are reflected in higher number of encoder tate. For thi reaon the orthogonal STBC introduced by Alamouti [5] ha gained in popularity, although no coding gain can be achieved. The cheme i compoed of two tranmit and an arbitrary number of receive antenna. The ource data are encoded in two ymbol period. In the firt interval, ymbol 1 i tranmitted over the firt antenna and ymbol 2 over the econd. In the econd interval, complex conjugated ymbol are tranmitted: 2 * over the firt antenna and 1 * over the econd. Decoding i performed with imple linear proceing at the receiver ide. Figure 1 depict a 2 1 Alamouti STC cheme. Figure 1: 2 1 Alamouti STC cheme 2.2 Sytem exploiting patial multiplexing Sytem exploiting patial diverity require multiple antenna at one ide of the communication channel only. However, if we want to achieve patial multiplexing gain by imultaneouly tranmitting multiple data tream over different antenna, multiple antenna are required at both ide of the channel, ince multiplexing gain can be at mot min(m,n). In [6], Fochini propoed a layered pace-time (LST) architecture referred to a V-BLAST (Vertical Bell lab LAyered Space-Time). The propoed cheme i hown in Figure 2. The input data tream i divided into M independent ubtream which are then mapped into modulated ignal and tranmitted imultaneouly over M tranmit antenna, each uing the ame frequency band. By tranmitting independent tream over each antenna, MIMO ytem can offer a linear increae in data rate within the ame bandwidth. Under uitable condition, uch a a rich cattering environment that reult in an uncorrelated Rayleigh fading channel, the receiver can eparate the data tream. Figure 2: MIMO ytem diagram with uncoded patial multiplexing The efficiency of uch a ytem i trongly dependent on the decoding algorithm elected, in which a compromie between complexity and optimality ha to be made. The optimal decoding algorithm i Maximum Likelihood (ML) decoding. ML detection require an exhautive earch over all poible combination of tranmitted ymbol and i often computationally too demanding. Other, uboptimal, approache are linear detection method like Zero Forcing (ZF) or Minimum Mean Square Error (MMSE) [3], iterative approache uch a Succeive Interference Cancellation (SIC) [7], and Sphere Detector [8]. 2.3 Adaptive MIMO ytem In a time varying channel, a tranmitter ha to be able to adapt to the change of the channel tate in order to optimize it performance [9]. Beamforming i a precoding ignal proceing method ued in multiple antenna ytem, and allow the formation of the deired beam in order to improve the ytem performance by cancelling out the interfering ignal and teering the beam in the direction of the choen uer. It i baed on knowledge of the intantaneou channel tate at the tranmitter, and can thu be adapted to change introduced into the channel [10]. Channel tate information (CSI) can be obtained via reciprocity (in TDD ytem) or via a return feedback link from the receiver. Under highly dynamic channel condition thi may not be poible, due to the delay in acquiring the CSI data. Moreover, the ue of reciprocity in TDD ytem require accurate calibration of all the element in the communication path, which i not eay to achieve in practice. The channel capacity would be optimally exploited if the MIMO ytem decided, dynamically, which of the three technique (diverity, patial multiplexing or beamforming) hould be ued. 3 IEEE tandard WiMAX i a metropolitan area network (MAN) wirele technology planned to provide high-peed communication over the lat mile. WiMAX i a ytem that conform to IEEE , a wirele communication tandard being tandardized in IEEE. For Fixed WiMAX, the (802.16d) [11] verion of tandard i ued. In 2005, an upgrade to IEEE wa approved under the form of IEEE

4 Multiple Antenna Technique in WiMAX Sytem e amendment, which add mobility feature to the original tandard [12]. The organization reponible for certification and interoperability of broadband wirele product i called WiMAX Forum. WiMAX Forum develop WiMAX ytem profile that define the mandatory and optional feature of the IEEE tandard that are neceary to build a WiMAX compliant air interface that can be certified by the WiMAX Forum. If we make a coare comparion of the functionalitie adopted in Fixed and Mobile WiMAX profile, we note that the main difference i in a multiple acce cheme and the upport of mobility. Fixed WiMAX upport OFDM a a phyical layer technology; however only one uer can tranmit at a time, while the multi-uer acce i done in the time domain. On the other hand, Mobile WiMAX upport Orthogonal Frequency Diviion Multiple Acce (OFDMA), which enable multiple uer to tranmit imultaneouly uing different carrier. Neverthele, ub-channelization can be implemented in Fixed WiMAX in order to achieve imultaneou multiple acce in the uplink. Apart from that, e tandard functionalitie enable full mobility, in contrat to d, in which mobility upport i limited to nomadic. 4 MIMO technique in WiMAX ytem One of the many advantage of OFDM technology lie in it robutne to multipath and the eae with which the multiple-antenna technique can be utilized to increae range and throughput [13]. Hence, WiMAX i ideally uited for operation in cluttered environment with high preence of ignal cattering, where benefit of multiple antenna ytem uch a diverity and patial multiplexing gain are ignificant. 4.1 Multiple antenna technique in Fixed WiMAX In Fixed WiMAX, the upport of MIMO technique i limited to a imple diverity cheme uing 2 1 Alamouti STC code (3), which provide maximum tranmit diverity gain for two antenna. Support for Beamforming (AAS Adaptive Antenna Sytem) i alo included in the OFDM-PHY; however, it i not ued in practice in the Fixed WiMAX certified equipment. 4.2 Multiple antenna technique in Mobile WiMAX Four multiple antenna technique are foreeen in Mobile WiMAX pecification, referred to a Matrix A, Matrix B, Adaptive Antenna Sytem and collaborative MIMO. Matrix A In Mobile WiMAX, the Alamouti STC cheme, propoed already in Fixed WiMAX ytem profile, i referred to a Matrix A: * 1 2 = * 2 1 A, (3) where the row repreent the tranmit antenna and the column repreent the ymbol period. Matrix A with two tranmit antenna i mandatory for Wave 2 certification of WiMAX ytem, while other STBC cheme for three or four tranmit antenna are optional. The Alamouti diverity cheme increae ytem reliability by mitigating fluctuation in the received ignal power, o Matrix A i very appropriate when the uer i highly mobile, with rapid ignal fading and multipath reception. The reduced fade margin allow the ue of a higher modulation level, cauing a certain increae in capacity a well. Matrix B While MIMO Matrix A implement rate 1 STC, MIMO Matrix B ue patial multiplexing. Support of Matrix B with two tranmit antenna i mandatory for Wave 2 certification: = 2 1 B. (4) It i a 2 2 MIMO technique where each ymbol i ent only once, o no redundancy i introduced during tranmiion. That mean that no diverity gain can be achieved, but two ymbol are ent in each ymbol period, which, under uitable condition, double the data rate. A explained in Section 2, patial multiplexing require at leat a many receive antenna a the number of independent data tream, o in thi cenario at leat two receive antenna at the mobile tation (MS) ide are required. The efficiency of Matrix B MIMO depend greatly on the preence of natural multipath, which implie independent fading of different path. Under thi condition two received ignal are not correlated and can be uccefully ditinguihed at the receiver. Thi kind of channel characteritic i common in dene urban area while, in area with LOS condition and limited multipath, Matrix B will perform poorly due to high ignal correlation. Although Matrix B enable a ignificant capacity increae, the ue of STC/MRC cheme (Matrix A in combination with MRC) offer better performance at low SNR. Since the capacity grow logarithmically with SNR, it prove more convenient to increae SNR in order to gain capacity a well a robutne. However,

5 28 Celcer, Plevel, Javornik, Fortuna, Kandu at high SNR, the linear nature of capacity growth enable Matrix B to outperform Matrix A. Adaptive Antenna Sytem (AAS) If more than two antenna are implemented at the bae tation, the additional degree of freedom can be utilized. AAS can ue beamforming in order to hape the tranmitted beam and to patially eparate different ignal coming from different direction. Beamforming can be ued, either in combination with either Matrix A or B or a a tand-alone technique, in order to eparate ignal from different uer and decreae interference. Again, the channel tate information ha to be available at the tranmitter, which can be a very demanding tak, epecially in a rapidly changing mobile environment. In practice, beamforming i more appropriate when there i a predominant angle of arrival, o the ytem can form an appropriate radiation pattern. Thu it will not be very efficient in Rayleigh fading environment (high multipath and no LOS component). Collaborative MIMO Collaborative MIMO i ued in the uplink a a requiite for increaed capacity of the ytem a a whole. It doe not reult in any per uer data rate increae, but it can double the cumulative capacity of the ector. Collaborative MIMO ue Space-Diviion Multiple Acce (SDMA), o that two ubcriber tation, each equipped with a ingle antenna, can tranmit imultaneouly uing the ame OFDM ubcarrier. Thi i imilar to the patial multiplexing ued in downlink with Matrix B, except that the tranmitter are well eparated in pace and thu the correlation i much lower. 5 WiMAX ytem meaurement In thi ection we decribe tet reult of Telima indoor Fixed WiMAX deployment with MRC and STC in Bangalore, India. Baed on the terrain meaurement we have developed a dedicated channel model and have performed extenive tet of MRC and STC performance with the channel imulator in the laboratory. Note that, at the bae tation ide, cro-polarized antenna were ued in order to exploit polarization diverity ince operator prefer to intall only a ingle antenna, intead of intalling two BS antenna eparated in pace in order to exploit patial diverity. The field meaurement actually howed that the gain of STC and MRC obtained with one cro-polarized antenna at the BS were imilar to thoe obtained with two patially eparated antenna. 5.1 Field meaurement reult Figure 3 how the average increae in CINR when STC i ued in the downlink, compared to the normal operation. Each location index correpond to one indoor tet location. The increae in average carrier to interference-plu-noie ratio (CINR) i around 4.5 db. A 3 db increae in CINR i achieved due to a twofold increae in tranmitted power. The remaining increae in CINR i due to lower frequency electivity in STC mode than in imple SISO communication. Very imilar reult were alo obtained for the MRC teting in the uplink, where the baic 3 db gain i achieved due to coherent combination of the received ignal. CINR improvement [db] Location Index Figure 3: Average STC gain in CINR High gain of average CINR can be achieved with STC (Figure 3); however, the gain meaured at different location can vary ignificantly. For ome location a gain of only 2 db i achieved, while for other location it i over 8 db. Thi can be explained by the fact that, at ome indoor ubcriber tation (SS) location, the ignal from the primary BS antenna i tronger than the one from the econdary antenna, while for other location the ignal from the econdary BS antenna i much tronger. Since the STC or MRC gain i actually a gain in db relative to the received CINR when only a primary tranmit antenna i ued, the gain are very different. Even a mall change of the indoor modem antenna poition reult in a ignificant variation in gain. However, the mot important obervation in the field wa that variation of the ignal quality i much lower when STC or MRC i ued. It wa alo etablihed that the gain of STC and MRC are highet in location where CINR i lowet, ince at thee location the primary antenna ignal i low. 5.2 Laboratory meaurement with a channel imulator Baed on the obervation of ignal variation, frequency electivity and correlation in the field, we modified a tandard SUI-3 channel model in order to imulate the non-los (NLOS) indoor Bangalore channel that wa oberved in Telima actual WiMAX network. There are three tap in the developed channel. The firt i Rice ditributed with Rice K factor 9. The econd and third tap are Rayleigh ditributed, with 5 db attenuation relative to the firt tap, and are delayed by 300 n and 900 n, repectively. Thi reult in a trong frequency

6 Multiple Antenna Technique in WiMAX Sytem 29 elective channel. The correlation of the channel correponding to each BS antenna wa et to 0.2. An Elektrobit Propim C2 Wideband Radio Channel Simulator wa ued in the Telima laboratory to imulate the channel. PER 1E+00 1E-01 1E-02 1E-03 1E avg RSSI [dbm] QPSK_1/2_noSTC QPSK_1/2_STC QPSK_3/4_noSTC QPSK_3/4_STC QAM16_1/2_noSTC QAM16_1/2_STC QAM16_3/4_noSTC QAM16_3/4_STC QAM64_2/3_noSTC QAM64_2/3_STC Figure 4: Packet Error Rate (PER) improvement with STC The Packet Error Rate (PER) veru average input ignal level (RSSI), for STC and normal non-stc (SISO) operation for different coding and modulation level, i demontrated in Figure 4. A channel bandwidth of 3.5 MHz wa ued and the packet ize wa 1460 Byte. A hown in (2), the difference in gradient of the PER curve repreent the diverity gain. The meaurement reult prove that high diverity gain are achieved, ince the PER decay i much fater for STC operation. For the 16QAM ¾ cheme it can be een that the gradient of the lope i nearly doubled, meaning that in thi cae a diverity gain cloe to 2 i achieved, which i the maximum poible gain achievable with two antenna. From Figure 4 it can be etimated that for PER 10-2 the gain of STC are around 5 db for QPSK ½, 7 db for QPSK ¾, 5 db for 16QAM ½, 10 db for 16QAM ¾ and 7 db for 64QAM R. It may eem urpriing that the gain and the lope are o different for different mode of operation. The reaon lie in the frequency electivity. Since higher order modulation are more enitive to frequency electivity, the gain of STC are higher for thoe modulation. But, more importantly, the Forward Error Correction (FEC) add robutne to frequency electivity, utilizing frequency diverity. It can be een clearly from Figure 4 that, in non-stc (SISO) operation, curve repreenting higher redundancy FEC mode are teeper than thoe for lower redundancy one. Thi i becaue tronger FEC coding exploit frequency diverity better. For example, comparing non-stc QPSK ½ and QPSK ¾, it can be een that QPSK ¾ perform quite poorly. Even more urpriingly, without STC at a lower PER area, the 16QAM ½ even outperform the QPSK ¾, although it ha higher pectral efficiency. The high redundancy FEC coding (e.g. ½) already effectively exploit the frequency diverity gain, o the gain of STC are lower for thoe mode. 6 Concluion Multiple antenna ytem, their main characteritic and their ue in both Fixed and Mobile WiMAX, have been briefly reviewed. The performance of Telima Fixed WiMAX ytem upporting STC in downlink and MRC in uplink ha been decribed and analyzed. The field reult how a great improvement of ignal quality with STC and MRC, but the gain differ greatly for different indoor location. The average gain in CINR oberved in the field wa 4.5 db and the highet over 10 db. Baed on the reult of field meaurement, a dedicated channel model wa developed in order to imulate the real propagation channel, either by computer or by hardware channel imulator. Laboratory meaurement done with the channel imulator, uing thi model, howed that the gain of STC i highly dependent on the PER at which it i oberved, and on the coding and modulation mode ued. The meaurement reult proved that diverity gain i achieved in all cae, ince the PER lope i much teeper in STC mode. It wa alo revealed that, in general in a dynamic channel, the coding-modulation mode with higher redundancy perform much better than thoe with lower redundancy. Since the performance of lower redundancy code i poor in uch a demanding channel, the gain of STC and MRC are greatet for thoe mode. The highet gain meaured wa for 16QAM ¾, i.e. 10 db at PER 10-2 and 12 db at Acknowledgement Thi work ha been upported in part by Telima d.o.o. 7 Reference [1] A. Goldmith, Wirele Communication, Cambridge Univerity Pre, [2] I. E. Telatar, Capacity of multi-antenna Gauian channel, European Tran. Telecomm., vol 10, pp , Nov [3] B. Vucetic, J. Yuan, Space-time coding, John Wiley and Son Inc., [4] L. Zheng and D. Te, Diverity and Multiplexing: a Fundamental Tradeoff in Multiple Antenna Channel, IEEE Tran. Inf. Theory, vol. 49, no. 5, pp , May [5] S. M. Alamouti, A Simple Tranmit Diverity Technique for Wirele Communication, IEEE J. Select. Area Commun., vol. 16, no. 10, pp , Oct [6] G. J. Fochini, Layered Space-Time Architecture for Wirele Communication in a Fading Environment When Uing Multiple Antenna, Bell Lab Tech. J., pp , Autumn 1996.

7 30 Celcer, Plevel, Javornik, Fortuna, Kandu [7] P. W. Wolnianky, G. J. Fochini, G. D. Golden, R. A. Valenzuela, V-BLAST: An Architecture for Realizing Very High Data Rate Over the Rich-Scattering Wirele Channel, invited paper, Proceeding of ISSSE Conference, Pia, Italy, Sep [8] B. Haibi, H. Vikalo, On the Sphere Decoding Algorithm. I. Expected Complexity, IEEE Tran. on Sig. Proc., vol. 53, no. 8, pp , Aug [9] L. Hanzo, C. H. Wong, M. S. Yee, Adaptive Wirele Tranceiver Turbo-Coded, Turbo-Equalized and Space-Time Coded TDMA, CDMA and OFDM Sytem, John Wiley & Son, [10] E. Biglieri, R. Calderbank, A. Contantinide, A. Goldmidth, A. Paulraj, H. V. Poor, MIMO Wirele Communication, Cambridge Univerity Pre, [11] IEEE Standard , IEEE Standard for Local and Metropolitan Area Network - Part 16: Air Interface for Fixed Wirele Acce Sytem. [12] IEEE Standard e-2005, Amendment to IEEE Standard for Local and Metropolitan Area Network - Part 16: Air Interface for Fixed Broadband Wirele Acce Sytem - Phyical and Medium Acce Control Layer for Combined Fixed and Mobile Operation in Licened Band [13] A. R. S. Bahai, B. R., Saltzberg, M. Ergen, Multi Carrier Digital Communication: Theory and Application of OFDM, Springer, Tine Celcer received hi B.Sc. degree in electrical engineering from the Univerity of Ljubljana, Slovenia, in Currently, he i a Ph.D. tudent and a Junior Reearcher in the Department of Communication Sytem at the Jožef Stefan Intitute. Hi reearch interet are in the field of tratopheric and terretrial wirele communication ytem, pecializing in MIMO ytem, OFDM technology and WiMAX ytem. Sre'o Plevel received hi B.Sc. degree in 2002 at the Faculty of Computer and Information Science, Univerity of Ljubljana and hi Ph.D. degree at the Faculty of Electrical Engineering, Univerity of Ljubljana, Slovenia in Currently, he work a a enior RF deigner at Telima d.o.o., company involved in WiMAX deployment. Prior to that, he wa a Junior Reearcher in the Department of Communication Sytem at the Jožef Stefan Intitute. Tomaž Javornik received hi B.Sc., M.Sc., and Ph.D. degree in electrical engineering from the Univerity of Ljubljana, Slovenia, in 1987, 1990 and 1993, repectively. He joined the Jožef Stefan Intitute in 1987, where he currently work a a reearcher in the Department of Communication Sytem. He i involved in the tudy of digital radio-relay ytem, modulation technique, coding, adaptive ignal proceing and digital mobile communication ytem. Carolina Fortuna received her B.Sc. degree in electrical engineering from the Univerity of Cluj-Napoca, Romania, in Currently, he i a Ph.D. tudent and a Junior Reearcher in the Department of Communication Sytem at the Jožef Stefan Intitute. Her reearch interet are in the field of tratopheric and terretrial wirele communication ytem and intruion detection ytem. Gorazd Kandu received hi B.Sc., M.Sc. and Ph.D. degree in electrical engineering from the Univerity of Ljubljana, Slovenia, in 1971, 1974 and 1991, repectively. He i currently the Head of the Department of Communication Sytem at the Jožef Stefan Intitute and Profeor at the Faculty of Electrical Engineering, Computer Science and Information Technology, Univerity of Maribor.