Tap chi Khoa h^c v^ Cong nghf 102 (2014) 043-049 A Comparision of 3GPP and Geometrical Channel Models on Correlated MIMO-OFDMA System So sanh mo hinh kenh 3GPP va mo hinh kenh vong tron he thong kenh hrong quan MIMO-OFDMA Nga Nguyen^ Bach Tran^, Van Due Nguyen''^* 'School of Electronics and Telecommunications, ^Vietnamese-German Center for Science and Technology Cooperation Hanoi University ofScience and Technology - No 1, Dai Co Viet Str., Hai Ba Trung Ha Noi, Viet Nam Received: May 6, 2014; accepted: August 25, 2014 Abstract This paper investigates the performance of the multi Input multi output orthogonal frequency division multiplexing access (MIMO-OFDMA) system in two geometrical stochastic and correlation stochastic models Performance symbol error rate (SER) of MIMO systems in realistic radio environments greatly depends on the spatial correlation because of the fading or cluster of scatterers in the multipath propagation environments It is assumed that all antennas have same radiation pattern and the spatial correlation depends on the position of an antenna arrays The SER performance results are depend on the correlation ratio of the channel The number of users in MAC frames is also calculated in the system Based on an algorithm of Vertical Bel! Laboratories Layered Space Time (V-BLAST), a successive cancellation method applied to the receiver we obtain the high SNR in correlation channel Our results are simulated on both of the Extended One-Ring channel model and the Spatial Correlation Model (SCM) based on long term evolution (LTE) standard for 2x2 MIMO-OFDMA system Keywords: MIMO-OFDMA, SCM, V-BLAST, ZF, Correlated channel Tom tat Bdi bdo khao sat hieu nSng h$ thing da anten phat da anten thu da truy nh$p phan chia theo tin so trirc giao (MIMO-OFDMA) di/a tren hai mo hinh kenh vong trdn va mo hinh kenh tuvng quan Ti so loi ki tij' (SER) cOa h§ thing da anten phat da anten thu mdi trwdng vo tuyin thi/c ti phu thuoc v^o tirang quan khong gian fading tin hieu ho$c chOm tan xa moi trudng da dudng truyin Gia thiit la cac anten cd phat xa nhw va ttrang quan khong gian phu thupc vi tri cua cac anten Do vay hi$u nang h§ thing cung phu thudc v^o he sd tuang quan kenh truyin Di/a tren giai thuat mS hoa VBLAST, mdt thuat todn tri^t nhieu hi$u qua du^ them vao b§n thu di dat duac ti s6 tin hieu/nhieu cao Cac kit qua duvc mo phdng dus tren md hinh tan xa mpt vong tron One-ring va mo hinh kenh tuang quan khong gian (SCM) v&i chuin LTE he thing anten phat'2 anten thu MIMO-OFDMA TLF khda: MIMO-OFDMA, SCM, V-BLAST, ZF, kenh tu'cng quan, 1, Introduction interference by using diversity methods Based on An orthogonal frequency division multiple access (OFDMA) technique is in [1] with proposed decentralized dynamic sub-carriers assignment algoridrai for multi-hop OFDMA time division duplexing networks (OFDMA/TDD) to solve the problem of hidden node and the threshold of accessing mnlti input multi ouJutorthoEonal frequency divisim '^f'Pfj"'S (MIMO-OFDM) the system can be achieved in both spatial and ircquency domains In [3] """""•"'i ' ' " ' " ' * = miHi-users detector and channel estimation of MIMO- SDMA-OFDM (space division multiple access) in Rayleigh fading channels, ^')'^ ^T'^J'J.?,!" F ' ' ' * ! , " " " ' ''""TT ''°' multi-cell OFDMA trequency division duplexmg ^ /,-M7rvx.( A /r7i-.TvJ c u I II • networks (OFDMA/FDD) Some channel allocation methods for multi-cells OFDMA networks with a reused frequency solve the interference problem that eliminates the intra-cell interference so increase the throughput of the sysiem V-BLAST in Rayleigh fading chamiel model i • rA^ , -.u •,• A- , u • • • i •" l^] ^ * " cognitive radio technique m single earner , , T , n ,~,r-i-.ô* i , m AC^T Jã ã r^-, r,n r ^ , ^ ; ? n ? t,'^ K ; n i m "^^'.' T ^^^ ^^^ ^"^ ^ ' J ^ ^ V 1^^' ' ' " ' " ° ^^^^; ^^'^^ detectors « ^ ^^^ rich-scattermg wireless or fading Rayleigh ^"^^""^'^ VBLAST method not only enables a complete combat the interference but also a communication technique to achieve high data rate streams which transmitted simultaneously from multiple antennas On the other hand, with the joint of MIMO enables application of multi-user and mitigates the ' Corresponding Author: Tel: (+84) 989,145.909 Email: due nguy en van l@hust, edu.vn ap ctii Knoa noc va »_ong ngn? iti: The correlation at the receiver and the transmitter side can considerably reduce tbe performance of the MIMO systems The correlation properties of MIMO - OFDM channel are investigated in suburban macrocell environment as [7] vidth some coding techniques such as space time block coding (STBC), V-BLAST, space-frequency block coding (SFBC) over SCM LTE, The geometrical one-ring with HIPERLAN2 standard and its spatial - temporal correlation function of MIMO - OFDM system with coding STBC are studied in [8], Tbe extended one-nng scattering model for frequency selective channel with vehicular A (EVA) - ITU model performance of transmit diversity algorithm in MIMO-OFDM systems with perfect channel State Infomiation (CSI) is in [9] The Fourth Generation Long Term Evolution (4G- LTE) cellular network uses the MIMO-OFDM technology for the downlink fransmission channel and the LTE performance has been studied in [10] [II] [12] Reference in [13], [14] study the SCM, which is designed for multiple antenna systems and proposed by The Third Generation Partnership Project (3GPP) with set of parameters of angles and environments for simulation Authors in [15] proposed the dynamic channel allocation (DCA) algorithm for MIMO - OFDMA with the using of ZF or MMSE equalizer, but not using VBLAST, in fading channel as Monte Carlo with HYPERLAN2 standard From that paper, maximum SNR sub-carriers are selected for the corresponding communication link so that the system performance can be significantly improved Authors in [16] study the geometrical One-ring model with HYPERLAN2 or LTE standard for MIMO - OFDMA using the ZF equalizer combined to VBLAST algorithm Userl DCA > ^ System model with correlated channel 2.1 System model We assume that the channel has perfect CSIi the exact position of MS is feedback to BS ; power conttol of the system is equal to the amoimi loss of the distance, at the receiver the chamiEl recovered in the best condition Multiusers MIMO OFDM system with Q w is assumed that both ttansimtter and receiver anten are two Based on DCA in [15] the stteam of sym pass through is assigned to each OFDMframei then modulated OFDM symbol After that the sym is inserted the guard interval, mapped and then h Q set of transmit antermas (Txl and Tx2), ' transmitter architecture of MIMO- OFDMA isini I At the receiver, from the Q set of rec£ antennas (Rxl and Rx2), the cyclic extensioi removable as it contains the channel spread and is mapped Then the FFT is taken in each of received antenna which receives a different i superimposition of faded versions of transmil signals The data passes to the de-multiplexing in( sub-stteams of equal length and then maps l symbol by the same constellation Tbe sub-ste then passes through signal processing or MIMl OFDM- VBLAST demodulation Again, the signal is allocated by subcai selection for MS in DCA scheme as in [15] befon to users MIMO-OFDMA receiver architecture using V-BLAST algorithm is shown in Fig.2 -4^ IFFTl llscr2 [> -^ FFT User Q -> -^ The structiu^ of this paper is as follow: Seci 2, we briefly infroduce MIMO- OFDMA syj model with correlated and uncorrelated channel two kinds of LTE channel models Section detection algorithms are studied Section simulation results and discussion are shown Fiaa conclusion is offered m Section IFFT Q Fig MIMO-OFDMA transmitter Fig MIMO-OFDMA receiver Tap chi Khoa hoc v^ Cdng nghf 102 (2014) 043-049 2.2 Correlated channel of SCM In [14], for an element Hnear BS array and a U element linear M S array, the chaimel coefficients for one of A'^ multipath components are given by a f/ -byS matrix of complex amplitudes We denote tbe correlation fimction between the two channels is a fimction of the distance of antenna elements in both transmit and receive sides which is described in [7], For the n"' multipath component (w = / N), the (u,s)''' component (s = S; u = 1, ,U) the normalized complex correlation - the spatial-temporal correlation fiinction is given as follow: (Ads, Adu, T ) = i ex^{jkM,sm{e„_^j^^o)) expOfcAd,sin(0„,^,,„^)) exp[;'fc||t;||cos(e„,^,^,^- , ) T ] (1) where Arfj and Arfu are the distances of ttansmitter and receiver antenna, respectively; T is the time delay of the channel When r = 0, the spatial cross correlation fimction (CCF) of wideband MIMO channel is given as follow: Ln=i ^ Lm=i {exp(jkAd^sin(6„,^,^,^))} ^ ' two channel models above to estimate the correlation charmel in the MIMO - OFDMA system The SCM with parameters for suburban macrocell environment with A'^ paths, sub-paths per path M= 20, bandwidth B < MHz as shown in reference [7], [12] and [13] 2.3 Correlated channel of Extend One-ring One-ring model is studied in [8] with HiperLAN standard Extended One-ring Channel Model is studied in [9] with coefficient, the complex channel gain, space-time-frequency cross correlation fiinction (CCF) of wideband MIMO channel with ttansmitand receive- antennas are also investigated One-ring MIMO channel model is one kind of geometticallybased stochastic models from which scattering is modelled with stochastic disttibution by applying the reflection, diffraction, and scattering angle of the signals In [11], the extended geometrical one-ring scattenng model for MIMO- OFDM system in a conditional of frequency selective channel is experiment with vehicular A (EVA) - ITU model performance of transmit diversity algorithm in the case of perfect CSL The space-time-frequency cross correlation fiinction (CCF) of wideband MIMO channel with MBS ttansmit- and M^s receive- antennas in [8], [11] is given by: p'^'^I^CAd^.AdJ = iSJl'=i,S = -T-n^i^l^^i " n , i ^ expyn^ Pslu^iAd^) exp{jkAd,sinli0,^^_^,^)) [cosia^oo) H (3) If Adu = 0- then Z^.i52!^.ie^p(y'fcAd,5in(0„,^,^„o)) (6) Ady = exp 1/77 -CCAOA)\ (7) The antenna element spacings at the MS and BS are Ad„, Ad^, respectively, or^o^^ a^gj, denote the multielement antenna tilt angles at the MS and BS, The quantity 0^'"'is the angle of arrival (AOA) of the n"' incoming wave at the MS, the corresponding angle of departure (AOD) denoted hy