Research For Synchronization, Channel Estimation And MIMO Decoding In SDM MIMO OFDM System

With the wireless data and multimedia application growing rapidly recently, the demands for high system capacity and QoS guarantee have become more and more urgent. Orthogonal Frequency Division Multiplexing (OFDM), as a promising transmission technique in PHY layer, has been attached much more research emphases and has been considered as the key component for the next generation wireless communication (4G).OFDM inherently attains the merits of Multiple Carrier Modulation (MCM), i.e. efficient parallel modulation and long symbol period, which makes the accurate symbol synchronization in assistance with the Guard Interval or Cyclic Prefix for the OFDM receiver. Meanwhile, with the parallel modulation, OFDM could utilize a frequency selective fading channel as several flat fading channels, thus improving the transmission efficiency tremendously.Based on the deep research into the reachable capacity of multi-path in space, Multiple Input Multiple Output (MIMO) has attracted enormous research interest. Among the many kinds of MIMO systems, the Space Division Multiplexing (SDM) MIMO system could make full uses of the multi-path scatter channel and transmit several independent data stream parallel, thus enlarging the system capacity greatly in the same expense of transmission power and bandwidth.Combined with several OFDM transceivers at the base-band parallel, the conventional single-carrier SDM MIMO system could be extended to SDM MIMO OFDM system. SDM MIMO OFDM system takes both advantages of OFDM and MIMO techniques, which are keeping the robust performance in the frequency-selective fading channel by OFDM and increasing the system capacity enormously by MIMO.The research work in this thesis begin with the base-band SISO OFDM system in which we make deep investigation into the frame synchronization, frequency synchronization as well as symbol synchronization, including algorithms simulation and hardware implementation.In the frequency synchronization, we design a combined correlation average and angle average structure to estimate the frequency offset accurately. Simulation results demonstrate that the ML frequency offset estimator utilizing the novel combined structure achieves more accurate synchronization results in comparison with utilizing the conventional correlation average only. Within the same requirement of estimation variance error, the novel could save up 2 — 3dB SNR approximately. After the implementation of the new designed frequency synchronization processing on FPGA broad, the test results on 6 sample frequency offset points between 100kHz and 200kHz show at most 0.33% estimation error compared to the simulation on Matlab.To eliminate the influence from the frequency offset, we design a novel symmetrical crosscorrelation scheme for symbol synchronization processing, which could remove the interference from frequency offset totally. Simulation results demonstrate that the novel synchronization scheme could capture the accurate symbol synchronization position even in the influence of a large frequency offset. Meanwhile the symmetrical cross correlation scheme is irrelevant to the specific preamble content and could be applied to many kinds of OFDM system directly, including both 802.16a and 802.11a standard.Based on the research experience on the synchronization processing in the SISO OFDM receiver, we further extend this research into 4*4 SDM MIMO OFDM system, in which we design a robust multiple receiver joint synchronization structure. The joint synchronization structure could make full use the training data on different receiver to get more accurate synchronization result compared to synchronization result on a single receiver separately. In the frame synchronization, we utilize RS, EGC and Hybrid RS/EGC algorithms for multiple receiver joint synchronization, and in the frequency synchronization, on the basis of the RS, EGC and Hybrid RS/EGC we design Hybrid OS/EGC as well as Hybrid OS/OC for improvement. Simulation results show that in the 4*4 system, the Hybrid OS/OC joint frequency synchronization scheme with only 2 receivers could achieve almost the same accurate estimation results as the EGC joint scheme which needs to utilize 4 receivers simultaneously.MIMO channel estimation and MIMO decoding are two critical components in SDM MIMO OFDM system. In this thesis, we first build a 4*4 SDM MIMO OFDM simulation model compatible to the 1EEE802.1 la specification and make comprehensive comparison and simulation of different MIMO channel estimation and MIMO decoding algorithms in further step. At the same time we also show the attainable system capacity of the 4*4 SDM system after extending the 802.1 la standard OFDM system.