The Detection Algorithm And Optimization On Generalized MIMO-OFDM Systems

Generally, the next generation mobile communication systems are calledB3G or 4G systems of which MIMO-OFDM is accepted as the basicframework. The channel capacity of Multiple Input Multiple Output (MIMO)systems increases approximately linearly with the minimum of the numbersof the transmit and receive antennas at high SNR. Orthogonal FrequencyDivision Multiplexing (OFDM) is an efficient and simple technique toovercome the effect of frequency selectivity caused by multipaths. Butordinary MIMO-OFDM can only obtain the space multiplexing gains, so weshould apply STBC/SFBC code to the system in order to obtain all theencode, multiplexing and diversity gains. More importantly, we shoulddesign efficient and dependable receive detection algorithms to exertgeneralized MIMO-OFDM's potencies.The main contents for this thesis are about the detection algorithms ofgeneralized MIMO-OFDM systems, as follows:1．Proposes systematically the extension method over MIMO-OFDM system,derives Multistage Parallel Interference Cancellation (MPIC) andsimplified Iterative Successive Interference Cancellation (SISIC)algorithm. Analyzes and obtains generalized MIMO-OFDM architecturewhich can be adaptive to different communication environment. Alsolinear and nonlinear detection (NLD) algorithms on generalizedMIMO-OFDM system are derived by combining both VBLAST andSTBC ideas. Proposes MPIC algorithm, which only have littleperformance loss but reduces computation complexity greatly bycancelling most of the matrix inversion or sort operation. Proposes SISICalgorithm, which can also reduces complexity by removing unnecessaryiterative process. Computer simulation and quantitative complexityanalysis proves the results above.2．Proposes simplified SDR detection (SSDR) algorithm based onpreprocessing and simplified PDA detection (SPDA) algorithm based onauxiliary variables while deriving relevant Turbo iterative detectionalgorithms. Applying quasi-maximum likelihood (QMLD) detection ideas,including semidefinite relaxation (SDR),probability data association(PDA) detection, to generalized MIMO-OFDM system, we proposes SSDR and SPDA algorithm. With the two proposed algorithm, the systemcan approach ML performance and enjoy a lower complexity comparedwith ordinary SDR and PDA algorithms. Also extends the applied range ofPDA algorithm by introducing MCMC theory, which can produce acomprehensive PDA algorithm. Researches the key problems on theiterative detection on generalized MIMO-OFDM system, proposes theiterative detection algorithm by applying simplified SDR and PDAalgorithm to the iterative configure, the proposed iterative algorithm canachieve a good performance. Computer simulation and quantitativecomplexity analysis proves the results above.3．Proposes combination algorithm of NLD and QMLD algorithms. Derivestwo kinds of combination algorithm based on SSDR algorithm, SPDAalgorithm and non-linear algorithm. By performance analysis andsimulation, we also derives optimal combination algorithm. Computersimulation and quantitative complexity analysis proves the results above.4．Proposes channel estimation algorithm for generalized MIMO-OFDMsystem. By analyzing the channel estimation and interpolation algorithmof MIMO system, we derives channel estimation algorithm forgeneralized MIMO-OFDM system.