Study On Signal Detection Technology Of MIMO-OFDM Systems In Time-varying Channels

Multiple input Multiple output (MIMO) technology has the advantage ofimproving the capacity of wireless communication system, while Orthogonal frequencyDivision Multiplexing (OFDM) technology can effectively against frequency selectivefading. MIMO-OFDM system which combines the good features of MIMO and OFDM,is regarded as the best choice of the future broadband wireless mobile communication.But the performance of MIMO-OFDM system is seriously degraded by time-varyingchannels, since the orthogonality between the OFDM sub-carrier is destroyed. Thesignal detection algorithms play a vital role in system performance. So it is important toresearch on the effective signal detection algorithms for the MIMO-OFDM systemworking in time-varying channels.The thesis started with the MIMO-OFDM system model and time-varying channel.With the analysis of time-varying channels’ influence on MIMO-OFDM system, thethesis focused on the iterative detection algorithms which behaved well on combatingthe system performance degradation caused by channel estimation error and inter-carrierinterference in time-varying channels. The main results are summarized as followed.①To get the quantitative analysis of ICI in mobile MIMO-OFDM system, thethesis presented the MIMO-OFDM system model and the TDL model of multipathchannel at first. The time-varying channel estimation methods about linear time-varyingmodel and basis expansion model are also described. And the model errors betweendifferent BEM were compared by simulation results.②the thesis compared the characteristics between different detection algorithm ofMIMO system through the theoretical analysis and simulation. The algorithms includedmaximum likelihood detection (ML), zero force detection (ZF), minimum mean squareerror detection(MMSE), ordered successive interference cancellation detection (OSIC)and parallel interference cancellation detection (PIC). To reduce unexpected effects ofthe channel estimation errors in time varying channels, the thesis adopted a MMSE-SICdetection scheme to distinguish received signals. In this method, a better nulling weightvector which minimizes the power of various terms generated by the channel estimationerrors was derived. The symbol ordering operation was also modified by choosing theminimum mean-square error symbol at each iteration. The simulation results show that,compared with the traditional detection scheme, the improved MMSE-SIC detection algorithm can effectively improve the system performance with the channel estimationerror.③In consideration of the ICI caused by time-varying channel and the deficiencyof PIC-DSC iterative detection scheme, the thesis put forward a MMSE-PIC-DSCdetection scheme. It utilized the channel frequency response and the MMSE criterion todesign detection filter. By using diagonal band-limited structure of ICI interferencereduction, the detection complexity was effectively reduced. The simulation resultsshow that, compared with PIC-DSC method, the proposed MMSE-PIC-DSC detectionalgorithm is not sensitive to time-varying channel estimation error, and effectivelyimproves the bit error rate performance of MIMO-OFDM system in time-varyingchannels.