Low Complexity Variational Bayes Iterative Receiver for MIMO-OFDM Systems

A low complexity iterative receiver is proposed in this paper for MIMO-OFDM systems in time-varying multi-path channel based on the variational Bayes (VB) method. According to the VB method, the estimation algorithms of the signal distribution and the channel distribution are derived for the receiver. With the aid of the soft-output QRD-M algorithm, whose complexity is fixed and relatively low, the signal distribution can be obtained conveniently. In particular, a sequential channel estimation algorithm, which completely avoids the computation of matrix inversion and multiplication, is introduced for the channel distribution estimation. Moreover, the distribution estimations of the signals and the channels are performed in a cyclical iteration way. The simulation results show that the performance loss of the proposed receiver is only ldB for fast varying channels and less than 0.5 dB for slow varying channels at the bit error rate of 10^-4 after 3 iterations, compared with the optimum receiver with perfect channel state information.