Low Complexity Semi-Blind Bayesian Iterative Receiver for MIMO-OFDM Systems

Based on the variational Bayes expectation-maximization (VBEM) algorithm, a low complexity semi-blind Bayesian iterative receiver with joint signal detection and channel tracking is proposed in this paper for MIMO-OFDM systems over time-varying multi-path channels. Since the VBEM algorithm provides distribution estimation of all parameters, the detection performance can be improved by taking the channel estimation error into account. In addition, with the aid of the soft information provided by the signal detector, the recursive VBEM (RVBEM) algorithm is introduced to track the time-varying channels. Due to the high complexity of the RVBEM algorithm, a novel time-frequency domain recursive VBEM (TF-LCRVBEM) algorithm with low complexity is further proposed. The TFLCRVBEM algorithm simply predicts the channel impulse responses (CIRs) on time domain and recursively refines them on all subcarriers. The complexity analysis results demonstrate that the TF-LCRVBEM algorithm totally avoids computation of matrix inversion and obtains linear complexity. Moreover, the simulation results show that the proposed receiver not only dramatically outperforms the conventional receiver, but also provides performance close to the optimal receiver with perfect channel state information (PCSI).