Hybrid MCMC and LMMSE Detector for MIMO Frequency-Selective Channels

Optimal detection in multiple-input multiple-output (MIMO) frequency-selective systems is known to have exponen- tial complexity due to the number of transmitter antennas and channel length. In this paper, we model the detection problem using factor graphs and apply the sum-product algorithm (SPA) to derive the optimal detector. Then we adapt the SPA to propose a hybrid suboptimal algorithm based on two known detectors: the Markov chain Monte Carlo (MCMC) detector initialized with the solution from the linear minimum mean square error (LMMSE) detector. The proposed algorithm achieves better performance than any of the two individually while preserving their lower complexity.