Graph based low complexity MIMO detection with intermediate hard-decision

Recently, a novel algorithm for low complexity, iterative MIMO detection has been devised based on pair-wise Markov random fields (MRF) in [11], where for two types of MRF, namely, the fully-connected and the ring-type MRF, they developed belief propagation based iterative algorithm run over the corresponding factor graphs. Since the factor graphs have only 2 edges per factor node, the computations are much easier than that of ML. In this paper, we further simplify the algorithm by making hard-decision in intermediate message passing. By doing so, the complexity can be further reduced at only a small performance degradation.