Approximate Belief-Selective Propagation Detector for Massive MIMO Systems

When faced with challenging antenna configu-rations or high-order modulations in realistic propagationenvironments, the Belief Propagation (BP) MIMO detector out-performs its linear counterparts. To mitigate the error floor issueand lower the complexity, a revised BP detector, named theBelief-selective Propagation (BsP) detector, has recently emergedby selectively utilizing trusted incoming messages for updates.Despite those promising potentials, the straightforward hardwareimplementation of the BsP detector still suffers from highcomplexity and necessitates further optimization. To bridge thegap between the BsP algorithm and implementation, this paperintroduces anapproximatebut implementation-friendly BsPdetector called aBsP, based on which the very first BsP hardwareis proposed. Two unexplored features:approximate initializa-tionandsimplified message updatessave the complexity (morethan 84%) with acceptable performance penalization. Multi-level optimization techniques involving group-layered messageupdating, approximate arithmetic circuits, and hybrid-precisequantization are developed to boost the hardware efficiency.A 128x8 256-QAM aBsP MIMO detector ASIC in 40 nmCMOS occupies an area of 0.68 mm(2)and reaches a throughputof 790.52 Mbps. Benchmarking with the recent arts, this workachieves 1.08xarea efficiency and 3.34xgate efficiency.