A low complexity soft-input soft-output MIMO detector which combines a Sphere Decoder with a Hopfield network

In this paper, a reduced complexity soft-input soft-output MIMO detector is presented. The detector combines a Sphere Decoder, a Hopfield neural network and an error correction code in an iterative structure (turbo). The simulation results demonstrate that with less computational complexity, the proposed system's performance equals that of a sphere decoder based max-log-map detector in ideal channel conditions. In correlated channel conditions, the system performs within 0.3 dB of the max-log-map detector at a reduced complexity. An adaptive extrinsic information scaling factor is also introduced to improve performance in non-ideal channel conditions.