A Low-Complexity Large-MIMO Approximate Message Passing Detector

Message passing algorithms (MPAs) are iterative algorithms widely used in multiple-input multiple-output (MIMO) detection, benefiting from their good trade-off between performance and complexity. In this paper, a low-complexity large-MIMO approximate message passing (LC-LAMA) detector with partial message updates is proposed. Also, we propose an evolutionary algorithm (EA) based method to optimize the number of updated messages in LC-LAMA for multiple scenarios. Numerical results exhibit that the proposed LC-LAMA detector presents comparable performance to that of state-of-the-art (SOA) detectors. The superior performance of LC-LAMA in mismatch scenarios illustrates its robustness. An analysis of complexity reveals the proposed LC-LAMA provides at most 72.7% complexity reduction as compared to other SOA detectors.