Low-Complexity Near-Optimal Demodulation for Spatial Modulation Based on M-Algorithm

Abstract

Spatial modulation (SM), renowned for its high energy efficiency, stands as a pivotal massive MIMO transmission technique. Leveraging the parallel processing capability and superior error rate performance, the M-algorithm to maximum likelihood (MML) algorithm has emerged as a competitive demodulation method for SM systems. In MML, judicious configuration of layer widths can effectively diminish complexity. Nevertheless, existing literature, to the best of our knowledge, relies on empirical settings for layer widths, lacking a formal optimization strategy. In this letter, we propose a layer width optimization algorithm based on mathematical analysis and Monte Carlo simulations. Numerical results have demonstrated that the MML algorithm coupled with our proposed layer width configuration method yields the symbol vector error rate (SVER) comparable to the state-of-the-art optimal demodulation, while achieving a noteworthy complexity reduction.