Low Complexity Joint MPA Detection for Downlink MIMO-SCMA

Abstract

The newly proposed sparse-code multiple-access (SCMA) has been shown to approach the single-user performance even the overload reaches 150% when message passing algorithm (MPA) detection is applied. Obviously, combined with multiple-input multiple-output (MIMO) techniques, the spectrum efficiency of SCMA system can be further improved. While maximum likelihood (ML) detection is optimal for MIMO-SCMA, the detection complexity would grow exponentially when the number of both the antennas and users increase. In this paper, we propose a novel low complexity joint MPA detection scheme for downlink MIMO-SCMA to utilize the sparsity of the channel and codeword. The analysis and simulation results show that compared with ML detection, the proposed joint MPA can decrease the complexity significantly while keep the BER performance unchanged.