Low-Complexity Expectation Propagation Detection for Uplink MIMO-SCMA Systems

We consider uplink sparse code multiple access (SCMA) systems associated with multiple input multiple output (MIMO), where the transmitters and the receiver are equipped with multiple antennas, for enhanced reliability (diversity gain) or improved data rate (multiplexing gain). For each diversity or multiplexing based MIMO scheme combined with SCMA, we develop low-complexity iterative detection algorithms based on the message passing algorithm (MPA) and the expectation propagation algorithm (EPA). We show that the MIMO-SCMA under EPA enjoys the salient advantage of linear complexity (in comparison to the MPA counterpart with exponential complexity) as well as enhanced error rate performances due to the MIMO transmission. We also show that the performance of EPA depends on the codebook size and the number of antennas.