Hybrid message passing based low complexity receiver for SCMA system over frequency selective channels

As the mobile communications develop rapidly, ever higher spectral efficiency is required. The sparse code multiple access (SCMA) has been recognized as a promising technology to further increase the access efficiency of wireless networks. In this paper, we consider the receiver design problem for SCMA system over frequency selective channels. The conventional minimum mean squared error (MMSE) detection method suffers from huge complexity due to the the multi-user and inter-symbol interferences. To this end, we propose a near optimal low complexity message passing receiver. By approximating the discrete loglikelihood ratio as Gaussian random variable, all messages on factor graph can be obtained as Gaussian distributions. Furthermore, we propose to introduce auxiliary variables to the factor graph and develop a novel hybrid belief propagation (BP) and expectation propagation (EP) receiver. Simulation results show that the proposed hybrid BP-EP method performs close to the MMSE-based receiver with reduced complexity. Also, compared to the orthogonal multiple access scheme, the considered SCMA system with the proposed receiver is able to support 50% more users.