Fractional Sparse Code Multiple Access and Its Optimization

This letter proposes a fractional sparse code multiple access (FrSCMA) system in which each user directly maps a varying number of error-correction coded symbols onto the resource elements per channel use. The dynamic sparse mapping leads to a fractional spreading factor, often smaller than that in the existing multidimensional constellation (MDC) mapping based sparse code multiple access (SCMA) approach. We develop a two-step optimization method to design the factor graph for maximizing the sum rate. Simulation results demonstrate that FrSCMA achieves higher sum rates than MDC-SCMA without increasing the decoding complexity.