A Nonbinary LDPC-Coded SCMA System with Optimized Codebook Design

Sparse code multiple access (SCMA) is a promising technique to meet the requirements for massive capacity in 5G communications. In this paper, a nonbinary LDPC-coded SCMA system with a turbo receiver is proposed. We also provide a heuristic approach based on Cartesian product and Euclidean distance to construct the SCMA codebook, which greatly reduces the detection complexity and can be easily extended to large-size constellations. Furthermore, the codebook design for large SCMA systems is presented, and the sparse structure is optimized via the construction of matrices with optimized cycle distribution. Simulation results show that the nonbinary LDPC-coded SCMA system is much more compatible with high-order modulation than its binary counterpart, and the proposed design outperforms the original SCMA codebook design in both the error probability and complexity even in large loading situations.