Enabling High Order SCMA Systems in Downlink Scenarios With a Serial Coding Scheme

Abstract

In the future fifth generation communication systems, sparse code multiple access (SCMA) shows strong competitiveness as a novel non-orthogonal multiple access (NOMA) technique. Attributing to multi-dimensional codebooks, the SCMA can obtain shaping gain and provide a better performance than some other NOMA schemes, such as pattern division multiple access, low-density signature (LDS), and so on. However, under higher user load, each orthogonal resource in high-order SCMA systems is occupied by more users and the constellations of these systems possess smaller Euclidean distances, both of which will degrade the link performance. The iterative algorithm employed in SCMA decoding, namely message passing algorithm, will also bring an exponential growth of computational complexity. In this paper, a scheme with serial codes for system order reduction in downlink scenarios, namely serial SCMA, is proposed to tackle these issues. Based on the idea of hierarchical coding, the processes of encoding and decoding in a high-order system are decomposed into several processes in low-order systems. Specific mapping modules are designed to recover the original user binary bits. Simulation results show that serial SCMA will substantially cut down the detection complexity and enjoy better link performance while maintaining sparser codebooks. Besides, it still inherits the advantages of original SCMA systems such as overloading and average aggregate energy efficiency over OMA and other NOMA schemes.