SIC vs. JD: Uplink NOMA techniques for M2M random access

Uplink grant-free non-orthogonal multiple access (NOMA) is promising for machine-to-machine (M2M) communications to meet the requirement of sporadic short packets transmission of massive devices. In this paper, successive interference cancellation (SIC) and joint decoding (JD), as two well-known decoding techniques for NOMA, are analyzed for M2M random access. Due to the sporadic small packet transmission, a block-fading multiple access channel (BF-MAC) is considered, and the frequency-selectivity is also introduced. The achievable individual outage rate regions of SIC and JD, with deterministic active users, are first obtained based on individual outage events, then the analysis is extended to random access and the achievable individual outage rates and throughputs are derived. Numerical results show that SIC can perform as well as JD with relative low user load and transmission powers over frequency-flat BF-MAC. But with higher transmission powers or frequency-selectivity, the outage-limited maximum throughputs of JD are significantly higher than that of SIC.