Network-coded uplink clustered NOMA relay networks: Models and performance comparisons

Abstract

In this paper, network sum-rate maximization (NSR-MAX) for network-coded (NC) uplink clustered non-orthogonal multiple-access (NOMA) relay networks is considered. In particular, the goal is to maximize network sum-rate via optimal power allocation, where the user clusters communicate with the base-station over an amplify-and-forward relay, subject to quality-of-service (QoS) constraints. Two NC-NOMA transmission models – namely NC-NOMA-I and NC-NOMA-II – are proposed, where in the first model, network-coding is applied at the relay only, while the second model applies network-coding for user clusters’ transmissions and at the relay to further minimize the transmission delay. The formulated NSR-MAX problem happens to be non-convex, and hence, is computationally-intensive. Thus, a low-complexity iterative two-layer algorithm (ITLA) is devised, which decouples the formulated problem and efficiently solves its over two-layers. Specifically, the inner-layer solves the NSR-MAX problem to obtain the optimal relay transmit power, while the outer-layer determines the optimal users’ transmit powers. Numerical results are presented, which illustrate that the proposed ITLA yields near-optimal solutions for all transmission models, in comparison to the formulated NSR-MAX problem (solved via global optimization package) as well as outperforming its OMA-based counterparts. Not only that, but the NC-NOMA-II is shown to be superior to the other NOMA and OMA schemes, while demonstrating resilience to timing offsets as well as successive interference cancellation (SIC) and channel state information (CSI) errors.