Performance Analysis and Optimization of Partitioned-IRS-Assisted NOMA Network

In this letter, we analyze the performance of a downlink non-orthogonal multiple access (NOMA) network that uses co-phasing at an intelligent reflecting surface (IRS) to communicate to a near-user (NU) and a far-user (FU) concurrently. Both users combine the direct and reflected signals. Considering cross-channel interference (CI) and a finite number of phase levels at the IRS, we derive closed-form expressions for the outage probability. Closed form expressions are also presented for throughput, energy efficiency (EE) and optimal power allocation. A low complexity procedure is suggested to determine the optimum partitioning so as to maximize the FU throughput while ensuring a desired throughput at the NU. The proposed IRS-NOMA (IN) system outperforms conventional NOMA (CN) and orthogonal multiple access (OMA) systems in terms of throughput and EE, and permits a broader range of NOMA power allocations. Simulations validate the derived expressions.