Low-Complexity Passive Beamforming for IRS-Aided Uplink NOMA

We consider nonorthogonal multiple access (NOMA) assisted by an intelligent reflecting surface (IRS) for the uplink of a cellular network. We aim to optimize the reflection phases of the IRS elements to maximize the sum rate at low complexity for both cases of continuous and discrete phases. The low complexity phase adjustment suggested in this paper results in a negligible loss when compared with the conventional method for the continuous phase case. It is also observed that a simple phase quantization of the solution for the continuous phase works well by using no less than two bits for discrete phase representation.