On Intelligent Reflecting Surface-Assisted NOMA Uplinks With Imperfect SIC

As the interplay between intelligent reflecting surface (IRS) and non-orthogonal multiple access (NOMA) can boost both the spectrum- and the power-efficiency, this letter investigates an IRS-NOMA system, taking into account the multi-group detection and imperfect successive interference cancellation (SIC). Specifically, a group-level SIC (GSIC) is conceived to remove the decoded groups' signals, whilst the joint design of power control and equalizers is leveraged to mitigate the aggregated interference caused by NOMA, imperfect SIC, and concurrent transmissions per group. According to the newly formulated multi-group NOMA framework, the transmit power and the phase shifts are optimized alternately to reduce the total energy consumption together with the minimum mean square error equalizer. Thereof, the power control and the phase shift subproblems are respectively addressed by a group-level parallel iteration method and an element-specific sequential rotation method. Compared with various existing methods, the superiority of the proposed IRS-NOMA scheme is demonstrated through simulations.