Dual-Polarized IRSs in Uplink MIMO-NOMA Networks: An Interference Mitigation Approach

In this work, intelligent reflecting surfaces (IRSs) are optimized to manipulate signal polarization and improve the uplink performance of a dual-polarized multiple-input multiple-output (MIMO) non-orthogonal multiple access (NOMA) network. By multiplexing subsets of users in the polarization domain, we propose a strategy for reducing the interference load observed in the successive interference cancelation (SIC) process. To this end, dual-polarized IRSs are programmed to mitigate interference impinging at the base station (BS) in unsigned polarizations, in which the optimal set of reflecting coefficients are obtained via conditional gradient method. We also develop an adaptive power allocation strategy to guarantee rate fairness within each subset, in which the optimal power coefficients are obtained via a low-complexity alternate approach. Our results show that all users can reach high data rates with the proposed scheme, substantially outperforming conventional systems.