Joint Precoding and User Grouping for RIS-Aided mmWave NOMA System

Reconfigurable Intelligent Surface (RIS) and Non-Orthogonal Multiple Access (NOMA) are two promising technologies for future Beyond Fifth-Generation (B5G) and the Sixth Generation (6G). Herein, we investigate the RIS-aided millimeter-wave (mmWave) NOMA system, where user grouping, passive beamforming, hybrid precoding, as well as power allocation are jointly optimized to maximize the system sum rate. We exploit the Riemannian conjugate gradient (RCG) algorithm to optimize the passive beamforming and then propose a dynamic user grouping scheme based on the singular value decomposition (SVD) of the user channel. Furthermore, the analog and digital precoding are designed according to the selected singular value vectors. Finally, the power allocation optimization problem is solved by an iterative optimization algorithm. Compare with the existing NOMA schemes, simulation results show that the proposed scheme can achieve better performance.