RIS-Assisted Broad Coverage for mmWave Massive MIMO System

Reconfigurable intelligent surfaces (RISs) make wireless environments controllable to ensure reliable communications in millimeter-wave (mmWave) multi-antenna systems. With the assistance of an RIS, we rebuild a coverage for users when the direct broadcast channel is blocked in a mmWave massive multiple-input multiple-output (MIMO) system. To obtain a broad coverage with no channel state information of the users, we generate a flat-top radiation pattern by jointly optimizing the precoding at the base station and the phase shifts of the RIS. Since the phase shifts of RIS are constrained by a constant modulus, we regard this non-convex constraint as a Riemannian manifold to facilitate the optimization. Then, we propose a manifold-based algorithm to efficiently search the optimal RIS phase shifts over the Riemannian manifold by using a conjugate gradient method. Moreover, a broad beam in fact disperses the transmit power over the space such that the power gain decreases as the coverage area increases. Therefore, we introduce beam sweeping to improve the performance of broad coverage. Numerical results show that the proposed algorithm can provide an adjustable broad coverage with small power fluctuations, leading to a reliable mmWave communication of broadcast channel.