Adaptive Hybrid Precoding for Multiuser Massive MIMO

Abstract

Hybrid precoding (HP) is widely utilized in millimeter wave-based massive MIMO systems with significantly reduced radio frequency (RF) chains, but it requires a large number of analog phase shifters (APSs) and RF adders to realize the connection between RF chains and antenna elements. In this letter, an adaptive hybrid precoding (AHP) is proposed to approach the performance of the conventional HP with reduced complexity. Different from the conventional HP where each antenna is connected to all RF chains through APSs and RF adders, the proposed AHP connects each antenna with only one RF chain through an adaptive connection network. This adaptive connection network and the phases of APSs are jointly designed, which is formulated as an optimization problem to maximize the users' average downlink achievable rate. Moreover, a multiuser adaptive analog precoding (MU-AAP) algorithm is proposed to provide a near-optimal solution to this joint-design problem. Simulation results verify the performance gain of the proposed AHP in typical multiuser massive MIMO scenarios.