Coordinated Hybrid Precoding for Energy-Efficient Millimeter Wave Systems

Hybrid precoding/combining architectures, which use fewer radio frequency (RF) chains than the number of antennas, have the potential to reduce energy consumption and provide high data rate for millimeter wave systems with large antenna arrays. In this paper, by allowing joint transmissions from multiple base stations (BSs), we jointly optimize over the hybrid precoders and the BSs sleep/active modes such that the total power consumption is minimized with per-user quality-of-service constraints. We present a decoupled hybrid beamforming scheme where the analog precoders are designed based on the equal gain transmission method and the digital precoders are found by solving a convex semidefinite problem. We define realistic models for the hardware power consumption and the channel's blocking probabilities to compare the performance of our scheme with the fully-digital precoding scheme and to assess the value of BS cooperation. The simulation results show that our proposed method results in almost the same RF transmit power as the fully-digital precoding scheme while saving considerable hardware power due to the reduced number of RF chains and digital-to-analog converters. Moreover, the network densification is shown to further reduce the total power consumption through the optimal cooperative BS active/sleep mode selection.