Energy-Efficient Multiuser Hybrid Precoding Based on Switch-Controlled Hilbert Network

Hybrid precoding is an extremely attractive technology in mmWave massive multiple-input multiple-output (MIMO) systems. Especially the partially-connected structure has attracted more and more attention due to its low energy consumption. Nonetheless, its energy efficiency (EE) is still limited by a complex phase shifter network consisting of high-resolution phase shifters. In order to further improve EE, this paper proposes two energy-efficient partially-connected hybrid precoding structures, a) fixed-connected and b) adaptive-connected, based on switch-controlled Hilbert network. In both structures, we jointly optimize the switch connection network and hybrid precoder to maximize EE. a) For the fixed-connected structure, we directly develop an effective multiuser phase adaptive selection (MUPAS) algorithm for the discrete hybrid precoding problem when the channel propagation environment is relatively stable. b) For the adaptive-connected structure, we adopt a switch connection state selection (SCSS) algorithm based on maximization channel gain criterion to dynamically adjust the connection between antennas and radio frequency links according to different channel conditions. When the connection state is obtained, the hybrid precoder can also be solved by the MUPAS algorithm. Simulation results demonstrate that our scheme achieves much higher EE than other traditional structures while maintaining an acceptable spectral efficiency.