Joint Quantized Constant Envelope Precoding and Antenna Selection for Massive MU-MIMO Downlink Systems Using Higher-Order QAM

This paper introduces the antenna selection (AS) mechanism into the quantized constant envelope (CE) precoding in massive multiuser multiple-input multiple-output systems to simultaneously reduce multiuser interference and hardware power consumption. However, this joint design problem requires optimizing over the precoding vector, the precoding factor, and the selected antenna subset. Unfortunately, these design parameters are connected. To address this problem, we propose an iterative algorithm developed from the alternating minimization (AltMin) framework to decouple the considered problem into two subproblems: the precoding factor design problem and the joint AS and the precoding vector design problem. The former can be easily solved approximately, whereas the latter incurs combinatorial complexity. We address this issue by using a simple yet effective coordinate update algorithm (CUA). We also develop a reduced-complexity version of CUA without performance loss. Moreover, the proposed CUA-based AltMin algorithm can be applied to the quantized CE precoding design problems without modification. The simulation results show that the proposed CUA-based AltMin algorithm considerably outperforms the existing algorithms for the quantized CE precoding design problems. In addition, when the AS mechanism is included in the quantized CE precoding design, CUA-AltMin with AS achieves the same or higher performance than CUA-AltMin without AS while providing further power savings.