Hybrid Precoder Design with Minimum-Subspace-Distortion Quantization in Multiuser mmWave Communications

Hybrid precoding has been a promising technique in millimeter wave (mmWave) communications, providing a balanced tradeoff between system performance and hardware complexity. Existing hybrid precoding schemes either require full channel state information (CSI) or use codebook-based design, causing large feedback overhead or degraded system performance, respectively. In this paper, we propose a balanced scheme with limited feedback CSI and element-level quantization. Our key idea is to maximize the system sum rate by using an adaptive baseband precoder and an eigenbeam radio frequency (RF) precoder. The adaptive baseband precoder can balance the effective channel gain and the multiuser interference according to the transmitted power. The eigenbeam RF precoder is optimized with only requiring a limited length of feedback vector. Considering the practical constraints on RF precoders, we then propose a quantization algorithm that minimizes the subspace distortion between the optimized eigenbeam RF precoder and the quantized one. The quantization algorithm also works for the case when RF chains are more than users. Extensive simulation results are provided and validate the effectiveness of the proposed hybrid precoding scheme.