Low-Complexity Hybrid Precoding with Adaptive Connection Network for mm-Wave Massive MIMO System

Abstract

Hybrid precoding (HP) plays a significant role in mm-Wave communication, which greatly reduces the number of radio frequency (RF) chains compared to full digital baseband precoding (FD-BP). Fully-connected HP can approach FD-BP in the performance closely, but it applies a large number of analog phase shifters (APSs), leading to a large power consumption. In order to overcome this difficulty, we propose a novel HP structure based on low power consumption inverters and adaptive connection network This IACN-based HP structure can be realized by digital chip with low computation complexity. Then, an IACN-HP algorithm was proposed on the basis of IACN-based structure, where the analog precoding in the IACN-HP scheme is matched with the subchannels of the different users. To implement the analog precoding, we group the antennas for each RF chain according to channel gain and determine the phase of each inverter corresponding to each antenna jointly. After that, the digital precoding is designed with zero-forcing criterion. The simulation and analysis results show that IACN-HP scheme has higher sum-rate than conventional sub-connected HP schemes. Besides, it also enjoys low power consumption and low computational complexity.