A Low-Complexity AEPDF-assisted Precoding Scheme for Massive MIMO Systems with Transmit Antenna Correlation

Abstract

Multi-user massive multiple-input multiple-output (MU M-MIMO) system employs array of hundred (or thousand) antennas and precoding/equalization capability at the base station (BS) has been adopted in recent 5G new radio (5G NR) standards. By taking advantage of the multiplexing, diversity and antenna gain inherited from this versatile system configuration, unprecedented enhancement in capacity, spectral efficiency and signal integrity can be realized. While many MU M-MIMO system related issues such as pilot contamination, channel estimation and precoding/detection algorithms have been extensively examined by research communities, the investigation of a low-complexity, effective precoding scheme aims to address performance degradation due to spatial correlation (i.e., packing vast number of antennas in a limited physical space) remained scarce. In this paper, an iterative precoding scheme based on Chebyshev acceleration is proposed for one-sided Kronecker model (i.e., Rayleigh flat fading channel with transmit (Tx) antenna correlation), often encountered in uniform linear array (ULA) configuration. From the asymptotic eigenvalue probability density function (AEPDF) constructed for the non-commutative Kronecker channel, near-optimal recurrence coefficients, initial and final estimates of the proposed Chebyshev iterative method can be obtained free from expensive eigenvalue computation. Simulation shows that the proposed algorithm achieves outstanding BER in all correlation scenarios (i.e., from low to high) while exhibiting much lower complexity compared to other state-of-the-art precoding schemes.