Model-Driven Deep Learning for Massive MU-MIMO with Finite-Alphabet Precoding

TL;DR

This paper introduces a model-driven deep learning approach for finite-alphabet precoding in massive MU-MIMO systems, effectively addressing hardware constraints and improving performance over traditional methods.

Contribution

It develops a novel deep learning network by unfolding an iterative algorithm specifically for finite-alphabet precoding in massive MU-MIMO systems.

Findings

Significant performance improvements over traditional techniques.

Reduced complexity and increased robustness to channel estimation errors.

Effective in Rayleigh fading channels.

Abstract

Massive multiuser multiple-input multiple-output (MU-MIMO) has been the mainstream technology in fifth-generation wireless systems. To reduce high hardware costs and power consumption in massive MU-MIMO, low-resolution digital-to-analog converters (DAC) for each antenna and radio frequency (RF) chain in downlink transmission is used, which brings challenges for precoding design. To circumvent these obstacles, we develop a model-driven deep learning (DL) network for massive MU-MIMO with finite-alphabet precoding in this article. The architecture of the network is specially designed by unfolding an iterative algorithm. Compared with the traditional state-of-the-art techniques, the proposed DL-based precoder shows significant advantages in performance, complexity, and robustness to channel estimation error under Rayleigh fading channel.