Joint Active and Passive Beamforming Design for IRS-Assisted Multi-User MIMO Systems: A VAMP-Based Approach

TL;DR

This paper proposes a novel VAMP-based joint active and passive beamforming optimization method for IRS-assisted multi-user MIMO systems, enhancing spectral efficiency and throughput.

Contribution

It introduces a VAMP-based iterative algorithm for phase shifter optimization and applies it within a BCD framework for joint beamforming design.

Findings

Achieves higher throughput than existing methods.

Validates robustness under different IRS phase shift models.

Demonstrates improved spectral efficiency through numerical simulations.

Abstract

This paper tackles the problem of joint active and passive beamforming optimization for an intelligent reflective surface (IRS)-assisted multi-user downlink multiple-input multiple-output (MIMO) communication system. We aim to maximize spectral efficiency of the users by minimizing the mean square error (MSE) of the received symbol. For this, a joint optimization problem is formulated under the minimum mean square error (MMSE) criterion. First, block coordinate descent (BCD) is used to decouple the joint optimization into two sub-optimization problems to separately find the optimal active precoder at the base station (BS) and the optimal matrix of phase shifters for the IRS. While the MMSE active precoder is obtained in a closed form, the optimal phase shifters are found iteratively using a modified version (also introduced in this paper) of the vector approximate message passing (VAMP) algorithm. We solve the joint optimization problem for two different models for IRS phase shifts. First, we determine the optimal phase matrix under a unimodular constraint on the reflection coefficients, and then under the constraint when the IRS reflection coefficients are modeled by a reactive load, thereby validating the robustness of the proposed solution. Numerical results are presented to illustrate the performance of the proposed method using multiple channel configurations. The results validate the superiority of the proposed solution as it achieves higher throughput compared to state-of-the-art techniques.