Low-Complexity Joint Active and Passive Beamforming Design for IRS-Assisted MIMO

TL;DR

This paper introduces low-complexity, closed-form joint active and passive beamforming methods for IRS-assisted MIMO systems by leveraging channel Kronecker structures, improving computational efficiency with minimal spectral efficiency loss.

Contribution

It proposes two novel closed-form beamforming algorithms exploiting channel structure, enhancing efficiency and robustness under imperfect channel estimation.

Findings

Kronecker structure exploitation reduces computational complexity.

Tensor-based method outperforms SVD in spectral efficiency under imperfect channels.

Proposed methods achieve near-optimal performance with lower complexity.

Abstract

In this letter, we consider an intelligent reflecting surface (IRS)-assisted multiple input multiple output (MIMO) communication and we optimize the joint active and passive beamforming by exploiting the geometrical structure of the propagation channels. Due to the inherent Kronecker product structure of the channel matrix, the global beamforming optimization problem is split into lower dimensional horizontal and vertical sub-problems. Based on this factorization property, we propose two closed-form methods for passive and active beamforming designs, at the IRS, the base station, and user equipment, respectively. The first solution is a singular value decomposition (SVD)-based algorithm independently applied on the factorized channels, while the second method resorts to a third-order rank-one tensor approximation along each domain. Simulation results show that exploiting the channel Kronecker structures yields a significant improvement in terms of computational complexity at the expense of negligible spectral efficiency (SE) loss. We also show that under imperfect channel estimation, the tensor-based solution shows better SE than the benchmark and proposed SVD-based solutions.