Joint Transceiver and Passive Beamforming Optimization for RIS-Assisted MIMO Systems

This paper investigates the transceiver and beamforming design for a reconfigurable intelligent surface (RIS)-assisted multiple-antenna communication system. We aim to minimize the weighted mean-squared error of the received data by jointly optimizing the precoder, the passive RIS beamforming, and the receiving filter of the system, subject to a transmit power constraint and unit-modulus constraints imposed on the RIS reflection coefficients. In order to tackle the nonconvexity of this design problem, we propose a computationally-efficient algorithm based on alternating optimization, where the solutions to the precoder, the RIS beamforming, and the receiving filter are derived in semi-closed forms and closed-form, respectively. The convergence of the proposed algorithm is theoretically proved. Numerical results validate the performance superiority of the proposed algorithm over various existing benchmark schemes.