LAUSR

Reconfigurable intelligent surface (RIS) is a cost-effective solution to support millimeter wave (mmWave) communications suffering from severe blockage. In this letter, we aim to maximize the spectral efficiency (SE) of the RIS-assisted mmWave multiple-input multiple-output (MIMO) system by jointly optimizing the hybrid transceiver and passive RIS under the nonconvex unit-modulus constraints. It is revealed that the considered system is equivalent to a conventional mmWave system with a Kronecker-structured hybrid transmit antenna array. Using this fact, we firstly propose to jointly design the RIS and hybrid transceiver based on the idea of unitary matching. Motivated by the asymptotic orthogonality of array steering vectors for very large systems, a low-complexity two-stage scheme is also developed by simply choosing the transmit/receive array steering vectors associated with the first several largest path gains as the analog precoder/combiner. Numerical results further demonstrate the superior performance of the proposed schemes in terms of the achievable SEs and average run time.