Transceiver and Phase Optimization for RIS-assisted Hardware-impaired MIMO Cellular Systems

We analyze the potential of using reconfigurable intelligent surfaces (RISs) as an alternative to classical active beamforming to encounter the effects of non-ideal hardware components in the transmitter chain of a base station (BS). In particular, we consider a RIS-assisted multi-user multiple-input multiple-output (MIMO) downlink system, where the received signal-to-noise ratio is not only affected by inter-user interference but also hardware impairments due to amplifier nonlinearity, phase noise, and I/Q channel imbalance. Accordingly, we propose a hybrid beamforming strategy by minimizing the sum of mean-squared errors of all estimated symbols subject to the total transmit power and element-wise phase-shift constraint to obtain the optimal beamforming vectors at the BS and phase shift matrices at the RIS. We show that this problem can be cast as a second-order cone programming problem, and a joint design of the active beamforming vectors and phase-shift matrices can be obtained through an iterative algorithm. Through numerical results, we demonstrate that by using RISs in a cellular communication system, it is possible to effectively mitigate the impact of hardware impairments and achieve better sum-rate and power-efficiency.