A novel non-iterative algorithm for the joint design of transceiver beamforming and surface reflection in an IRS-enhanced MIMO system

Abstract Intelligent reflecting surface (IRS) is an emerging and low-cost revolutionary technology that can be deployed in future communication systems to enhance data transmission performance. An IRS comprises a lot of small, passive, and low-cost elements that can smartly reflect the incident signal from the transmitter to the receiver. In this paper, an IRS-enhanced multiple-input multiple-output (MIMO) downlink network is considered, in which a cell-edge user receives the data signal from a base station (BS). In this network, the direct path between the BS and the user is blocked by an obstacle, and the user receives only the tunable reflected signal from the IRS. To achieve spatial diversity, beamforming is applied to the antennas at the BS and the user. The goal is to jointly design the transceiver beamforming vectors and the IRS reflection coefficients so that the signal-to-noise ratio (SNR) of the user is maximized. Since the IRS elements are passive, the amplitude of the IRS reflection coefficients must be equal to or smaller than one. The constrained SNR optimization problem is non-convex. We propose a three-step procedure to obtain an effective sub-optimal solution for this problem. Accordingly, an innovative non-iterative algorithm is proposed to design the problem parameters. Simulation results show that the IRS-enhanced MIMO downlink system, in which the proposed non-iterative algorithm is used to develop the network parameters, outperforms the conventional network without IRS in terms of bit error rate (BER).