Low Complexity Passive Beamforming Algorithms for Intelligent Reflecting Surfaces with Discrete Phase-Shifts over OFDM Systems

Abstract

In recent literature, intelligent reflecting surfaces (IRS) based wireless system design has been a significant point of excitement among the wireless community. We consider the problem of configuring the IRS elements efficiently and effectively for a practical IRS with discrete phase shifts and coupled elements deployed in an orthogonal frequency-division multiplexing (OFDM) based environment. We propose two near-optimal low complexity extremely scalable heuristic algorithms to design phase shifts at IRS in an OFDM system when the number of bits used to configure the IRS is limited, and the reflected channels via the IRS are spatially correlated. We benchmark the sum data rate performance of our algorithms against the theoretical upper bound and the time performance against the existing successive convex approximation. Results indicate that 4 bits are sufficient to obtain theoretically optimum sum data rates and that our proposed algorithms obtain a good trade-off between complexity and performance.