Impact of Phase Optimization on the Cascade Channel and System Performance of RIS-Assisted Wireless Networks

In this work, we investigate the impacts of the Reconfigurable Intelligent Surface (RIS) phases or reflection matrix on the effective cascade (i.e., joint transmitter (TX)-to-RIS and RIS-to-receiver (RX)) channel and on the system performance for RIS-aided wireless networks. Using channel gain, signal-to-noise ratio and spectral efficiency as metrics, we demonstrate that our physical model-based RIS phases employing an optimized Fourier Transform (FT)-based beam-forming algorithm outperforms the cosine similarity (CS) phase algorithm. Unlike the CS algorithm where the RIS phases are deduced from the joint incident and reflected channel vectors, the optimized FT-based phases are blind to the cascade channels. We benchmark our results with random RIS phases as well as the mathematical upper bound. The results show that the FT phases outperform the CS and random phases, and approach the mathematical upper bound performance under diverse system parameters and configurations.