Spectral Efficiency vs Complexity in Downlink Algorithms for Reconfigurable Intelligent Surfaces

TL;DR

This paper explores the tradeoff between spectral efficiency and computational complexity in downlink algorithms for reconfigurable intelligent surfaces, proposing gradient ascent codesign methods and comparing their performance.

Contribution

It introduces new gradient ascent codesign algorithms for RIS-assisted downlink communication and compares their efficiency and complexity.

Findings

Two gradient ascent algorithms achieve optimal spectral efficiency-complexity tradeoffs

Simulation results demonstrate the effectiveness of proposed algorithms in practical scenarios

Comparison of seven algorithms highlights the best tradeoffs in performance and complexity

Abstract

Reconfigurable Intelligent Surfaces (RIS) are an emerging technology that can be used to reconfigure the propagation environment to improve cellular communication link rates. RIS, which are thin metasurfaces composed of discrete elements, passively manipulate incident electromagnetic waves through controlled reflective phase tuning. In this paper, we investigate co-design of the multiantenna basestation beamforming vector and multielement RIS phase shifts. The downlink narrowband transmission uses sub-6 GHz frequency bands, and the user equipment has a single antenna. Subject to the non-convex constraints due to the RIS phase shifts, we maximize the spectral efficiency or equivalent channel power as a proxy. Our contributions in improving RIS-aided links include (1) design of gradient ascent codesign algorithms, and (2) comparison of seven codesign algorithms in spectral efficiency vs. computational complexity. In simulation, the best spectral efficiency vs. computational complexity tradeoffs are shown by two of our proposed gradient ascent algorithms.