Reconfigurable Intelligent Surface assisted Multi-user Communications: How Many Reflective Elements Do We Need?

TL;DR

This paper investigates the relationship between the number of reflective elements in reconfigurable intelligent surfaces and the system sum-rate in multi-user communications, providing guidelines on the minimum elements needed for acceptable performance.

Contribution

It derives the asymptotic capacity with zero-forcing precoding and determines the number of RIS elements required to achieve a desired sum-rate ratio.

Findings

Asymptotic capacity derived for RIS-assisted systems.

Number of elements needed for target sum-rate ratio identified.

Numerical results confirm theoretical analysis.

Abstract

Reconfigurable intelligent surfaces (RISs) consisting of multiple reflective elements are a promising technique to enhance communication quality as they can create favorable propagation conditions. In this letter, we characterize the fundamental relations between the number of reflective elements and the system sum-rate in RIS-assisted multi-user communications. It is known from previous works that the received signal-to-noise ratio~(SNR) can linearly increase with the squared number of RIS reflective elements, but how many elements are sufficient to provide an acceptable system sum-rate still remains an open problem. To this end, we derive the asymptotic capacity with zero-forcing (ZF) precoding, and then discuss how many reflective elements are required so that the ratio of the system sum-rate to the capacity can exceed a predefined threshold. Numerical results verify our analysis.