Active and passive beamforming in RIS-assisted cell-free massive MIMO systems: an edge computing perspective

Abstract

In the rapidly advancing field of edge computing, improving the end-to-end transmission rate is crucial to accommodating the needs of latency-sensitive applications. To address this, this article introduces Reconfigurable Intelligent Surfaces (RIS) to examine the challenge of maximizing the minimum attainable rate among users in a cell-free massive MIMO system from an edge computing perspective. In this article, a framework is proposed to improve the end-to-end user transmission rate by alternately optimizing the precoding matrix of Access Points (APs) and the phase shift matrix of the RIS. For the optimization of the APs’ precoding matrix, this framework utilizes a Second Order Cone Programming (SOCP) method. In order to optimize the continuous phase shifts at the RIS, this framework uses a Semidefinite Relaxation (SDR) technique. For the optimization of the discrete phase shifts at the RIS, a projection-based method is proposed in this framework. By integrating these two forms of beamforming, the proposed framework significantly improves the end-to-end transmission rate, meeting the critical requirements of latency-sensitive applications in edge computing scenarios.