RIS-Assisted Joint Communication, Sensing, and Multi-Tier Computing Systems

Abstract

This paper investigates the application of Reconfigurable Intelligent Surfaces (RIS) in Joint Communication, Sensing, and Multi-tier Computing (JCSMC). An RIS-assisted JCSMC framework is proposed, wherein a full-duplex multi-antenna Base Station (BS) is employed to sense targets and provide edge computation services to User Equipment (UE). To enhance computational efficiency, a Multi-Tier Computing (MTC) architecture is adopted, enabling joint processing of computing tasks through the deployment of both the BS and the Cloud Servers (CS). Meanwhile, this paper studies the potential advantages of RIS in the proposed framework. It can assist in enhancing the efficiency of resource sharing between sensing and computing functions and then maximize the ability of computing the offload. This study aims to maximize the computation rate by jointly optimizing the BS transmission beamformer, RIS reflection coefficients, and computational resource allocation. The ensuing non-convex optimization problems are addressed using an alternating optimization algorithm based on Block Coordinate Ascent (BCA) for partial offloading mode, which ensures convergence to a local optimum, then extending the proposed joint design algorithms to the scenario with imperfect Self-Interference Cancellation. The effectiveness of the proposed algorithm was confirmed by analyzing and contrasting the simulation results with the benchmark scheme. The simulation results show that, when the BS resources are limited, utilizing MTC architecture can significantly improve the computation rate. In addition, the proposed RIS-assisted JSCMC framework is superior to other benchmark schemes in dealing with resource utilization between different functions, achieving superior computing power while maintaining sensing quality.