Sum Secrecy Rate Optimization in RIS-Assisted ISAC Systems: A Manifold-Based Framework

Abstract

We investigate enhancing the sum secrecy rate (SSR) of a reconfigurable intelligent surface (RIS)-assisted integrated sensing and communication (ISAC) system in a multi-cluttered environment with multiple eavesdropping targets. We formulate an SSR maximization problem by jointly optimizing the ISAC base station (BS) beamformer, the RIS phase shifts, and artificial noise (AN) covariance matrix used by the ISAC BS to disrupt the eavesdropping targets. A novel low-complexity manifold-based approach is proposed to solve this problem, incorporating signal-to-clutter-plus-noise ratio (SCNR) and signal-to-interference-plus-noise ratio (SINR) as constraints for sensing and communications, respectively. The proposed alternating optimization (AO) algorithm on Riemannian manifolds achieves high secrecy rates with a reduced complexity when compared to classical optimization approaches.