Secure Transmission in ISAC Systems Aided by Active STAR‐RIS

ABSTRACT Integrated sensing and communication (ISAC) is a pivotal technology for sixth‐generation (6G) networks. Intelligent reflecting surfaces (IRS), particularly the simultaneous transmitting and reflecting IRS (STAR‐RIS), enable dynamic channel control to enhance ISAC performance. However, conventional passive STAR‐RIS is constrained by limited signal gain. While the emerging active STAR‐RIS addresses this limitation via signal amplification, it introduces heightened power consumption and security risks, especially when sensing targets act as potential eavesdroppers. This paper investigates the secure transmission problem in an active STAR‐RIS‐aided ISAC downlink system. Our objective is to maximize the sum secrecy rate by jointly optimizing the base station beamforming vectors, the sensing signal covariance matrix and the coefficients of the active STAR‐RIS for both reflection and transmission, while satisfying practical constraints on power and sensing performance. To solve this non‐convex problem, we propose an efficient two‐layer alternating optimization algorithm that decomposes it into tractable subproblems. These subproblems are solved using semidefinite relaxation and a novel eigenvalue‐penalty‐based method. Numerical simulations demonstrate that the proposed active STAR‐RIS scheme significantly outperforms baseline architectures (passive STAR‐RIS, active RIS and passive RIS) in achieving a superior balance between communication security and sensing capability.