SPEB Minimization for STAR-RIS-Assisted Cooperative ISAC With Active and Passive Sensing

We consider the joint design of a simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS)-assisted cooperative integrated sensing and communication (ISAC) system to realize concurrent high-precision sensing and reliable communication for future sixth-generation networks. Specifically, we consider a STAR-RIS-assisted cooperative ISAC system with the sensing-at-STAR-RIS architecture. The STAR-RIS not only assists severely blocked non-line-of-sight communication users but also allows passive sensing of the target. Thus, in addition to monostatic active sensing of the target at the base station, the proposed architecture utilizes the echoes received at the STAR-RIS to significantly enhance the sensing performance. To characterize the sensing performance, we derive the fundamental squared position error bound (SPEB) as a key localization metric for the considered STAR-RIS-assisted cooperative ISAC system. We investigate the optimal design of the transmit beamforming matrix, sensing covariance matrix, and STAR-RIS coefficients to minimize the SPEB of the target while guaranteeing communication quality-of-service requirements. In order to analytically solve the highly-coupled non-convex optimization problem, we adopt an iterative algorithm based on the double-loop penalty dual decomposition framework and block coordinate descent method. Numerical results corroborate the effectiveness of the proposed architecture by demonstrating that it outperforms the conventional RIS and random STAR-RIS phase-shift baselines by achieving 24.62% and 40.08% reductions in the SPEB, respectively, for the considered parameter values.