IRS-Aided Secure Sensing for Surveillance Area Coverage: Framework and Algorithm Design

This paper proposes a novel IRS-aided framework for secure sensing, which aims to minimize the worst-case Cramér-Rao bound (WC-CRB) within an entire surveillance area by optimizing the IRS reflecting beamforming, enabling reliable and secure localization of arbitrary and unknown targets. Specifically, we first establish a general IRS-aided localization coverage model and derive the closed-form expression for the CRB of an arbitrary point, which reveals the relationship between the localization error bound and the Fisher information of the angle of arrival (AOA), angle of departure (AOD) and delay. To solve this challenging min-max optimization problem, we design efficient algorithms for different area types. For sector area, we first represent the Fisher information as trigonometric polynomials, then construct the WC-CRB coverage constraint as a non-negativity problem of these polynomials, and finally approximate it as an efficiently solvable semidefinite program (SDP). For the more challenging case of arbitrarily shaped area, we propose a two-tiered solution comprising a low-complexity heuristic algorithm based on geometric approximation and a high-performance detailed design that accurately solves the problem by decomposing the irregular boundary into multiple continuous segments. Numerical simulations validate the superiority of the proposed framework, demonstrating that our designs significantly outperform various benchmark schemes in terms of robustness and performance uniformity. The results show that the framework not only effectively reduces the WC-CRB but also achieves a highly uniform performance coverage across the entire area, providing a reliable and efficient solution for practical localization security applications.