Sensing-Informed Robust Beamforming for Secure ISAC Against Mobile Passive Eavesdroppers
TL;DR
Results demonstrate that explicitly coupling sensing-derived localization accuracy with communication-side CSI uncertainty provides an effective sensing-informed design paradigm for robust secure ISAC under the adopted geometry-based propagation and sensing assumptions.
Abstract
Integrated sensing and communication (ISAC) enables wireless systems to simultaneously perform environmental sensing and data transmission, creating new opportunities for sensing-assisted physical-layer security. However, existing secure ISAC designs generally assume either perfect knowledge of the eavesdropper’s channel or predefined channel state information (CSI) uncertainty models, making them unsuitable for passive mobile eavesdroppers whose CSI must be reconstructed indirectly from sensing observations. This paper proposes a sensing-informed secure ISAC framework that establishes an explicit sensing-to-security coupling by constructing the communication-side CSI uncertainty directly from sensing-derived localization accuracy. A Cramér–Rao bound (CRB)-guided adaptive sensing policy first schedules sensing updates according to the predicted localization uncertainty of the mobile eavesdropper. A geometry-based channel model is then combined with first-order uncertainty propagation to derive a confidence-qualified sensing-informed CSI uncertainty model that analytically maps localization uncertainty into communication-channel uncertainty. Building upon this uncertainty representation, the robust secrecy beamforming problem is successively reformulated through covariance reformulation, deterministic secrecy reformulation, the S-procedure, and semidefinite relaxation, resulting in a tractable semidefinite programming (SDP) formulation. Numerical results demonstrate that the proposed adaptive sensing policy substantially reduces sensing overhead while maintaining localization accuracy close to continuous sensing. Furthermore, the proposed sensing-informed robust beamforming framework consistently achieves higher secrecy rates, lower secrecy outage probabilities, and more efficient transmit-power utilization than representative benchmark schemes under imperfect CSI and varying target mobility. These results demonstrate that explicitly coupling sensing-derived localization accuracy with communication-side CSI uncertainty provides an effective sensing-informed design paradigm for robust secure ISAC under the adopted geometry-based propagation and sensing assumptions.