Joint Beamforming Design for Semi-IRS-Aided Secure ISAC with Assistance of Sensing

In this paper, we investigate the sensing-assisted secure transmission for an integrated sensing and commu-nication (ISAC) system, where the semi-intelligent reflecting surface (IRS) is deployed to facilitate the secure communication and sensing. First, we adopt the multiple signal classification algorithm to estimate the target's direction with Cramér-Rao bound (CRB) serving as the performance metric. After obtaining the information about the target, which may be a potential eavesdropper, the expression of secrecy rate related to the eavesdropper's estimation accuracy can be formulated. Then, we formulate a weighted optimization problem to maximize the secrecy rate and minimize the CRB by jointly designing the active beamforming, artificial noise and phase shifts of IRS. In particular, by considering the estimation errors, we introduce a series of beampattern constraints about the main beam in order to cover all possible angles of target, ensuring that the security is guaranteed in all directions. Accordingly, the sensing function can improve the secrecy performance through the increasing estimation accuracy in each iteration. Simulation results reveal effectiveness of the proposed scheme and the utility of sensing as a facilitator for physical layer security.