Secure Beamforming for NOMA-ISAC With Multicast and Unicast Communications

This letter investigates the issue of secure beamforming for non-orthogonal multiple access (NOMA)-assisted integrated sensing and communication (ISAC), where some targets acquiring group-oriented information from the base station (BS) of the ISAC system are cooperative. To secure the privacy of served users from malicious targets, artificial noise (AN) carrying no information is often transmitted in conventional security schemes. In this letter, we propose to employ multicast signals to convey group-oriented information to cooperative targets, disrupt malicious targets, and simultaneously sense all targets. Meanwhile, unicast signals are transmitted for privacy information of the served users. We formulate a robust secure beamforming problem that deals with imperfect channel state information (CSI) and the uncertainty regarding the angles of malicious targets. To solve this problem, we propose a penalty-based iterative algorithm by transforming the non-convex optimization problem into a convex one with semidefinite relaxation (SDR) and examining the relationship between the trace, the largest eigenvalue and the rank of a positive semidefinite matrix. Numerical results are provided to demonstrate that, with lower computational complexity, the proposed scheme delivers superior performance, compared to state-of-the-art schemes.