Robust Secure ISAC: How RSMA and Active RIS Manage Eavesdropper's Spatial Uncertainty
Authors:
A. Abdelaziz Salem,
Saeed Abdallah,
Mohamed Saad,
Khawla Alnajjar,
Mahmoud A. Albreem
Abstract:
Incorporating rate splitting multiple access (RSMA) into integrated sensing and communication (ISAC) presents a significant security challenge, particularly in scenarios where the location of a potential eavesdropper (Eve) is unidentified. Splitting users' messages into common and private streams exposes them to eavesdropping, with the common stream dedicated for sensing and accessible to multiple…
▽ More
Incorporating rate splitting multiple access (RSMA) into integrated sensing and communication (ISAC) presents a significant security challenge, particularly in scenarios where the location of a potential eavesdropper (Eve) is unidentified. Splitting users' messages into common and private streams exposes them to eavesdropping, with the common stream dedicated for sensing and accessible to multiple users. In response to this challenge, this paper proposes a novel approach that leverages active reconfigurable intelligent surface (RIS) aided beamforming and artificial noise (AN) to enhance the security of RSMA-enabled ISAC. Specifically, we first derive the ergodic private secrecy rate (EPSR) based on mathematical approximation of the average Eve channel gain. An optimization problem is then formulated to maximize the minimum EPSR, while satisfying the minimum required thresholds on ergodic common secrecy rate, radar sensing and RIS power budget. To address this non-convex problem, a novel optimization strategy is developed, whereby we alternatively optimize the transmit beamforming matrix for the common and private streams, rate splitting, AN, RIS reflection coefficient matrix, and radar receive beamformer. Successive convex approximation (SCA) and Majorization-Minimization (MM) are employed to convexify the beamforming and RIS sub-problems. Simulations are conducted to showcase the effectiveness of the proposed framework against established benchmarks.
△ Less
Submitted 21 July, 2024;
originally announced July 2024.
