MA-Enabled ISAC Systems for Robust Sensing under Jamming Attacks

Abstract

This thesis explores how integrated sensing and communication (ISAC) systems can be designed to remain reliable and efficient in adversarial environments. ISAC represents a growing field in wireless networks that merges communication and radar sensing within a shared infrastructure allowing systems to transmit information while observing their surroundings at the same time. Although significant progress has been made in improving ISAC performance, relatively few studies have focused on securing its sensing function. To address this gap, this thesis proposes an ISAC framework that enhances sens- ing performance in a multistatic ISAC setting, particularly under jamming attacks. A multi-antenna base station (BS) simultaneously serves communication users and probes a target in the presence of multiple jammers, while multiple spatially dis- tributed receiver BSs collect the reflected echoes and forward them to a central fusion unit for joint processing. The system design aims to maximize a fused Echo Spectral Efficiency (ESE) metric, subject to transmit power constraints and user quality-of-service (QoS) requirements. The proposed framework is first developed using fixed antennas and then extended to incorporate movable antennas (MAs) at both the transmitters and receivers. This additional degree of freedom enhances the system’s ability to mitigate jamming and improve sensing. To efficiently solve the resulting high-dimensional non-convex optimization prob- lem, a gradient-based meta-learning (GML) approach is proposed, enabling rapid adaptation to varying conditions. By jointly optimizing beamforming and antenna positions, the approach improves resilience to jamming, enhances sensing gains, and makes more efficient use of spatial resources compared to fixed-antenna designs. Furthermore, it achieves near-optimal sensing performance, reaching up to 93.7% of the performance obtained by the fmincon optimizer, while significantly reduc- ing computational complexity. This demonstrates the effectiveness of our proposed solution as a scalable approach for practical ISAC deployments.