Multistatic ISAC for Improved Sensing Under Jamming Attacks: A Gradient-Based Meta-Learning Approach

Abstract

Motivated by the growing vulnerability of integrated sensing and communication (ISAC) systems to jamming attacks, this work explores multistatic architectures as a promising approach to improve robustness in adversarial environments, specifically in the sensing functionality. We consider a multistatic ISAC framework in which a transmit base station (BS) simultaneously performs communication and sensing, while a set of spatially distributed receiver BSs cooperatively process the reflected echoes from the target through a fusion center under sensing-directed jamming. To fully leverage the spatial diversity in this architecture, we formulate a joint optimization problem that maximizes the fused echo signal-to-interference-plus-noise ratio (SINR), subject to the BS power budget and per-user quality-of-service (QoS) constraints by determining the precoding vector at the transmit BS and the receive beamforming vector at the receiving BSs. The resulting problem is non-convex due to the complex coupling between variables, making it challenging to solve using traditional optimization methods. To address this, we propose a gradient-based meta-learning (GML) approach that learns efficient update rules for rapid convergence. Simulation results validate the effectiveness of the proposed approach, achieving up to 95.68 percent of the optimal performance while substantially mitigating the impact of jamming. The findings highlight that multistatic reception with meta-learning offers a scalable solution to security threats in ISAC systems.