Secure UAV-Enabled ISAC Against Active Mobile Eavesdroppers: A Stackelberg Game Approach

Abstract

This paper studies secure UAV-enabled integrated sensing and communication (ISAC) against an active mobile aerial eavesdropper. We consider a three-dimensional adversarial scenario where an information UAV (I-UAV) simultaneously serves a ground user and senses a target, while an eavesdropping/jamming UAV (E-UAV) adapts its trajectory and jamming power. The defender-attacker interaction is formulated as a Stackelberg game, in which the I-UAV jointly optimizes confidential beamforming, artificial-noise covariance, and trajectory, and the E-UAV optimizes jamming power and mobility. To solve the resulting non-convex problems, we develop an alternating optimization framework based on DC reformulation, SDR, SCA, and SOCP-compatible surrogate subproblems. Simulation results show that the proposed scheme adaptively coordinates trajectory and power allocation, maintains positive secrecy rates under active jamming, and achieves smoother utility convergence than a Nash-game benchmark.