“Lure the Enemy in Deep”: Confronting Rogue UAV Through Diverse Hybrid Jamming

This research addresses the challenge in capturing rogue unmanned aerial vehicle (UAV) strategies based on communication and navigation jamming technology in the absence of observable UAV communication performance parameters. The jamming involves remote communication jamming (RCJ) power control, as well as navigation jamming (NJ) and navigation deception (ND) combination, hence named as the diverse Hybrid jamming (DHJ) strategy. Firstly, we construct the DHJ model, which analyzes the dynamic impact mechanism of DHJ on UAV flight control systems based on common control protocols. Subsequently, we introduce the concept of the Situation Estimation Ring (SER) to overcome reward-driven deficiencies arising from incomplete observation and optimize features of the state space, which essentially considers the flight status feedback mechanism of the UAV to DHJ attacks and will be used to construct the DHJ Markov Decision Process (DHJ-MDP) mathematical framework. Then, the DHJ attack process against UAV is formulated as a game framework, where the counter-UAV system (CUS) generates DHJ strategy solely based on spatial positional states to deceive and capture the UAV, and the UAV adopts flight control strategies to evade jamming based on communication quality and navigation positioning information. Finally, the analysis of Nash equilibrium (NE) properties explain that the SER-based dynamic spatial positioning of UAV can be used for CUS to derive DHJ equilibrium strategies, thereby providing a theoretical foundation for solving these strategies using knowledge-based deep reinforcement learning (DRL) algorithms, with the algorithm’s complexity also being proven. Simulation and comparison experiments demonstrate that the solved DHJ strategy effectively drives UAV toward deceptive destinations under incomplete observation states, achieve the ’lure the enemy in deep’ effect and the proposed SER-base algorithms enhances algorithm’s convergence speed.