Game Theoretic Non-cooperative Dynamic Target Tracking for Directional Sensing-Enabled Unmanned Aerial Vehicles

Abstract

In this article, a game theoretic non-cooperative dynamic target tracking algorithm that empowers defensive unmanned aerial vehicles (UAVs), with directional sensing capabilities to track and collect information on intrusive UAVs, is proposed. Specifically, defenders aim to maximize the collection of identity information from intruders possessing anti-tracking and evading capabilities, while simultaneously preventing their entry into protected areas. Game theory is employed to determine the optimal confrontation paths for defenders against the intruders. The probability perception model is utilized for evaluating the dynamic target tracking capability and designing a tracking merit function to assess tracking performance, taking into account both the target's position and the perception relative angle. Furthermore, considering the dynamic interactive behaviors between intruders and defenders, the iterative linear quadratic game (ILQG) algorithm is employed to solve the Nash equilibrium of the non-cooperative target tracking game. Through simulation experiments, the effectiveness of the proposed algorithm in accomplishing multi-agent dynamic target tracking tasks is demonstrated and the performance of the algorithm under varying parameters in the intruder's cost function is evaluated, which represent different intrusion intentions.