Real-Time Target Tracking for Autonomous UAVs in Adversarial Environments: A Gradient Search Algorithm

摘要

In this paper, we present a real-time target tracking strategy for autonomous UAV operations in adversarial environments. The strategy generates the commanded heading and speed within the dynamic constraints of the UAV (i) to ensure that the UAV does not enter restricted regions (ii) to maintain the pursuit of the target (iii) to minimize the threat exposure level of the UAV. Probabilistic threat exposure map (PTEM) of the area of operation is generated by using a set of Gaussian probability distribution functions. PTEM defines various types of threats in a single framework and gives the risk of exposure to these sources of threat as a function of position. A steepest gradient search approach is utilized to determine in which direction the UAV should move to minimize the threat exposure or maximize the likelihood of avoiding a restricted region. Simulation results show the capability of the strategy to track a maneuvering target in an adversarial environment