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

Abstract

This paper presents a rule-based intelligent guidance strategy for autonomous pursuit of mobile targets by unmanned aerial vehicles (UAVs) in an area with threats, obstacles, and restricted regions. The probabilistic threat exposure map (PTEM) is used as the mathematical formulation of the area of operation for the guidance strategy to make intelligent decisions based on a set of defined rules. The rules are developed for three objectives in the order of priority as: 1) avoid obstacles/restricted regions; 2) maintain the target proximity; 3) minimize UAV threat exposure level. A least-square estimation and kinematic relations are used to estimate/predict the target states based on noisy position measurements. The work presented herein addresses the same problem as in a previous work by the authors, and aims at improving the computational efficiency without compromising the performance. Simulation results of several pursuit scenarios demonstrate the full capabilities of the strategy and the improvement over the previous work