A hierarchical decision-making algorithm of UAV for collision avoidance in target tracking
Yunyun Zhao, Xiangke Wang, Lincheng Shen, Weiwei Kong · 2017
Effective target tracking and obstacle avoidance strategies are essential to the success of unmanned aerial vehicle (UAV) missions. In this paper, we consider the UAV action decision-making problem for tracking a ground moving target, while avoiding dynamic ellipsoidal obstacle detected en route. The studied problem was formulated as a Partially Observable Markov Decision Process (POMDP), and then in this framework, we get the optimal or sub-optimal strategy to achieve the target tracking in the barrier environment. We first improved the conditions and methods of collision detection based on collision cone, and use common penalty function method to ensure tracking safety. Further taking into account that the UAV needs to react more quickly to obstacles than target tracking, we propose an alternative hierarchical decision-making approach. For the entire tracking process, the total information gain in the UAV-to-target observation is used as a tracking optimization criterion in case of no obstacle. Simulations are presented to demonstrate the comparison between the proposed approach with the improved penalty function approach. The results show that our proposed algorithm is superior.