Air Shepherd: Trajectory Prediction-Based Target Localization and Circumnavigation in Cluttered Environments

This paper proposes a trajectory prediction-based target localization and circumnavigation pattern for cluttered three-dimensional environments, which is more realistic and suitable for more complex environments than traditional patterns. The main work of the paper consists of two parts: tracking based on trajectory prediction and circumnavigation based on broadcast information. On the one hand, the tracking Autonomous Aerial vehicle (AAV) obtains target trajectory prediction based on the B-spline curve, and then achieves target localization and tracking through front-end search and back-end optimization. On the other hand, without communicating with each other, a distributed control strategy is presented so that the multiple circumnavigation AAVs can achieve target circumnavigation and reciprocal avoidance by only observing the status of adjacent AAVs. In the simulation, obstacle avoidance vehicles moving freely at different speeds are selected as targets in two scenarios and the simulation results are given to verify the effectiveness of the proposed approach. Furthermore, a hardware-in-the-loop experiment and a overall system validation experiment are designed to verify the feasibility of the algorithm. Note to Practitioners—Practical tracking scenarios often involve numerous obstacles, and it is difficult to determine the location of the target being tracked and surrounded. Many existing methods assume that the trajectory or position of the tracked target is already known, but in reality, this assumption is difficult to fulfill due to the agility of the target and the complexity of the environment. Compared to existing work, this paper proposes an ‘air shepherd’ mechanism, in which a tracking AAV plays the role of a ‘sheepdog’, and based on trajectory prediction, the target can still be tracked even when it is out of sight. The circumnavigation AAVs play the role of the ‘sheep’, achieving pursuit and encirclement without requiring communication between each other while ensuring reciprocal avoidance.