Steady Tracker: Tracking a Target Stably Using a Quadrotor

Maneuvering target tracking of Unmanned Aerial Vehicle(UAV) in cluttered environments is a challenging issue owing to the unknown motion intention of the target and the complex moving environments. As the complexity of the environment increases, stable and secure target tracking is increasingly difficult to guarantee. To address the issue, this paper proposes a stable quadrotor tracking solution. The proposed solution contains two parts: target motion prediction and tracking path searching. The target motion prediction method predicts the future target motion based on the obtained target observations while considering observation noise and prediction errors. The tracking path searching method utilizes a sampling-based search method, using the homotopy of paths to ensure that the tracking path and the target position are in the same space. Finally, simulations, real-world experiments and statistical analysis verify the correctness and effectiveness of the proposed approach.