Research On Multi-UAV Cooperative Standoff Tracking Method Based On Airborne Vision

Multi-UAV system will become an important combat unit in future wars.The use of multi-UAV system to track and monitor targets is an important application of UAVs.As the most common UAV task load,camera have played a key role in obtaining target information.The thesis takes fixed-wing UAVs as the research object,considers the application scenarios of visual-based standoff tracking of the target by multi-UAV system.The main contributions are as follows:(1)Linear rotation space based visual tracking algorithm(LARST)is proposed.In order to solve the tracking drift caused by the out-of-plane rotation of the target image,the linear rotation space is put forward and its update mechanism is given.The quality evaluation of the tracking results is introduced to determine the timing of updating the template or re-testing.The framework of linear rotation space based visual tracking method is given,and the visual tracking experiment is carried out using the drone perspective images database.Experimental results show that the proposed algorithm has a significant suppression effect on target out-of-plane rotation and tracking drift.(2)A modified Lyapunov guidance vector fields is proposed to achieve faster convergence to standoff circle and phase control.The thesis discusses the effect of introducing guidance parameter c on trajectory convergence.Based on the gradient descent method,an algorithm for searching optimal guidance parameter c is constructed.For multi-UAV cooperative tracking,a new phase adjustment algorithm based on speed control is proposed,and its stability is analyzed using the Lyapunov stability principle.The standoff tracking experiment of single UAV and the cooperative phase-keeping experiment of multi-UAV were conducted to verify the availability and effectiveness of the proposed algorithm.(3)A high-fidelity semi-physical simulation system were constracted and hardwarein-the-loop(HIL)experiments were carried out.The HIL simulation experiments demonstrated that the proposed algorithm can implement effective standoff tracking of targets in four motion states.This result shows that the multi-UAV cooperative standoff tracking algorithm proposed in this paper has excellent practicability.The thesis studies two important parts in the multi-UAV cooperative standoff tracking application based on airborne vision: visual object tracking and cooperative standoff tracking,which verifies the feasibility of related technologies and lays the foundation for subsequent practical applications.