Multi-UAVs Cooperative Target Tracking

Compared with single UAV, multiple UAVs collaborated in target tracking is much more robust and have better performance in tracking. The key of multiple UAVs collaborated in target tracking is ground target prediction and track planning. As for target prediction, this paper gives a method of rolling forecast based on gray Verhulst model, which could update the history data continuously and predict the value of the forecast gradually, make it to become more precise and cost less time. As for track planning, firstly, this paper let the predicted position as the next target position, then allocate the targets and generate the track. Hungary algorithm allocate the best position to every UAVs, CAPT (Concurrent Assignment and Planning of Trajectories) algorithm could give the safest trail to the target, ensuring the security and efficiency of reforming the UAVs formation. As for flight control, this paper used PX4, for it is reliable and classic. Besides, this paper also builds a simulation environment in Gazebo-ROS, simulating six drones simultaneously, which verified the reliability of the algorithms and could reform the formation of the UAVs with higher timeliness.