Dynamic Target Tracking Under The Collaboration Of Drones And Unmanned Vehicles

With the development of science and technology,people's requirements for robots are becoming higher,which makes a single robot have certain limitations in completing complex tasks.The collaborative system of unmanned aerial vehicle(UAV)and unmanned ground vehicle(UGV)has been widely concerned by scholars at home and abroad because of its high complementarity in load,speed and communication.The cooperative system of UAV and UGV mainly involves scientific problems such as perception,decision making and positioning.In this paper,the collaborative system of UAV and UGV is taken as the research object to solve the problem of how to complete the dynamic target tracking task together.This paper mainly studies three sub-problems: target tracking based on vision?co-location?environment modeling and path planning.The main research are as follows:First,the collaborative system of UAV and UGV adopts a distributed framework to define the task of each agent.The experiment platform of UAV and UGV is built,and the real-time communication between them is completed by using ROS(robot operating system).Secondly,the improved KCF algorithm proposed in this paper is used to track the dynamic target by using the visual field advantage of UAV.The position information of the dynamic target in the UAV coordinate system is solved according to April Tag,which is taken as the navigation target point of the UGV,and the improved A* algorithm is used to UGV's path planning.Thirdly,this paper solves the problem of co-location between UAV and UGV.Extended Kalman filter is used as the framework of the heterogeneous robot system.During collaborative positioning,the observation information provided by the UAV to the UGV may conflict with the positioning information of UGV itself,which may lead to the divergence of the EKF algorithm in serious cases.Based on the static game theory of complete information,the UGV makes a judgment on whether to choose the collaborative positioning,which improves the positioning accuracy of the UGV.Fourthly,this paper solves the problem of map construction in collaboration between UAV and UGV.Firstly,the stereo camera carried by the UAV is used to construct a rough two-dimensional grid map for the navigation of UGV.Secondly,while tracking the dynamic target,the UGV updates the map by using the single-line lidar carried by itself to get an accurate grid map.When obstacles change,the coordinate of the dynamic target is refreshed and the path of UGV is replanning.Fifth,this paper takes DJI Robotmaster S1 as the dynamic target,and verifies the correctness of all the algorithms proposed in this paper under the real environment.The results show that the proposed algorithms can meet the requirements of dynamic target tracking tasks.