Research Of Cooperative Ground Target Tracking For Multiple UAVs In Complicated Environment

With the rapid development of UAV technology and application, being a subtask of theUAVs application, subject of ground target cooperative tracking has received manyattention and research interest of many experts and scholars. Focusing on groundmaneuvering target collaborative tracking for multi UAVs in complicated environment, thisthesis studied two issues of state fusion estimation of ground maneuvering target andtrajectory planning of observation for UAVs.Firstly, this thesis analyzes the framework of solving the problem for multi UAVscollaborative target tracking under complex environment, and presents a diagram ofhierarchical distributed initiative solving structure and simple explanation. There are manyconstraints for UAV target tracking process, including the UAV kinematic constraints,observing restriction of sensors, communication topology changes and UAV airspaceconstraints. In response to these constraints, the thesis carries out a detailed description ofmathematical modeling and provides a mathematical foundation for the subsequent targetstate fusion estimation and tracking path planning.Secondly, the method of state fusion estimation of a “smart” anti-tracking maneuveringtarget on ground is deeply studied. Observation and estimation of the target state is animportant basic condition to complete target tracking mission, especially for highlymaneuvering target. The ability of effective estimation for the target state determines thefeatures of real-time target tracking for UAVs. Based on the analysis of Extended KalmanFilter (EKF) and Unscented Information Filter (UIF) algorithm, this thesis extends theMinimax Filter (MF) to nonlinear system and presents in the form of distributedconsistency MF filter to apply in target state estimation for multi UAVs. The method solvesthe following problems:1) In the circumstance of tracked maneuvering target is ‘Smart’,that can make counter-track and counter-surveillance, how to keep tracking target state,continuously;2) In the condition of communication topology changes and limitedmeasurement of sensors, how to estimate target state, cooperatively;3) Distributedconsensus target fusion estimation for multiple UAVs. Finally, UAV trajectory planning algorithm for maneuvering target tracking in presenceof static obstacles and dynamic threaten sources is studied. UAV will be restricted byairspace during flight, this thesis combines collision avoidance potential function and therelative velocity space dynamic programming method with Lyapunov Guidance VectorField(LGVF) to solve the UAVs target tracking process in the presence of static obstaclesand dynamic threaten sources, on which the LGVF guides the UAV to track a ground targetin a specified standoff circle.