Ground Target Tracking With UAV Based On Kalman Filter

One of the main applications of UAV is to provide support for surveillance and monitoring mission in modern warfare and future military or civilian uses. If a UAV is able to trace a ground target precisely, extend the monitoring range and avoid threats effectively, the efficiency of surveillance, even the military attack missions will be improved greatly.Therefore the goal of this research is to work on a plan which can make the UAV trace a ground moving target autonomously, and can also generate a flying path based on the mission requirements and practical circumstances.The main content of the thesis is as follows: First, the structure of the target tracing system is designed which includes: ground target tracing module, ground threats avoiding module and global path planning module; Second, the target model is developed which is a IMM structure of the combination of CT model and Wiener velocity model, and IMM-UKF is presented as the target state estimation filter of the ground target tracing module. Third, two kinds of path planning methods are proposed. In a secure situation, Lyapunov vector field is constructed based on Lyapunov stability theory in order to make the UAV fly around the target, and the target will always stays in the sensors' monitoring range through setting the radius of the circular flying path. Thus UAV can trace the target and the monitoring range is extended. On the other hand, velocity direction vector field is proposed to generate the flying path in an environment with threats. After analyzing the influence of the target and the danger zone exert on UAV, the pulling vector field, avoidance vector field and guidance vector field are constructed, pulling vector field is used to pull UAV toward the target, avoidance vector field is to push UAV from the danger zone and guidance vector field is to guide UAV to round the danger zone. The velocity direction of UAV is the result of the superposition of these three vector fields. The practical performances of UAV are also considered as the constraints in this method. Fourth, the global path was generated based on the target state estimation from ground target tracing module, and combined circling flying mode or velocity direction vector field.