Adaptive visual servo control of UAV Ground-Target-Autonomous-Tracking System

A novel adaptive servo control method is proposed for UAV GTATS(Ground-Target-Autonomous-Tracking System),which consists of basic control law for UAV and visual tracking controller for GTATS. The adaptive servo control method only depends on target information in the image plane and Kalman filtering technology. Based on this proposed method, a dynamic motion target can be tracked without target's 3D velocity. Synchronously, in order to estimate the optimal system state and target image velocity which is used later by the visual tracking controller, a self-tuning Kalman filter is adopted to estimate interesting parameters on-line in real-time. Further, Because the visual tracking controller is working entirely in image space, the dynamic characteristics of the image signal are analyzed and a kinematics model is developed for the target in the image plane by the geometrical relations among the UAV, the target and the camera. Finally, The performance of the controller is demonstrated by theoretical stability analysis.