Abstract：This paper presents a direct image adaptive control scheme for visual tracking of moving target. It is different from feature based visual servo. The manipulating process is as follows: first of all, move the arm to the desired manipulating position and orientation with respect to the object of interest then sample the image of the object by eye-in-hand camera. In the manipulating process, the scheme, which compares the real time sampled-image with the desired one, forms the image error. Then the control action composed of image error feedback and adaptive motion compensation-steers the gripper to track the moving target in order that the image error is eliminated. If the manipulating-process does not satisfy very well the initial error condition, the control law can be used for to track the moving target with accurate and robust properties.
1 Vidyasager.System Theory and Robotics.IEEE Control System Magazine,1987,2(7):16～17 2 Papanikolopoulos N P,Khosla P K,Kanade T.Visual Tracking of a Moving Target by a Camera Mounted on a Robot:a Combination of Control and Vision.IEEE Trans on Robotics and Automation,1993,1(9):14～35 3 Hashimoto K,Noritsugu T.Observer-based Control for Visual Servoing.Preprints of 13th IFAC World Congress,1996,453～458 4 Horn B K P.Robot Vision.The MIT Press,1986 5 李介谷.计算机视觉的理论和应用.上海交通大学出版社,1991