Positioning Research For Robot Adaptive Control Based On Image

Robot visual servoing is that, image is extracted by visual sensor, image feedbackinformation is processed quickly and efficiently, and robot is controlled by the processingresults. In recent years, due to its high accuracy, high flexibility and strong robustness, ithas attracted widely attention and research. In this paper, based on binocular vision modelirrespective of depth, robot visual servo control based on image is studied, and it canfinish positioning function of stationary target. On the basis of giving a tutorialintroduction to robot visual servoing control, the main work is summarized as follows:First of all, for binocular hand-eye and six-DOF manipulator system, vision model,manipulator kinematics model, and manipulator dynamics model are established. So therelevant concepts of manipulator visual servo control understand in order to carry onresearch to prepare.Secondly, according to the adaptive law of parameter, image Jacobian matrix isestimated online. A relatively simple adaptive proportional controller is put forward toposition static target by manipulator, and its stability has been proved by Lyapunovtheorem. So overall understanding for visual servo control research has been established.Again, manipulator dynamics characteristic is further considered in the paper. ImageJacobian matrix is estimated online by adaptive law of parameter. An adaptive PDcontroller is proposed to in order to make manipulator more effective complete visualpositioning tasks. Its stability has been proved by Lyapunov theorem.Finally, the joint angular velocity is difficult to accurately measure so that thisinfluences the performance of the system. The joint angular velocity of manipulator isestimated online by sliding-mode observer, and image Jacobian matrix is estimated onlineby adaptive law of parameter. An adaptive sliding-mode PD controller is put forward torealize manipulator position static target, and its stability has been proved by Lyapunovtheorem.