Research on Robot Grasping Moving Object Based on Dynamic Visual Servo
|Course||Pattern Recognition and Intelligent Systems|
|Keywords||Moving target capture Stereo Vision Dynamic visual servo Visual impedance CMAC network compensation|
Robots capture moving target robot visual servo control, complete tracking of moving targets and crawl, is a cutting-edge application subject of the intelligent robot, there is a good prospect in the field of industrial, aerospace and entertainment. Robot to capture the goals of the movement, has laid a good foundation for the robot to perform more complex tasks. Visual servo control with traditional sensor-based robot control compared with the more obvious advantages: greater flexibility, higher precision, more robust robot calibration error. Kinetic visual servo based on visual inspection, to calculate the acceleration of the robot and the joint torque, machine vision of the role of force feedback. This paper studies the the robot dynamic visual servo control technology, through the kinetic visual servo robots more flexible quickly moving target capture task, and optimized dynamic visual servo. Thesis, completed the following work: 1, treated according to the binocular stereo vision technology to capture the movement of the target three-dimensional position of the positioning in the camera coordinate and world coordinate the relationship between calibration based on the Kalman filter moving target visual feedback information filtering and estimated processing. 2, the robot unit composed of modules analysis discussed servo unit module dynamic and static accuracy and improve the accuracy of the commonly used methods. Analysis of the relations and the mechanical structure of the relationship of the geometry of the six degrees of freedom of the robot, the establishment of the kinematics and dynamics of the robot control system model, and using OpenGL model visualization simulation of robot control. 3, the optimization methods, robot target locus the decoupling joint coordinates of the trajectory. By Fuzzy Bang-Bang control the robot, the gripper to the time the best way to track and close to the moving target. In order to improve the stability of the system, when the position and the desired position of the joints of the robot is less than a given threshold value, a blur band integral servo control. 4, to study the dynamics of the robot visual servo control, and position-based visual impedance control algorithm for the robot to capture the moving target. The CMAC network compensation impedance control output of the visual optimization, and finally moving target capture control simulation experiments.