This thesis presents the robot vision control schemes using batch and data moving schemes of Newton-Raphson (N-R) method, and the recursive scheme of Extended Kalman Filter (EKF) method for a moving target tracking, and slender bar placement in the un...
This thesis presents the robot vision control schemes using batch and data moving schemes of Newton-Raphson (N-R) method, and the recursive scheme of Extended Kalman Filter (EKF) method for a moving target tracking, and slender bar placement in the uncertainty of circumstance. The vision system model used for this study involves the six camera parameters(). explains the uncertainty of camera's orientation and focal length, and explains the unknown relative position between the camera and the robot. In order to develop the robot vision control scheme, the batch and moving data schemes of N-R, and a the recursive schemes of EKF are developed for the estimation of the six camera parameters. Then, based on the estimated six parameters using three cameras, control schemes of the robot‘s joint angles are developed for the moving targets tracking and slender bar placement using both N-R and EKF methods. Especially, for the placement of slender bar in the uncertainty of circumstance, the discontinuous trajectory caused by obstacles is divided into three obstacle regions, which are beginning region, middle region, and near target region. Each obstacle region involves of 10 obstacles. Then, the effects of number of obstacles using the proposed robot's vision control schemes are investigated in each obstacle region.
Finally, in order to evaluate its strengths and weaknesses of three robot vision control schemes. the proposed three robot's vision control schemes is demonstrated experimentally by performing the moving target tracking, and the slender bar placement in discontinuous trajectory by obstacle.