Evaluation of two robot vision control algorithms developed based on N-R and EKF methods for the rigid-body placement

Abstract

There are many difficulties to overcome for the application of the vision system in the actual industry. The advantages of the proposed vision control algorithms are to make the robot move actively in the uncertain circumstance, even if the relative position between camera and robot, and the focal length and orientation of cameras are unknown. The reason is that the vision system model used for this study involves six camera parameters explaining the above uncertainty. Thus, the study estimates camera parameters and robot joint angles using the two different methods, such as N-R and EKF methods Finally, the positional accuracy and processing time of two vision control algorithms based on the two method are evaluated by performing the robot's rigid body placement task experimentally.