Visual feedback control for tracking and intercepting 3D moving object

This paper describes the visual feedback control system that can track and intercept a 3D moving object on conveyor belt system. We assume that the 3D object consists of quadratic surfaces. Laser range finder is located on an end-effector of the robot manipulator to obtain 3D range data of the object. We have used the surface segment fitting technique to determine the relative pose of the 3D object and robot. The pose estimation of object is converted to nonlinear optimization problem to minimize an error between the measured surface patch and the surface from the CAD model. We have performed computer simulations and experiments using integrated robot-conveyor and vision system. It demonstrates that our algorithm can track and intercept the 3D object moving on a conveyor belt