Skip to Main Content
Summary form only given as follows. This paper addresses the problem of picking up moving objects in pseudo-random motion. There are many industrial tasks in which visual servicing is required to provide sophisticated guidance information for either the tracking or grasping of objects in motion. In the case of grasping, there will come a point at which the view of the object being tracked will become obscured by the actuating mechanism itself, whether the vision system is mounted on or off the actuation mechanism. Thus, it becomes necessary to predict the future state of an object that is desired to be grasped.