Attractive Force Guided Grasping for a Teleoperated Manipulator System Based on Monocular Vision

Abstract

A method is presented here, based on the monocular vision, to provide haptic/force information for the operator which is used to signal to the operator the posture of end effecter relative to the object to be grasped. During the process of grasping, the information concerning the relative posture between the end effecter and the object is of key importance, especially when the end effecter is in the vicinity of the object. Although video information plays a major role in monitoring collision between robot and the object and fine-tuning the end effecter towards the object, the operator should make strict observations about the video images to take precautions against collision between robot and the object. As a result, operators might miss or neglect the advance notice of collision. In view of this problem, vision information obtained from a monocular camera is transformed into haptic information to which operators are more sensitive. Operators can give large master commands confidently without having to consider collision, and accordingly become released from the anxiety about the collision. At the same time, the efficiency of the telerobotic system is enhanced greatly.