Rope handling by a robot with visual feedback

Abstract

Visual information is necessary for a robot to be able to manipulate flexible objects such as a rope because flexible objects change their shape during motion. Actually direct visual feedback and verification in three-dimensional space are the keys to performing successful manipulation of such flexible objects. In this paper, important problems in hand-eye coordination are discussed and a rope handling experiment is described. The hand-eye system consists of a robot vision system, a general purpose manipulator and a Lisp system. The main visual functions adopted in the experiment are local image processing along a linear region and stereo vision. The experiment involves the performance of tasks such as inserting a rope into a ring and tying a rope. By coordinating the general purpose arm with stereo vision, our robot succeeds in manipulating a flexible rope.