Grasping and Coordinated Manipulation by a Multifingered Robot Hand

Abstract

A new avenue of progress in the area of robotics is the use of multifingered robot hands for fine motion manipulation. This paper treats two fundamental problems in the study of multi fingered robot hands: grasp planning and the determination of coordinated control laws with point contact models. First, we develop the dual notions of grasp stability and grasp manipulability and propose a procedure for task modeling. Using the task model, we define the structured grasp quality measures, and using these measures we then devise a grasp planning algorithm. Second, based on the assumption of point contact models, we develop a computed torque-like con trol algorithm for the coordinated manipulation of a multi fingered robot hand. This control algorithm, which takes into account both the dynamics of the object and the dynamics of the hand, will realize simultaneously both the position trajec tory of the object and any desired value of internal grasp force. Moreover, the formulation of the control scheme can be easily extended to allow rolling and sliding motion of the fingers with respect to the object.