On some weaknesses existing in optimal grasp planning

Optimal grasp planning has been investigated for over two decades. Still some important weaknesses in the previous work are worthy of notice: (i) The kinematic structure and the geometric configuration of a robot hand were ignored. Fingers were assumed to be capable of contacting an object anywhere. This is unrealistic. (ii) The grasp quality criterion was general and often did not match the task requirement. (iii) The criterion depends on the choice of unit and coordinate frame and lacks a clear physical meaning. This paper tries to remedy them. First, a general technique is proposed to find all feasible grasps on an object conforming to the robot hand. Next, for a specified external wrench or an external wrench set of a certain task, the maximum equilibrating contact force is adopted as the grasp quality criterion. Having an evident meaning, it is independent of the choice of unit and coordinate frame. Finally, an algorithm is presented for seeking the globally optimal grasp for which the value of the criterion is minimal among the feasible ones.