Abstract
A theory for identification and classification of two-fingered grasps is described and demonstrated. It is shown that optimal finger positions are obtained from the pairwise intersections of three subsets in the configuration space for the gripper: the symmetry and antisymmetry sets, and the critical set of the grasp map. Furthermore the three sets themselves form boundaries in a natural partition of configuration space. The partition corresponds to a classification of possible grasps according to their stability properties. An implementation of the theory has been developed for an ADEPT robot with visual sensing.

This publication has 10 references indexed in Scilit: