A grasp criterion for robot hands considering multiple aspects of tasks and hand mechanisms

Abstract

In this paper, a task-oriented grasp criterion for robot hands is discussed. The objectives of grasp criteria proposed in the past are stable grasping and the task that require adequate forces and moments on the grasping object. These criteria do not consider the other parameters such as positions and velocities on the grasping object and the mechanical property of robot hands. We propose a task-oriented grasp criterion that evaluates the feasibility of the task with considering hand mechanisms. Our proposed method find the grasping position by considering the efficiency of grasping object's positions, velocities, and forces. As a result, this criterion could detect the grasping position like humans do.