Adaptive synergy control for a dexterous artificial hand based on grasped object orientation

Abstract

An adaptive synergy controller is presented which autonomously modulates the finger synergies of a dexterous robotic hand according to the relative orientation of a grasped object. The adaptive synergy controller is derived from approximating the human motion of unscrewing a bottle cap with sinusoids to replicate the task with a robotic hand. In preliminary experiments, the finger joint motions of nine human test subjects were analyzed as they unscrewed a bottle cap with five different orientations of their hand with respect to the bottle. Results from these experiments show that the nature of the motions, while similar between participants, changed significantly as the orientation angle was varied. Data from human experiments were used to develop an adaptive synergy controller that autonomously modulates the artificial robotic finger motions in accordance with the orientation angle of the manipulator with respect to the grasped object. By choosing appropriate phase and amplitude parameters for the sinusoids used to drive the adaptive synergy controller, the complex motions involved in performing this task can be controlled by a single input. Experimental results of the adaptive synergy controller show that the control strategy successfully allows a dexterous robotic hand to unscrew objects in multiple orientations using only a single input.