Approach for Real In-Hand Dexterity Evaluation: Application to the RoBioSS Hand

Abstract

The development of flexible multidigital grippers with both adaptive grasping and in-hand manipulation capabilities remains a complex issue for human-like dexterous manipulation. After four decades of research in dexterous manipulation, many robotic hands have been developed. The evaluation of these hands remains a key challenge. The design of the hand (degree of actuation, number of fingers, and joints) does not guarantee the real behavior of the hand and its performance. For comparison and evaluation purposes, the literature often refers to a potential dexterity, but very rarely provides an evaluation of the real dexterity of these grippers. The RoBioSS hand offers a high level of dexterity with inside-hand manipulation capabilities. In order to characterize the performance of the hand with regard to the state of the art, this article proposes an approach based on the dexterity analysis. It is based on the use of a known reference object and on defined grasps and movements of the object. The results presented allow us to evaluate grasps capabilities and quality of the object's movement inside the hand.