Intelligent Grasping of an Underactuated Hand for Space Robots

Minzhou Luo,Chaohong Lu,Tao Mei,Shihong Zha
2005-01-01
Abstract:Based on underactuation principle, an underactuated robot hand is developed to obtain practical applications in space service. This kind of robot hand can adapt the grasped object shape and output large grasping force. In this paper, we focus on the problem of intelligent grasping of the underactuated robot hand, thus is characterized by autonomous grasping decision and control. Firstly, the characteristics and versatile grasp patterns of the underactuated robot hand are presented. Secondly, the grasping rules are extracted from the original experiment sample data. By fuzzy neural network, grasping map is established between the comprehensive fuzzy features, including grasping task and object features, and the grasping patterns. According to the selected grasping pattern, the control strategies and architecture based on force sensors feedback are proposed. At last, the experimental results demonstrate that autonomous decision and control method can achieve versatile grasping pattern and stable grasping.
What problem does this paper attempt to address?