Object shape recognition and grasping by five-fingered robotic hand based on E-ANFIS model

This paper presents a method for object shape recognition and the real-time grasping of unknown plane objects only using the positions information of robotic hand fingers. We aim to recover the shape, position and orientation of the grasped object by random five grasping points and the vision system in Cartesian space. A highly intelligent information fusion model which is combined with the expert system and multiple ANFIS models (E-ANFIS model) is presented to solve the problem above and the validity of the proposed method is demonstrated by simulations.