Grasp synthesis based on tactile sensation in robot manipulation of arbitrary located object

This paper analyzes grasp synthesis in multi-fingered robot arm equipped with a tactile sensor to handle objects located at arbitrary positions. We developed an 11-dof multi-fingered arm and a novel optical three-axis tactile sensor system based on an optical waveguide transduction method. The tactile sensor can simultaneously acquire normal and shearing forces. We analyzed normal and shear force distribution to grasp real object. The slippage direction acquired from the shear force data was used to control arm movements in the incipient grasp and release motion. Normal force is used to distinguish object stiffness and to generate optimum grasp pressure. The analysis results are compiled in a control algorithm inside the robot control system. The algorithm was evaluated in experiments with real objects, and the results revealed good performance of the robot arm when handling arbitrary located objects.