Fuzzy logic grasp control using tactile sensors

This paper presents a method to control a two-fingers gripper for safe grasping of fragile and delicate objects like glass, fruits or vegetables. The proposed control algorithm adjusts the motion of the fingers of a gripper using tactile feedback, so that an object is grasped with the minimal required force without the measurement of this force. The rules set of the control law is derived empirically by investigation of the human finger skills. The formulation of these rules and the inference mechanism are based on fuzzy logic. The controller receives the object velocity and acceleration and the detection of incipient slippage and adjusts the fingers motion. The main advantage of this method is that the designed controller is quite robust because knowledge concerning the size, the weight or the surface texture of the grasped object is not required. The object is stably grasped by applying a relatively low grasping force. Simulated experiments are presented to demonstrate the efficiency of the introduced fuzzy controller of robot grippers, while its performance is compared with a PID controller. (C) 2001 Elsevier Science Ltd. All rights reserved.