A cluster based fuzzy controller for grasp and lift fragile objects

In this paper, an intelligent control architecture based on Fuzzy Logic, for a robotic grasping system capable of lifting and manipulating an uknown and fragile object is presented. A Fuzzy Logic Controller that adjusts the closing velocity of the fingers of a gripper based on clustering is designed. Initially an input-output data set is collected from successful grasp and lifting of a wide range of objects. This input-output data set are used to train the fuzzy controller based on subtractive clustering. The conceptual design of the proposed fuzzy controller is based on the investigation of the neurophysiology of the human fingers in grasping objects without or with minimal knowledge of their characteristics. Particularly for fragile objects the human fingers apply just the required force to hold the object. The proposed fuzzy controller imitates the behaviour of the human control system for grasp and lifting objects but still it is far from reaching this behaviour. Simulated experiments are presented to demostrate the efficiency of the proposed intelligent controller for a wide range of objects weight and texture which are not used in the training of the controller.