A Hierarchical Self-tuning Fuzzy Controller to Grasp and Lift Fragile Objects

N.I. Glossas and N.A. Aspragathos (Greece)


Robot Grasping, Self-tuning fuzzy control, Case-Based Reasoning.


This paper presents the development of a fuzzy control algorithm for a robot gripper, which is able to grasp safely fragile and delicate objects. The goal of this approach is to develop grippers, which can grasp an unknown object applying the minimal required force. The introduced control system is a two-level hierarchical fuzzy controller. The lower level part is a fuzzy controller that adjusts the motion of the fingers of the gripper tuning the applied grasping force. The higher part is a fuzzy decision maker based on Case-Based Reasoning (CBR), which modifies the characteristics of the lower level controller when the grasped object is changed. The proposed approach includes two phases: an off line training phase and an on-line operation phase. In the first phase, a genetic algorithm is used to optimize the membership functions of the fuzzy controller for a variety of grasped objects, and the obtained results are saved in a database. In the second phase, the CBR is used to determine the appropriate characteristics of the fuzzy controller in order to grasp objects of a wide mass and texture range. Simulated experiments are presented to demonstrate the efficiency of the introduced Self Tuning Fuzzy Logic Controller (STFLC) of a robot gripper, while its performance is compared with a simple Fuzzy Controller.

Important Links:

Go Back