Fuzzy neural friction compensation method of robot manipulation during position/force control

Abstract

Position and force controls are important and fundamental tasks of robot manipulators. In order to control position of the robot which simultaneously applies force to the environment, the friction between the robot and the environment has to be compensated. However, the friction force varies according to the applied force to the environment. Therefore, it is difficult to compensate the friction effectively with conventional controllers if we do not know the friction coefficient. Many researches have been done on fuzzy neural control, the combination of neural networks and fuzzy control, in order to make up for each other's weak points. The fuzzy neural control is expected to perform more sophisticated control than conventional control in an unknown environment. In this paper, we propose a new friction compensation method using the fuzzy neural network which contains a specialized neuron for friction compensation and a switch-learning. Simulation has done using a 3DOF planar robot manipulator to confirm the effectiveness of the proposed method.