Application of CMAC control in servo system with friction

To resolve the wave distortion caused by friction when angular velocity passes through zero, this paper presents a digital tracking controller based on CMAC neural network (cerebellum model articulation controller) after thoroughly consideration of the friction factors in the servo system. The tracking controller consists of the feedforward controller, the PD controller and the CMAC controller. The neural network controller percepts and self-learns the moment force of friction and other disturbance, and further provide appropriate compensation. The proposed controller is shown to be effective and valid to meet the requirement of real time work, and has engineering reference value to improve dynamic tracking performance of high precise servo system.