A SELF-ORGANIZING FUZZY CEREBELLAR MODEL ARTICULATION CONTROLLER BASED OVERLAPPING GAUSSIAN MEMBERSHIP FUNCTION FOR ROBOTIC SYSTEM WITH UNCERTAINTIES

Abstract

This paper proposes an effective intelligent control system for nonlinear objects. The selected object is n-link robotic manipulator model. The robotic manipulator system is always facing uncertain changes in its dynamics. To solve this problem, an intelligent control system consisting of a new self-organizing fuzzy cerebellar model articulation controller (NSOFC) plays a key role built by a cerebellar model articulation controller (CMAC) with a sliding mode control (SMC) to estimate the ideal controller and the compensator for eliminating the approximation error. The remarkable thing about the conventional cerebellar model (CMAC) is the reuse of previous data to synchronously mix the current state for more accurate tracking error checking. The control system not only adjusts model parameters on-line based on Lyapunov stability theory but also restructures the main controller – increasing or decreasing layers automatically. Finally, the robot manipulator 2 DOF is provided to demonstrate the effectiveness of the proposed control system through the experimental results.