AN INNOVATIVE RECURRENT CEREBELLAR MODEL ARTICULATION CONTROLLER FOR PIEZO-DRIVEN MICRO-MOTION STAGE

This paper proposes an innovative recurrent cerebellar model articulation control system (IRCMACS) for a micro-motion stage powered by linear piezoelectric motor (LPM). In this control system, the fast learning, generalization capability of the CMAC and good dynamic response of recurrent network are incorporated to handle nonlinear properties, disturbances of the LPM and noise to achieve high-precision positioning. According to getting rid of local minima problem, convergence speed of the system, the momentum and proportional elements are added into learning rules of the conventional CMAC to update parameters of the controller. The effects of disturbances and noise are vanished by an estimated bound compensator controller in the sense of the Lyapunov-Like Lema theory. The experimental results for the micro-motion stage system are shown to verify the effectiveness and applicability of the proposed control system for model-free nonlinear systems.