Identification of Dynamic Model for Ultra-Precision Positioning System

Abstract

Abstract: In order to improve the control precision of the piezoceramic actuator in precision position system, a dynamic modeling of Ultra-precision Positioning System is proposed. A two-step identification strategy is adopted in this paper. Step 1: The linear model was identified based on step input response by least square method. Step 2: According to the step.1 result, the designed nonlinear observer was applied to estimate the output of hysteresis by Fast Tracking Differentiator. The polynomials are used consistently approximate continuous functions of f(.) and g(.) in the Duhem modeling based on the theorem of the Weierstrass polynomial approximation theory. Recursive least square methods are developed to identify α and the coefficients of the polynomials in Duhem model. At last , the simulation results show applicability, rapid convergence velocity and high calculation accuracy of the algorithm.