Model based control of a piezo-actuated axis

Abstract

This paper presents the physical state space model of the static und dynamic behaviour of a piezo-actuated axis. The starting model parameters are fitted to measured data by an optimisation algorithm based on Newton's method. The state space description of the axis is used to design a controller architecture with a model based state observer. The measured closed loop frequency response of the model based control is compared to simple linear controllers like PID elements. To reduce noise, the feedback signals for the model based control are estimated with a Kalman filter which bases on the linear model of the open-loop system.