A robust extended Kalman filter for modeling piezoelectric actuators

Abstract

This paper presents an improved hysteretic model for a piezoelectric actuator (PEA) based nano-positioning system. PEAs exhibit hysteresis nonlinearities, which can dramatically limit control performance and accuracy of the nano-positioning system. The paper provides a design of a hysteretic nonlinearity estimator using a robust extended Kalman filter to determine the best estimate for the hysteresis model. A good quality and an accurate level of the model is an important task and a prerequisite that may significantly affect the control performance in nano-positioning systems even using an advanced controller. With the respect to the hysteresis model considered in [1], the proposed approach in this paper demonstrates a significant improvement to have perfectly matched the experimental data of the measured hysteresis curve of the displacement by comparing the one presented in [1].