Identification and control of linear dynamics with input Preisach hysteresis

Abstract

In this paper, the identification and control of linear dynamics with input Preisach hysteresis is presented without the measurement of hysteresis. Simulations indicate that the control action for the same reference can converge to different values and switch between several values under different PID controllers. Firstly, least squares estimation by singular value decomposition(SVD) is applied to get the density function of hysteresis but with an unseparated factor where the truncation of singular values is a trade-off between the truncation error and norm of pseudo-inverse, then the identification of the linear dynamics can be given by least squares estimation. Finally, the linearization of hysteresis with the identification result is presented to improve the tracking performance of complex signals, and the performance of the PID controllers with linearization is better than that without linearization. © 2010 AACC.