Identification of dynamics and hysteresis of Shape Memory Alloy (SMA) actuator using Laguerre filters

Abstract

Recursive approximation of a hysteretic system namely a Shape Memory Alloy (SMA) Actuator using Laguerre filters is experimentally tested in this paper. This approach permits us to capture simultaneously the Nonlinearity and the Dynamic of the system as opposed to most of the strategies dealt within the literature, where hysteresis is first compensated, then the dynamic is controlled. Two schemes using Linear Laguerre filter and Nonlinear Volterra Laguerre filters have been implemented to identify the SMA actuator with Hysteresis. An output prediction error model using classical Recursive Least Square (RLS) has been used to approximate the Hysteretic SMA. The Linear and Nonlinear Laguerre models have proven to be efficient with the choice of appropriate Laguerre pole. The two models have been compared and studied under different circumstances.

Online Estimation and Identification of Shape Memory Alloy-Actuated Flexible Structures Through Unscented Kalman Filtering

“…However, online identification schemes which require efficient recursive processing are few and limited to only black-box methods. An example is the study carried out by Kannan et al (2008) using linear Laguerre and non-linear VolterraLaguerre filters and the classic recursive least square (RLS) algorithm to identify a SMA actuator. Black-box methods are more suited to single-input single-output (SISO) systems; and even if some are capable of handling MIMO systems, they can model only the system inputoutput channels that are appropriately excitable and measurable.…”