Observer-Based Identification of Nonlinear System Parameters

Abstract

This paper deals with an observer-based nonlinear system parameter identification method utilizing repetitive excitation. Although methods for physical parameter identification of both linear and nonlinear systems are already available, they are not attractive from a practical point of view since the methods assume that all the system, x, and the system input are available. The proposed method is based on a “sliding observer” and a least-square method. A sufficient condition for the convergence of the parameter estimates is provided in the case of “Lipschitz” nonlinear second-order systems. The observer is used to estimate signals which are difficult or expensive to measure. Using the estimated states of the system with repetitive excitation, the parameter estimates are obtained. The observer based identification method has been tested on a half car simulation and used to identify the parameters of a half car suspension test rig. The estimates of nonlinear damping coefficients of a vehicle suspension, suspension stiffness, pitch moment inertia, equivalent sprung mass, and unsprung mass are obtained by the proposed method. Simulation and experimental results show that the identifier estimates the vehicle parameters accurately. The proposed identifier will be useful for parameter identification of actual vehicles since vehicle parameters can be identified only using vehicle excitation tests rather than component testing.