Continuous Nonlinear SISO System Identification using Parameterized Linearization Families
Abstract
This paper presents a new approach to the modeling and identification of continuous nonlinear dynamic systems in terms of linear local models. In this approach, each local model is associated with a member of the linearization family of the original nonlinear system. Based on this family, a nonlinear model can be constructed, constituting an approximation of the nonlinear system around the entire equilibrium manifold. As a result, empirical model interpolation procedures are not necessary. It is also shown how this method can be used for plant identification. A numerical example demonstrates the efficiency of the method.
