A Nonlinear System Identification Method based on Local Linear PLS Method

Takashi Shikimori, Hideo Muroi, and Shuichi Adachi


Identification, Nonlinear systems, Local linear model, Partial Least-Squares method, Recursive Least-Squares method


Nonlinear system identification is one of the most important topics in system identification theory. In this paper, a new nonlinear system identification method using Partial Least-Squares (PLS) method is proposed, which is called a local linear PLS method because it is based on local models. The proposed method consists of two steps. First step is to identify local linear models by using the conventional Recursive Least-Squares (RLS) method. Second step is to identify a virtual system which describes a nonlinearity of the identified object, by PLS method. The effectiveness of the proposed method is shown through numerical simulations.

Important Links:

Go Back