Improved Initialization for Nonlinear State-Space Modeling

TL;DR

This paper introduces a new initialization algorithm for nonlinear state-space models that improves estimation accuracy by separately identifying linear and nonlinear components using simple regression methods, validated on benchmark and real data.

Contribution

The paper presents a novel initialization approach that transforms nonlinear dynamic problems into static ones for efficient parameter estimation.

Findings

Effective initialization improves model estimation accuracy.

Method validated on Wiener-Hammerstein benchmark.

Successful application to crystal detector identification.

Abstract

This paper discusses a novel initialization algorithm for the estimation of nonlinear state-space models. Good initial values for the model parameters are obtained by identifying separately the linear dynamics and the nonlinear terms in the model. In particular, the nonlinear dynamic problem is transformed into an approximate static formulation, and simple regression methods are applied to obtain the solution in a fast and efficient way. The proposed method is validated by means of two measurement examples: the Wiener-Hammerstein benchmark problem, and the identification of a crystal detector.