Improved Initialization for Nonlinear State-Space Modeling

• This paper discusses a novel initialization algorithmfor the estimation of nonlinear state-space models. Good initialvalues for the model parameters are obtained by identifyingseparately the linear dynamics and the nonlinear terms inthe model. In particular, the nonlinear dynamic problem istransformed into an approximate static formulation, and simpleregression methods are applied to obtain the solution in a fast andefficient way. The proposed method is validated by means of twomeasurement examples: the Wiener-Hammerstein benchmarkproblem, and the identification of a crystal detector.

Ämnesord

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Reglerteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Control Engineering (hsv//eng)

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Signalbehandling (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Signal Processing (hsv//eng)

Nyckelord

nonlinear modeling

Multilayer perceptrons

state-space models

system identification

nonlinear dynamical systems

Publikations- och innehållstyp

art (ämneskategori)

ref (ämneskategori)