Identification of Hammerstein models for control using ASYM

Y. Zhu

Research output: Contribution to journalArticleAcademicpeer-review

48 Citations (Scopus)
1 Downloads (Pure)

Abstract

Identification of single-input single-output Hammerstein models is studied in this work. The basic idea here is to extend the recently developed asymptotic method (ASYM) of linear model identification to include input non-linearity in the model set. First identification test design will be discussed. In parameter estimation, prediction error criterion is used in order to maintain consistence when the process is operating in closed-loop. A relaxation iteration scheme is proposed by making use of a model structure in which the error is bilinear in the parameters. The order of the linear part and nonlinear part are determined by looking at an output error related criterion which is control-relevant. The frequency domain upper error bound of the linear part will be derived and used for model validation. Simulation study will be used to illustrate the method and comparisons with other methods are also given.
Original languageEnglish
Pages (from-to)1692-1702
JournalInternational Journal of Control
Volume73
Issue number18
DOIs
Publication statusPublished - 2000

Fingerprint

Dive into the research topics of 'Identification of Hammerstein models for control using ASYM'. Together they form a unique fingerprint.

Cite this