Inversion of non-linear stochastic models for the purpose of parameter estimation

DOI10.1080/00207170110090651MaRDI QIDQ3151509zbMATH OpenOpenAlexFDO

Authors Ola Markusson, Håkan Hjalmarsson

Publication date 16 October 2002

Published in International Journal of Control (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1080/00207170110090651

zbMATH Keywords

feedbackinversionstochastic model

Mathematics Subject Classification ID

Applications of mathematical programming (90C90) Estimation and detection in stochastic control theory (93E10)

Review text

Estimation methods (such as the maximum likelihood (ML) method and the prediction error method) of nonlinear stochastic models require inversion of the model, which can be quite a difficult task. In this paper the authors show that model inversion can be implemented in numerical software by means of a feedback connection based on the model. Using the feedback configuration this inversion can be done without involving an equation solver if the model can be partitioned as a linear part in parallel with a nonlinear part where the time-delay in the nonlinear part is larger than in the linear one. Another benefit of the feedback structure is that the computation of the derivatives, used for optimization of the cost function, can be implemented as a linear time-varying filter. The feedback configuration was used in ML estimation on two numerical examples.

Recommendations

• Linear prediction error methods for stochastic nonlinear models
• scientific article; zbMATH DE number 4145158
• Recursive prediction error parameter estimator for non-linear models
• An innovation representation for nonlinear systems with application to parameter and state estimation
• Nonlinear estimation

Cites work

• Gradient expressions for a closed-loop identification scheme with a tailor-made parametrization
• scientific article; zbMATH DE number 1001723 (Why is no real title available?)
• scientific article; zbMATH DE number 432498 (Why is no real title available?)
• scientific article; zbMATH DE number 439900 (Why is no real title available?)
• scientific article; zbMATH DE number 48093 (Why is no real title available?)
• scientific article; zbMATH DE number 3997615 (Why is no real title available?)
• scientific article; zbMATH DE number 852532 (Why is no real title available?)
• Interactive system identification: prospects and pitfalls
• Maximum likelihood and prediction error methods
• Nonlinear black-box modeling in system identification: A unified overview
• Nonlinear black-box models in system identification: Mathematical foundations
• Nonlinear model-based control using second-order Volterra models
• Right inverse of a discrete time non-linear system

Cited in

(2)

• A general framework for approximated model stable inversion
• From experiment design to closed-loop control

This page was built for publication: Inversion of non-linear stochastic models for the purpose of parameter estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3151509)