A Bayesian approach for model identification of LPV systems with uncertain scheduling variables

Farshid Abbasi, J. Mohammadpour, Roland Toth, N. Meskin

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

7 Citations (Scopus)

Abstract

This paper presents a Gaussian Process (GP) based Bayesian method that takes into account the effect of additive noise on the scheduling variables for identification of linear parameter-varying (LPV) models in input-output form. The proposed method approximates the noise-free coefficient functions by a local linear expansion on the observed scheduling variables. Therefore, additive noise on the scheduling variables is reconstructed as a corrective term added to the output noise that is proportional to the squared gradient obtained from the posterior of the Gaussian Process. An iterative procedure is given so that the obtained solution converges to the best estimation of the coefficient functions according to the given measure of fitness. Moreover, the expectation and covariance functions estimated by GP are modified for the noisy scheduling variable case to include the noise contribution on the estimated expectation and covariance functions. The model training procedure identifies noise level in the measurements including outputs and scheduling variables by estimating the noise variances, as well as other defined hyperparameters. Finally, the performance of the proposed method is compared to the standard GP approach through a numerical example.

LanguageEnglish
Title of host publication2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Pages789-794
Number of pages6
ISBN (Electronic)978-1-4799-7885-4
ISBN (Print)978-1-4799-7884-7
DOIs
StatePublished - 2015
Event54th IEEE Conference on Decision and Control (CDC 2015) - "Osaka International Convention Center", Osaka, Japan
Duration: 15 Dec 201518 Dec 2015
Conference number: 54
http://www.cdc2015.ctrl.titech.ac.jp/

Conference

Conference54th IEEE Conference on Decision and Control (CDC 2015)
Abbreviated titleCDC 2015
CountryJapan
CityOsaka
Period15/12/1518/12/15
Internet address

Fingerprint

Linear Parameter-varying Systems
Model Identification
Bayesian Approach
Identification (control systems)
Gaussian Process
Scheduling
Covariance Function
Additive noise
Additive Noise
Output
Hyperparameters
Bayesian Methods
Coefficient
Iterative Procedure
Fitness
Directly proportional
Gradient
Converge
Numerical Examples
Term

Cite this

Abbasi, F., Mohammadpour, J., Toth, R., & Meskin, N. (2015). A Bayesian approach for model identification of LPV systems with uncertain scheduling variables. In 2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan (pp. 789-794). [7402326] Piscataway: Institute of Electrical and Electronics Engineers. DOI: 10.1109/CDC.2015.7402326
Abbasi, Farshid ; Mohammadpour, J. ; Toth, Roland ; Meskin, N./ A Bayesian approach for model identification of LPV systems with uncertain scheduling variables. 2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan. Piscataway : Institute of Electrical and Electronics Engineers, 2015. pp. 789-794
@inproceedings{8963a93116bf49049cbb6bf24202d43c,
title = "A Bayesian approach for model identification of LPV systems with uncertain scheduling variables",
abstract = "This paper presents a Gaussian Process (GP) based Bayesian method that takes into account the effect of additive noise on the scheduling variables for identification of linear parameter-varying (LPV) models in input-output form. The proposed method approximates the noise-free coefficient functions by a local linear expansion on the observed scheduling variables. Therefore, additive noise on the scheduling variables is reconstructed as a corrective term added to the output noise that is proportional to the squared gradient obtained from the posterior of the Gaussian Process. An iterative procedure is given so that the obtained solution converges to the best estimation of the coefficient functions according to the given measure of fitness. Moreover, the expectation and covariance functions estimated by GP are modified for the noisy scheduling variable case to include the noise contribution on the estimated expectation and covariance functions. The model training procedure identifies noise level in the measurements including outputs and scheduling variables by estimating the noise variances, as well as other defined hyperparameters. Finally, the performance of the proposed method is compared to the standard GP approach through a numerical example.",
author = "Farshid Abbasi and J. Mohammadpour and Roland Toth and N. Meskin",
year = "2015",
doi = "10.1109/CDC.2015.7402326",
language = "English",
isbn = "978-1-4799-7884-7",
pages = "789--794",
booktitle = "2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan",
publisher = "Institute of Electrical and Electronics Engineers",
address = "United States",

}

Abbasi, F, Mohammadpour, J, Toth, R & Meskin, N 2015, A Bayesian approach for model identification of LPV systems with uncertain scheduling variables. in 2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan., 7402326, Institute of Electrical and Electronics Engineers, Piscataway, pp. 789-794, 54th IEEE Conference on Decision and Control (CDC 2015), Osaka, Japan, 15/12/15. DOI: 10.1109/CDC.2015.7402326

A Bayesian approach for model identification of LPV systems with uncertain scheduling variables. / Abbasi, Farshid; Mohammadpour, J.; Toth, Roland; Meskin, N.

2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan. Piscataway : Institute of Electrical and Electronics Engineers, 2015. p. 789-794 7402326.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - A Bayesian approach for model identification of LPV systems with uncertain scheduling variables

AU - Abbasi,Farshid

AU - Mohammadpour,J.

AU - Toth,Roland

AU - Meskin,N.

PY - 2015

Y1 - 2015

N2 - This paper presents a Gaussian Process (GP) based Bayesian method that takes into account the effect of additive noise on the scheduling variables for identification of linear parameter-varying (LPV) models in input-output form. The proposed method approximates the noise-free coefficient functions by a local linear expansion on the observed scheduling variables. Therefore, additive noise on the scheduling variables is reconstructed as a corrective term added to the output noise that is proportional to the squared gradient obtained from the posterior of the Gaussian Process. An iterative procedure is given so that the obtained solution converges to the best estimation of the coefficient functions according to the given measure of fitness. Moreover, the expectation and covariance functions estimated by GP are modified for the noisy scheduling variable case to include the noise contribution on the estimated expectation and covariance functions. The model training procedure identifies noise level in the measurements including outputs and scheduling variables by estimating the noise variances, as well as other defined hyperparameters. Finally, the performance of the proposed method is compared to the standard GP approach through a numerical example.

AB - This paper presents a Gaussian Process (GP) based Bayesian method that takes into account the effect of additive noise on the scheduling variables for identification of linear parameter-varying (LPV) models in input-output form. The proposed method approximates the noise-free coefficient functions by a local linear expansion on the observed scheduling variables. Therefore, additive noise on the scheduling variables is reconstructed as a corrective term added to the output noise that is proportional to the squared gradient obtained from the posterior of the Gaussian Process. An iterative procedure is given so that the obtained solution converges to the best estimation of the coefficient functions according to the given measure of fitness. Moreover, the expectation and covariance functions estimated by GP are modified for the noisy scheduling variable case to include the noise contribution on the estimated expectation and covariance functions. The model training procedure identifies noise level in the measurements including outputs and scheduling variables by estimating the noise variances, as well as other defined hyperparameters. Finally, the performance of the proposed method is compared to the standard GP approach through a numerical example.

UR - http://www.scopus.com/inward/record.url?scp=84962034503&partnerID=8YFLogxK

U2 - 10.1109/CDC.2015.7402326

DO - 10.1109/CDC.2015.7402326

M3 - Conference contribution

SN - 978-1-4799-7884-7

SP - 789

EP - 794

BT - 2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

ER -

Abbasi F, Mohammadpour J, Toth R, Meskin N. A Bayesian approach for model identification of LPV systems with uncertain scheduling variables. In 2015 54th IEEE Conference on Decision and Control (CDC), 15-18 December 2015, Osaka, Japan. Piscataway: Institute of Electrical and Electronics Engineers. 2015. p. 789-794. 7402326. Available from, DOI: 10.1109/CDC.2015.7402326