A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation

Amir Valibeygi, M. Hadi Balaghi I., Krishna Vijayaraghavan

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

1 Citation (Scopus)

Abstract

A recently developed nonlinear H observer and Extended Kalman Filter (EKF) offer two filters for state estimation in nonlinear systems. The Riccati equation that arises while developing the nonlinear H observer is compared with the Riccati equation arising from the Extended Kalman Filter (EKF). Variations between the two Riccati equations translate into the differences in the performance of these alternative estimation methods. The H filter offers faster convergence of the estimation covariance at large estimation errors during the transience of the filter. The Extended Kalman Filter, on the other hand, maintains higher levels of optimality at steady state at the expense of higher computational load. An LMI formulation for the H filter is also presented that allows leveraging the bound on the nonlinearity to seek a stable filter for nonlinear systems.

Original languageEnglish
Title of host publication2017 American Control Conference, ACC 2017
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Pages5211-5216
Number of pages6
ISBN (Electronic)978-1-5090-5992-8
ISBN (Print)978-1-5090-4583-9
DOIs
Publication statusPublished - 29 Jun 2017
Event2017 American Control Conference, ACC 2017 - Seattle, United States
Duration: 24 May 201726 May 2017

Conference

Conference2017 American Control Conference, ACC 2017
CountryUnited States
CitySeattle
Period24/05/1726/05/17

Fingerprint

Riccati equations
Extended Kalman filters
State estimation
Nonlinear systems
Control nonlinearities
Error analysis

Cite this

Valibeygi, A., Balaghi I., M. H., & Vijayaraghavan, K. (2017). A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation. In 2017 American Control Conference, ACC 2017 (pp. 5211-5216). [7963764] Piscataway: Institute of Electrical and Electronics Engineers. https://doi.org/10.23919/ACC.2017.7963764
Valibeygi, Amir ; Balaghi I., M. Hadi ; Vijayaraghavan, Krishna. / A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation. 2017 American Control Conference, ACC 2017. Piscataway : Institute of Electrical and Electronics Engineers, 2017. pp. 5211-5216
@inproceedings{d060278b61f54554a7f6af884c05a8f9,
title = "A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation",
abstract = "A recently developed nonlinear H∞ observer and Extended Kalman Filter (EKF) offer two filters for state estimation in nonlinear systems. The Riccati equation that arises while developing the nonlinear H∞ observer is compared with the Riccati equation arising from the Extended Kalman Filter (EKF). Variations between the two Riccati equations translate into the differences in the performance of these alternative estimation methods. The H∞ filter offers faster convergence of the estimation covariance at large estimation errors during the transience of the filter. The Extended Kalman Filter, on the other hand, maintains higher levels of optimality at steady state at the expense of higher computational load. An LMI formulation for the H∞ filter is also presented that allows leveraging the bound on the nonlinearity to seek a stable filter for nonlinear systems.",
author = "Amir Valibeygi and {Balaghi I.}, {M. Hadi} and Krishna Vijayaraghavan",
year = "2017",
month = "6",
day = "29",
doi = "10.23919/ACC.2017.7963764",
language = "English",
isbn = "978-1-5090-4583-9",
pages = "5211--5216",
booktitle = "2017 American Control Conference, ACC 2017",
publisher = "Institute of Electrical and Electronics Engineers",
address = "United States",

}

Valibeygi, A, Balaghi I., MH & Vijayaraghavan, K 2017, A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation. in 2017 American Control Conference, ACC 2017., 7963764, Institute of Electrical and Electronics Engineers, Piscataway, pp. 5211-5216, 2017 American Control Conference, ACC 2017, Seattle, United States, 24/05/17. https://doi.org/10.23919/ACC.2017.7963764

A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation. / Valibeygi, Amir; Balaghi I., M. Hadi; Vijayaraghavan, Krishna.

2017 American Control Conference, ACC 2017. Piscataway : Institute of Electrical and Electronics Engineers, 2017. p. 5211-5216 7963764.

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

TY - GEN

T1 - A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation

AU - Valibeygi, Amir

AU - Balaghi I., M. Hadi

AU - Vijayaraghavan, Krishna

PY - 2017/6/29

Y1 - 2017/6/29

N2 - A recently developed nonlinear H∞ observer and Extended Kalman Filter (EKF) offer two filters for state estimation in nonlinear systems. The Riccati equation that arises while developing the nonlinear H∞ observer is compared with the Riccati equation arising from the Extended Kalman Filter (EKF). Variations between the two Riccati equations translate into the differences in the performance of these alternative estimation methods. The H∞ filter offers faster convergence of the estimation covariance at large estimation errors during the transience of the filter. The Extended Kalman Filter, on the other hand, maintains higher levels of optimality at steady state at the expense of higher computational load. An LMI formulation for the H∞ filter is also presented that allows leveraging the bound on the nonlinearity to seek a stable filter for nonlinear systems.

AB - A recently developed nonlinear H∞ observer and Extended Kalman Filter (EKF) offer two filters for state estimation in nonlinear systems. The Riccati equation that arises while developing the nonlinear H∞ observer is compared with the Riccati equation arising from the Extended Kalman Filter (EKF). Variations between the two Riccati equations translate into the differences in the performance of these alternative estimation methods. The H∞ filter offers faster convergence of the estimation covariance at large estimation errors during the transience of the filter. The Extended Kalman Filter, on the other hand, maintains higher levels of optimality at steady state at the expense of higher computational load. An LMI formulation for the H∞ filter is also presented that allows leveraging the bound on the nonlinearity to seek a stable filter for nonlinear systems.

UR - http://www.scopus.com/inward/record.url?scp=85027024834&partnerID=8YFLogxK

U2 - 10.23919/ACC.2017.7963764

DO - 10.23919/ACC.2017.7963764

M3 - Conference contribution

AN - SCOPUS:85027024834

SN - 978-1-5090-4583-9

SP - 5211

EP - 5216

BT - 2017 American Control Conference, ACC 2017

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

ER -

Valibeygi A, Balaghi I. MH, Vijayaraghavan K. A comparative study of Extended Kalman Filter and an optimal nonlinear observer for state estimation. In 2017 American Control Conference, ACC 2017. Piscataway: Institute of Electrical and Electronics Engineers. 2017. p. 5211-5216. 7963764 https://doi.org/10.23919/ACC.2017.7963764