Change search
ReferencesLink to record
Permanent link

Direct link
Extended accuracy analysis of a covariance matching approach for identifying errors-in-variables systems
Uppsala Univ, Dept Informat Technol, Div Syst & Control, SE-75105 Uppsala, Sweden..
Uppsala Univ, Dept Stat, SE-75120 Uppsala, Sweden.;BI Norwegian Business Sch, Dept Econ, N-0442 Oslo, Norway..
Karlstad University, Faculty of Technology and Science, Department of Physics and Electrical Engineering. (Matematisk modellering)
2014 (English)In: Automatica, ISSN 0005-1098, E-ISSN 1873-2836, Vol. 50, no 10, 2597-2605 p.Article in journal (Refereed) PublishedText
Abstract [en]

A covariance matching approach for identifying errors-in-variables systems is analyzed for the general case. The asymptotic covariance matrix of the jointly estimated system parameters, noise variances and auxiliary parameters is derived. An algorithm for how to compute this covariance matrix from given system descriptions is also provided. The results generalize previous known special cases. Using Monte Carlo analysis, we illustrate the proposed algorithm. The results suggest close agreement between the theoretical and empirical accuracy. (C) 2014 Elsevier Ltd. All rights reserved.

Place, publisher, year, edition, pages
Pergamon Press, 2014. Vol. 50, no 10, 2597-2605 p.
Keyword [en]
System identification, Errors-in-variables models, Linear systems, Covariance functions, Covariance matrix
National Category
Mathematics
Identifiers
URN: urn:nbn:se:kau:diva-41497DOI: 10.1016/j.automatica.2014.08.020ISI: 000344207300018OAI: oai:DiVA.org:kau-41497DiVA: diva2:923161
Available from: 2016-04-25 Created: 2016-04-11 Last updated: 2016-06-07Bibliographically approved

Open Access in DiVA

No full text

Other links

Publisher's full text

Search in DiVA

By author/editor
Mossberg, Magnus
By organisation
Department of Physics and Electrical Engineering
In the same journal
Automatica
Mathematics

Search outside of DiVA

GoogleGoogle Scholar

Altmetric score

Total: 6 hits
ReferencesLink to record
Permanent link

Direct link