Virtual Closed Loop Parameter Estimation: A Review

Dionisio Bernal, Martin Dalgaard Ulriksen

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review


The use of poles as targets for estimating parameters of linear systems is attractive from a variance perspective but the limitations are well-known, namely: 1) the number of poles in the identifiable bandwidth is often smaller than the number of parameters to be identified and 2) the sensitivity of many of the poles can be small. One way to mitigate these limitations, initially put forth about 20 years ago, is to operate in closed-loop. In the closed-loop alternative the number of poles can be increased by aggregating results from multiple gains, and there is the added benefit that some control over the sensitivities can be realized by gain design. The closed-loop strategy has thus far, however, had little practical impact. Presumably because the complexity and cost overhead of the real time operation has been deemed too high for off-line monitoring. It has been recently noted, however, that closed-loop poles for specified gains can be extracted from input-output data by appropriate signal processing and it follows, therefore, that the closed-loop strategy can be virtually implemented. This paper summarizes the theory presented in [10] and illustrates the scheme in a numerical example.
Original languageEnglish
Title of host publicationInternational Conference on Structural Engineering Dynamics (ICEDyn) 2019
Number of pages7
Place of PublicationViana do Castelo, Portugal
Publication date2019
Article number33
Publication statusPublished - 2019
EventInternational Conference on Structural Engineering Dynamics (ICEDyn) 2019 - Viana do Castelo, Portugal
Duration: 24 Jun 201926 Jun 2019


ConferenceInternational Conference on Structural Engineering Dynamics (ICEDyn) 2019
CityViana do Castelo

Bibliographical note

Proceedings published on a USB.


  • Closed-loop eigenstructure
  • Pole sensitivity
  • Parameter estimation

Cite this