Finite Model Order Optimal Input Design for Minimum Variance Control
2009 (English)In: European Control Conference, 2009, 454-459 p.Conference paper (Refereed)
It is well known that if we intend to use aminimum variance controller to stabilize a minimum phaseplant (with exactly one time delay), which is designed basedon a model obtained from an identification experiment, thebest experiment which can be performed on the system todetermine such a model (subject to output power constraints, orfor some specific model structures) is to use the true minimumvariance controller. This result has been derived under severalcircumstances, first using asymptotic (in model order) varianceexpressions but also more recently for ARMAX models offinite order. In this paper we re-approach this problem byusing a recently developed geometric approach to varianceanalysis, which is non asymptotic in model order, with whichwe generalize some of the previous results established in theliterature. We also believe the new derivations to be moretransparent than earlier contributions.
Place, publisher, year, edition, pages
2009. 454-459 p.
Research subject SRA - ICT
IdentifiersURN: urn:nbn:se:kth:diva-73588ScopusID: 2-s2.0-84955208693ISBN: 978-3-9524173-9-3OAI: oai:DiVA.org:kth-73588DiVA: diva2:488993
European Control Conference, ECC 09. Budapest, Hungary. 23-26 August 2009
QC 201204262012-02-022012-02-022015-05-27Bibliographically approved