Adaptive input design in system identification
2005 (English)In: 2005 44TH IEEE CONFERENCE ON DECISION AND CONTROL & EUROPEAN CONTROL CONFERENCE, Seville, 2005, Vol. 2005, 4988-4993 p.Conference paper (Refereed)
Recently there have been significant developments in re-casting experiment design problems in system identification as convex optimization problems. The practical implementation of these methods is hampered by the fact that typically the "data" in the optimization problem depend on the to be identified system. In this contribution we propose an adaptive certainty equivalence solution based on a recursively identified model. The input design is adapted by taking one Newton step using data from the last identified model.
Place, publisher, year, edition, pages
Seville, 2005. Vol. 2005, 4988-4993 p.
, IEEE Conference on Decision and Control, ISSN 0191-2216
Identification (control systems), Mathematical models, Optimization, Problem solving, Newton step, Optimization problem, Re-casting experiments, Adaptive control systems
Research subject SRA - ICT
IdentifiersURN: urn:nbn:se:kth:diva-60545DOI: 10.1109/CDC.2005.1582952ISI: 000240653704138ISBN: 0780395689ISBN: 9780780395688ISBN: 0-7803-9567-0OAI: oai:DiVA.org:kth-60545DiVA: diva2:478685
44th IEEE Conference on Decision Control/European Control Conference (CCD-ECC), Seville, SPAIN, DEC 12-15, 2005
FunderSwedish Research Council, 621-2001-2147
QC 201201192012-01-162012-01-132012-01-19Bibliographically approved