Order and structural dependence selection of LPV-ARX models revisited
2012 (English)In: Decision and Control (CDC), 2012 IEEE 51st Annual Conference on, IEEE , 2012, 6271-6276 p.Conference paper (Refereed)
Accurate parametric identification of Linear Parameter-Varying (LPV) systems requires an optimal prior selection of model order and a set of functional dependencies for the parameterization of the model coefficients. In order to address this problem for linear regression models, a regressor shrinkage method, the Non-Negative Garrote (NNG) approach, has been proposed recently. This approach achieves statistically efficient order and structural coefficient dependence selection using only measured data of the system. However, particular drawbacks of the NNG are that it is not applicable for large-scale over-parameterized problems due to computational limitations and that adequate performance of the estimator requires a relatively large data set compared to the size of the parameterization used in the model. To overcome these limitations, a recently introduced L1 sparse estimator approach, the so-called SPARSEVA method, is extended to the LPV case and its performance is compared to the NNG.
Place, publisher, year, edition, pages
IEEE , 2012. 6271-6276 p.
, IEEE Conference on Decision and Control. Proceedings, ISSN 0191-2216
ARX model, compressive system identification, identification, Linear parameter-varying systems, order selection, sparse estimators
IdentifiersURN: urn:nbn:se:kth:diva-104648DOI: 10.1109/CDC.2012.6426552ScopusID: 2-s2.0-84874235701OAI: oai:DiVA.org:kth-104648DiVA: diva2:565705
51st IEEE Conference on Decision and Control, CDC 2012;Maui, HI;10 December 2012 through 13 December 2012
FunderICT - The Next Generation
QC 201303042012-11-082012-11-082013-04-15Bibliographically approved