Parameter estimation in non-linear state-space models by automatic differentiation of non-linear kalman filters

dc.contributorAalto-yliopistofi
dc.contributorAalto Universityen
dc.contributor.authorGorad, Ajinkyaen_US
dc.contributor.authorZhao, Zhengen_US
dc.contributor.authorSärkkä, Simoen_US
dc.contributor.departmentDepartment of Electrical Engineering and Automationen
dc.contributor.groupauthorSensor Informatics and Medical Technologyen
dc.date.accessioned2020-12-31T08:41:38Z
dc.date.available2020-12-31T08:41:38Z
dc.date.issued2020-09en_US
dc.description.abstractIn this article, we propose automatic differentiation based methods for parameter estimation in non-linear state-space models. We use extended Kalman filter and cubature Kalman filters for approximating the negative log-likelihood (i.e., the energy function) of the parameter posterior distribution and compute the gradients and Hessians of this function by using automatic differentiation of the filter recursions. The proposed approach enables computing MAP estimates and forming Laplace approximations for the parameter posterior without a need for implementing complicated derivative recursions or manual computation of Jacobians. The methods are demonstrated in parameter estimation problems on a pendulum model and coordinated turn model.en
dc.description.versionPeer revieweden
dc.format.extent6
dc.format.mimetypeapplication/pdfen_US
dc.identifier.citationGorad, A, Zhao, Z & Särkkä, S 2020, Parameter estimation in non-linear state-space models by automatic differentiation of non-linear kalman filters . in Proceedings of the 2020 IEEE 30th International Workshop on Machine Learning for Signal Processing, MLSP 2020 ., 9231844, IEEE International Workshop on Machine Learning for Signal Processing, IEEE, IEEE International Workshop on Machine Learning for Signal Processing, Espoo, Finland, 21/09/2020 . https://doi.org/10.1109/MLSP49062.2020.9231844en
dc.identifier.doi10.1109/MLSP49062.2020.9231844en_US
dc.identifier.isbn9781728166629
dc.identifier.issn2161-0363
dc.identifier.issn2161-0371
dc.identifier.otherPURE UUID: 56583262-88b7-4130-bf3f-fa2ae9df237den_US
dc.identifier.otherPURE ITEMURL: https://research.aalto.fi/en/publications/56583262-88b7-4130-bf3f-fa2ae9df237den_US
dc.identifier.otherPURE LINK: http://www.scopus.com/inward/record.url?scp=85096479735&partnerID=8YFLogxKen_US
dc.identifier.otherPURE FILEURL: https://research.aalto.fi/files/53868447/ELEC_Gorad_etal_Parameter_Estimation_in_Non_Linear_MLSP2020_acceptedauthormanuscript.pdfen_US
dc.identifier.urihttps://aaltodoc.aalto.fi/handle/123456789/101493
dc.identifier.urnURN:NBN:fi:aalto-2020123160314
dc.language.isoenen
dc.publisherIEEE
dc.relation.ispartofIEEE International Workshop on Machine Learning for Signal Processingen
dc.relation.ispartofseriesProceedings of the 2020 IEEE 30th International Workshop on Machine Learning for Signal Processing, MLSP 2020en
dc.relation.ispartofseriesIEEE International Workshop on Machine Learning for Signal Processingen
dc.rightsopenAccessen
dc.subject.keywordAutomatic differentiationen_US
dc.subject.keywordCubature Kalman filteren_US
dc.subject.keywordExtended Kalman filteren_US
dc.subject.keywordNon -linear state space modelen_US
dc.subject.keywordParameter estimationen_US
dc.titleParameter estimation in non-linear state-space models by automatic differentiation of non-linear kalman filtersen
dc.typeA4 Artikkeli konferenssijulkaisussafi
dc.type.versionacceptedVersion

Files