Show simple item record

dc.contributor.authorUlbrich, Stefan
dc.contributor.authorRuiz de Angulo García, Vicente
dc.contributor.authorAsfour, Tamim
dc.contributor.authorTorras, Carme
dc.contributor.authorDillmann, Rüdiger
dc.contributor.otherInstitut de Robòtica i Informàtica Industrial
dc.date.accessioned2013-02-12T18:29:11Z
dc.date.available2013-02-12T18:29:11Z
dc.date.created2012
dc.date.issued2012
dc.identifier.citationUlbrich, S. [et al.]. General robot kinematics decomposition without intermediate markers. "IEEE transactions on neural networks", 2012, vol. 23, núm. 4, p. 620-630.
dc.identifier.issn1045-9227
dc.identifier.urihttp://hdl.handle.net/2117/17693
dc.description.abstractThe calibration of serial manipulators with high numbers of degrees of freedom by means of machine learning is a complex and time-consuming task. With the help of a simple strategy, this complexity can be drastically reduced and the speed of the learning procedure can be increased: When the robot is virtually divided into shorter kinematic chains, these subchains can be learned separately and, hence, much more efficiently than the complete kinematics. Such decompositions, however, require either the possibility to capture the poses of all endeffectors of all subchains at the same time, or they are limited to robots that fulfill special constraints. In this work, an alternative decomposition is presented that does not suffer from these limitations. An offline training algorithm is provided in which the composite subchains are learned sequentially with dedicated movements. A second training scheme is provided to train composite chains simultaneously and online. Both schemes can be used together with many machine learning algorithms. In the simulations, an algorithm using Parameterized Self-Organizing Maps (PSOM) modified for online learning and Gaussian Mixture Models (GMM) were chosen to show the correctness of the approach. The experimental results show that, using a two-fold decomposition, the number of samples required to reach a given precision
dc.format.extent11 p.
dc.language.isoeng
dc.rightsAttribution-NonCommercial-NoDerivs 3.0 Spain
dc.rights.urihttp://creativecommons.org/licenses/by-nc-nd/3.0/es/
dc.subjectÀrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic
dc.subject.lcshMachine learning
dc.subject.otherlearning (artificial intelligence) robot kinematics robots PARAULES AUTOR: KB-maps
dc.titleGeneral robot kinematics decomposition without intermediate markers
dc.typeArticle
dc.subject.lemacAprenentatge automàtic
dc.contributor.groupUniversitat Politècnica de Catalunya. ROBiri - Grup de Robòtica de l'IRI
dc.identifier.doi10.1109/TNNLS.2012.2183886
dc.description.peerreviewedPeer Reviewed
dc.subject.inspecClassificació INSPEC::Cybernetics::Artificial intelligence::Learning (artificial intelligence)
dc.relation.publisherversionhttp://dx.doi.org/10.1109/TNNLS.2012.2183886
dc.rights.accessOpen Access
local.identifier.drac9955383
dc.description.versionPostprint (author’s final draft)
dc.relation.projectidinfo:eu-repo/grantAgreement/EC/FP7/270273/EU/Robots Bootstrapped through Learning from Experience/XPERIENCE
dc.relation.projectidinfo:eu-repo/grantAgreement/EC/FP7/247947/EU/Gardening with a Cognitive System/GARNICS
local.citation.authorUlbrich, S.; Ruiz De Angulo, V.; Asfour, T.; Torras, C.; Dillmann, R.
local.citation.publicationNameIEEE transactions on neural networks
local.citation.volume23
local.citation.number4
local.citation.startingPage620
local.citation.endingPage630


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record