Robust visual tracking for manipulators with unknown intrinsic and extrinsic parameters

在线阅读 下载PDF 导出详情
摘要 Thispaperaddressestherobustvisualtrackingofmulti-featurepointsfora3Dmanipulatorwithunknownintrinsicandextrinsicparametersofthevisionsystem.Thisclassofcontrolsystemsarehighlynonlinearcontrolsystemscharacterizedastime-varyingandstrongcouplinginstatesandunknownparameters.ItisfirstpointedoutthatnotonlyistheJacobianimagematrixnonsingular,butalsoitsminimumsingularvaluehasapositivelimit.Thisprovidesthefoundationofkinematicsanddynamicscontrolofmanipulatorswithvisualfeedback.Second,theEulerangleexpressedrotationtransformationisemployedtoestimateasubspaceoftheparameterspaceofthevisionsystem.Basedonthetworesultsabove,andarbitrarilychosenparametersinthissubspace,thetrackingcontrollersareproposedsothattheimageerrorscanbemadeassmallasdesiredsolongasthecontrolgainisallowedtobelarge.Thecontrollerdoesnotusevisualvelocitytoachievehighandrobustperformancewithlowsamplingrateofthevisionsystem.TheobtainedresultsareprovedbyLyapunovdirectmethod.Experimentsareincludedtodemonstratetheeffectivenessoftheproposedcontroller.
机构地区 不详
出版日期 2007年04月14日(中国期刊网平台首次上网日期,不代表论文的发表时间)
  • 相关文献