Abstract. Bayesian networks for the static as well as for the dynamic case have gained an enormous interest in the research community of machine learning and pattern recognition. Although the parallels between dynamic Bayesian networks and Kalman tikers ar
ApplicationofBayesiancontrollerstodynamicsystems
RainerDeventer1,JoachimDenzler1,andHeinrichNiemann1
FriedrichAlexanderUniversit¨atErlangenN¨urnberg,91058Erlangen,GermanyAbstract.Bayesiannetworksforthestaticaswellasforthedynamiccasehavegainedanenormousinterestintheresearchcommunityofmachinelearningandpatternrecognition.AlthoughtheparallelsbetweendynamicBayesiannetworksandKalman ltersarewell-knownsincemanyyears,Bayesiannetworkshavenotbeenappliedtoproblemsintheareaofadaptivecontrolofdynamicsystems.Inourworkweexploitthewell-knownsimilaritiesbetweenBayesiannetworksandKalman lterstomodelandcontrollineardynamicsystemsusingdynamicBayesiannetworks.Theanalyticalmodelsarecomparedwithmodelsbeingtrainedwithstepandimpulseresponse.Theexperimentsshowthattheanalyticalmodelaswellasthetrainedmodelaresuitableforcontrolpurposes,whichleadstotheideaofselfadaptivecontrollers.
Keywords:DynamicBayesianNetworks,Kalman lter,DynamicSystem,Con-troller
1Introduction
Bayesiannetworks(BN)forthestaticaswellasforthedynamiccasehavegainedanenormousinterestintheresearchcommunityofarti cialintelli-gence,machinelearningandpatternrecognition.Recently,BNhavebeenappliedalsotostaticproblemsinproduction,sinceproductionprocessesbecomemoreandmorecomplexsothatanalyticalmodelingandmanualde-signaretooexpensive.OneexampleforthesuccessfulapplicationofBNtoastaticsysteminproductionarethequalityevaluationandprocessparameterselectioninordertoreachanacceptablequalitylevel[?].
AlthoughtheparallelsbetweenBNandKalman ltersarewell-knownsincemanyyears,BNhavenotbeenappliedtoproblemsintheareaofadap-tivecontrolofdynamicsystems.Adaptivecontrolofdynamicsystemsisonemajorprobleminproductionprocesses.Comparedtoclassicalcontrolmeth-odsBNhavetheadvantagethatthemodel(ofthestaticordynamicsystem)canbetrainedfromexamplesifthemodelisnotavailableinanalyticalform.DuringtrainingmissinginformationcanbehandledwhichmakesBNsupe-riortootherselfadaptivesystemslikearti cialneuralnetworks.Finally,BNcanalsocalculatethemostsuitableinputwhichleadstoadesiredoutputgiventheinformationthatisenteredasevidenceintheBN.
Dynamic Bayesian Network (DBN) with Structure Expectation...Systems, Department of Computer Science, Tsinghua ...application of the SEM to learning Bayesian ...
A simulation example is given for demonstration of the application of the ...controllers together in such a way that the global closed-loop adaptive ...
Application of Dynamic M... 11页 20财富值 Application of Bayesian ... 暂无...on Dynamical Systems and its Application to the Analysis of Dynamic Scenes ...
Adaptive NN Control of Dynamic Systems with Unknown Dynamic Friction In this paper, based on the dynamic LuGre friction model, adaptive NN controllers are ...
An empirical study of dynamic bayesian networks for...Therefore our application deals with functional and...Student Modeling: An Overview of Systems and ...
10页 免费 Learning the Dynamic Bay... 暂无评价 8页 免费 Learning Bayesian networ... 10页 1财富值 PAPER Structure Learning... 暂无评价 12页 免费喜欢...
However, the application of classical controllers is unsatisfactory as HVAC ...Modeling and optimization of HVAC systems using a dynamic neural network. ...
continuous processes coupled with digital controllers...systems modeled by dynamic Bayesian networks in [...2 Section 4 demonstrates the application of the ...
Dynamic Bayesian Networks Xiangyang Li, Member, ...One important application is to design and ...However, the development of such systems faces ...
A new algorithm for dynamic decoupling control of HPMSM using fuzzy controllers...controllers”, The 14th IEEE International Conference on Fuzzy Systems, 25,...
热门文档