资源描述:
《Parametric Identification of Hammerstein Systems With Consistency Results Using Stochastic Inputs》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、474IEEETRANSACTIONSONAUTOMATICCONTROL,VOL.55,NO.2,FEBRUARY2010ParametricIdentificationofHammersteinSystems[15]–[17],theidentificationproblemofthefollowingHammersteinWithConsistencyResultsUsingStochasticInputssystemisconsidered:Wen-XiaoZhao 111 0 (1) 0Abs
2、tract—Theextendedleastsquaresalgorithmisappliedtoidentifywhere,111aretheunknowncoefficientsofthepolyno-theHammersteinsystem,wherethenonlinearstaticfunction isex-mial .Themethodsusedin[2],[15],[16]areit-pressedasalinearcombinationofbasicfunctionswithunknowncoeffi-erativ
3、e.Letusdenotetheunknownparameters,111,,cients.Strongconsistencyoftheestimatesisestablishedandtheirconver-gentratesareobtainedaswell.111, ,111byforsimplicityofnotations.In[15],forthedataset withfixedsize,theestimateswith IndexTerms—ARMAX,extendedleasts
4、quares(ELS),Hammersteinsomeinitialvalueareobtainedbyminimizingsomecostfunc-system,strongconsistency.tion 0 0I.INTRODUCTION (2)TheHammersteinsystemconsistingofastaticfunction 1fol-lowedbyalineardynamicsystem(seeFig.1)isacommo
5、nlyusedandtheprocess(2)isrepeateduntiltheestimatesconverge.However,modelinengineeringpractice.Therearemanyequipmentsorphe-thisalgorithmmaydivergeifisnotappropriatelychosenasnomenacanbemodeledasHammersteinsystems.Forexample,theshownin[16].Thisdifficultyisovercomein[2]byanormalize
6、dtech-distillationcolumninchemicalengineering[8]andthepowerampli-nique.However,thesealgorithmsarenonrecursive,i.e.,toupdatethefierinelectronicscircuits[12]canbothbemodeledasHammersteinestimateswhilenewdataarriving,onehastocarryoutthewholeiter-systems.Inbiologicalcybernetics,theHammerst
7、einsystemisausefulativeprocedure(2).toolformodelingthedynamicsofthecortexofhumanandanimalsTheinstrumentalvariablemethodisusedin[17]toidentifysystem[11].Soitsidentificationissuenaturallyreceivesmuchattentionfrom(1).Buttheselectionofinstrumentalvariableswithrequiredproper-boththeoristsan
8、dengi