简介：Thispaperaddressesanonlinearfeedbackcontrolproblemforthechaoticarchmicro-electro-mechanicalsystemwithunknownparameters,immeasurablestatesandpartialstate-constraintsubjectedtothedistributedelectrostaticactuation.Toreflectinherentpropertiesanddesigncontroller,thephasediagrams,bifurcationdiagramandPoincaresectionarepresentedtoinvestigatethenonlineardynamics.TheauthorsemployasymmetricbarrierLyapunovfunctiontopreventviolationofconstraintwhenthearchmicro-electro-mechanicalsystemfacessomelimits.AnRBFneuralnetworksystemintegratingwithanupdatelawisadoptedtoestimateunknownfunctionwitharbitrarilysmallerror.Toeliminatechaoticoscillation,aneuro-adaptivebacksteppingcontrolschemefusedwithanextendedstatetrackingdifferentiatorandanobserverisconstructedtolowerrequirementsonmeasuredstatesandprecisesystemmodel.Besides,introducinganextendedstatetrackingdifferentiatoravoidsrepeatedderivativeforthevirtualcontrolsignalassociatedwithconventionalbackstepping.Finally,simulationresultsarepresentedtoillustratefeasibilityoftheproposedscheme.