摘要
Basedonhighorderdynamicneuralnetwork,thispaperpresentsthetrackingproblemforuncertainnonlinearcompositesystem,whichcontainsexternaldisturbance,whosenonlinearitiesareassumedtobeunknown.Asmoothcontrollerisdesignedtoguaranteeauniformultimateboundednesspropertyforthetrackingerrorandallothersignalsintheclosedloop.Certainmeasuresareutilizedtotestitsperformance.Noaprioriknowledgeofanupperboundonthe'optimal'weightandmodelingerrorisrequired;theweightsofneuralnetworksareupdatedon-line.Numericalsimulationsperformedonasimpleexampleillustrateandclarifytheapproach.
出版日期
2005年02月12日(中国期刊网平台首次上网日期,不代表论文的发表时间)