|
自动化学报 2005
未知输出反馈非线性时滞系统自适应神经网络跟踪控制, PP. 799-803 Keywords: Nonlineartime-delaysystems,neuralnetwork,backstepping Abstract: ?Anadaptiveoutputfeedbackneuralnetworktrackingcontrollerisdesignedforaclassofunknownoutputfeedbacknonlineartime-delaysystemsbyusingbacksteppingtechnique.Neuralnetworksareusedtoapproximateunknowntime-delayfunctions.Delay-dependentfiltersareintroducedforstateestimation.Thedominationmethodisusedtodealwiththesmoothtime-delaybasisfunctions.Theadaptiveboundingtechniqueisemployedtoestimatetheupperboundoftheneuralnetworkreconstructionerror.BasedonLyapunov-Krasoviskiifunctional,thesemi-globaluniformultimateboundedness(SGUUB)ofallthesignalsintheclosed-loopsystemisproved.Thearbitraryoutputtrackingaccuracyisachievedbytuningthedesignparametersandtheneuralnodenumber.Thefeasibilityisinvestigatedbyanillustrativesimulationexample.
|