应用信度分配的模糊CMAC实现非线性系统的容错控制

？Theadaptivefault-tolerantcontrolschemeofdynamicnonlinearsystembasedonthecreditassignedfuzzyCMACneuralnetworkispresented.Theproposedlearningapproachusesthelearnedtimesofaddressedhypercubesasthecredibility,theamountsofcorrectingerrorsareproportionaltotheinversionofthelearnedtimesofaddressedhypercubes.Withthisidea,thelearningspeedcanindeedbeimproved.BasedontheimprovedCMAClearningapproachandusingtheslidingcontroltechnique,theeffectivecontrollawreconfigurationstrategyispresented.Thesystemstabilityandperformanceareanalyzedunderfailurescenarios.ThenumericalsimulationdemonstratestheeffectivenessoftheimprovedCMACalgorithmandtheproposedfault-tolerantcontroller.