www.bewet.fora.pl

loczytzn

An adaptive fuzzy neural control system design

At k = 100 when the closed-loop system output Y (k) and the reference trajectory in Figure 4 Identification of L00 points off the closed-loop system output y (k) (solid line) and reference trajectory ((dashed line) y (k) compared to curve. If offline identification first 500 points, the curve will be more close to these ideals. 8 Conclusion self-learning self-organization of the super-ball coverage of the input space is divided into one area,[link widoczny dla zalogowanych], corresponding to 0, the antecedent of fuzzy rules the Department of conditions. The field of fuzzy sets for the A ', the different categories do not overlap each other over the ball. This fuzzy neural network based clustering is not a BP network, it is not a small error by Iterative methods, there is no local minimal problems. You can manipulate the data directly from the field to learn fuzzy control rules obtained. and the neural network structure and algorithm is not implied value can not be explained,[link widoczny dla zalogowanych], but the obvious display of the network in the form of rules. In particular, additional learning need for additional study, for all the network without re-learning,[link widoczny dla zalogowanych], learning fast. similar to the example paper, an artificial neural network, the number of iterations at least five thousand times more than hundred times to learn this method has been identified only good results. real-time control of the dozens of recognition very good times. References [i] ReilIyD.CooperL, ELbaumCAneuralmodelforcategory [earning.BiolCybern, 1982,[link widoczny dla zalogowanych],45 (I) :35 - 4L [2] Dou Zhen. fuzzy logic control technology and its application, Beijing: Beijing University of Aeronautics Published Society, L995L52 ~ 165 ��continued on page 27 1-way Wushan: large delay system theory and digital simulation DIGITALSIMULATIONANDTHEORETICALANALYSISOFLARGEDELAYSYSTEMOwngWushan (College. fElectronic & ElectricalEngine ~ ng.ShanghaiUnlve ~ ity. fEng] neefinRSci iv Shartghai200065) AbstractThisthesisdescribedthetheoryofthepure-delaysystemandcolvedthecontradictione-tweentheadvancedcomputercontrolsystemanddomesticapparatusesoffirstandsecondgenera-tion,[link widoczny dla zalogowanych], Section l3 andbetweentheadvancedcontroltechnlogyandthetraditionalmodeofmanagementinoldenterprisesKeywordsDigitalsimulationLargedelaysystemAutomaticproportion on the next page ASYSTEMDESIGNFORADAPTIVEFUZZYNEURALCONTRoLXuJingbo ��CollegeofElectronic & ElectricalEnr song ng, s} langhUnivemtyEngineeringScienceShanghai200065) AbstractThispaperhaspresentedamethodofdesigningalearningfuzzylogicreasoningnetworksys-tern. hiscapableofextactingexpertknowledgeautomatically. hneedsfewiterationsinlearningandiseasyforadditionallearning. Thefuzzycontrolrulesareproducedinaself-organizingwayTheeffectivenessisshownbynumericalsimulationresultsKeywordsFuzzyeontro [AdaptivecontrolLearningnetwork

相关的主题文章：

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]