标签: neural
相关资源
-
NeuralControlofHighPerformanceDrives-AnApplicationtothePMSynchronousMotorDrive
-
NeuralAidedDiscretePIDactiveControllerforNon-LinearHystereticBase-Isolationbuilding
-
ExpertSystem,FuzzyLogic,andNeuralNetworkApplicationsinPowerElectronicsandMotionControl(BimalK.Bose)
-
DEVELOPMENTOFPIDNEURALNETWORKCONTROLSYSTEMWITHVIRTUALINSTRUMENT
-
DesignofNeuralNetworksBasedFRITPIDControllersandItsApplications
-
Designandanalysisofneural／fuzzyvariablestructuralPIDcontrolsystems
-
BPneuralnetworktunedPIDcontrollerforpositiontrackingofapneumaticartificialmuscle
-
AnovelalgorithmforwaveletneuralnetworkswithapplicationtoenhancedPIDcontrollerdesign
-
Amodifiedneuralnetworkcontrollerconfiguration
-
FPGA-basedElmanneuralnetworkcontrolsystemforlinearultrasonicmotor(林法正)
-
ArtificialNeuralNetworks-TheTutorialWithMatlab
-
COMBININGIMAGEENTROPYWITHTHEPULSECOUPLED NEURALNETWORKINEDGEDETECTION 2010
-
AbstractInrecenttwentyyears,thetechniqueoffacedetectionandfacerecognition,asoneoftheimportantresearchareaofcomputervisionandimageunderstanding,attractsmoreandmoreattention.Ingeneral,facedetectioningray-levelstillimagesismoredifficultthanthatincolorimages.Thereforethispaperbrieflysurveysthisareaandindicatessomeissuesforexploration.……
-
TraditionalDSPisbasedonalgorithms,changingdatafromoneformtoanotherthroughstep-bystepprocedures.Mostofthesetechniquesalsoneedparameterstooperate.Forexample:recursivefiltersuserecursioncoefficients,featuredetectioncanbeimplementedbycorrelationandthresholds,animagedisplaydependsonthebrightnessandcontrastsettings,etc.Algorithmsdescribewhatistobedone,whileparametersprovideabenchmarktojudgethedata.Theproperselectionofparametersisoftenmoreimportantthanthealgorithmitself.Neuralnetworkstakethisideatotheextremebyusingverysimplealgorithms,butmanyhighlyoptimizedparameters.Thisisarevolutionarydeparturefromthetraditionalmainstaysofscienceandengineering:mathematicallogicandtheorizingfollowedbyexperimentation.Neuralnetworksreplacetheseproblemsolvingstrategieswithtrial&error,pragmaticsolutions,anda"thisworksbetterthanthat"methodology.ThischapterpresentsavarietyofissuesregardingparameterselectioninbothneuralnetworksandmoretraditionalDSPalgorithms.CHAPTERNeuralNetworks(andmore!)26TraditionalDSPisbasedonalgorithms,changingdatafromoneformtoanotherthroughstep-by-stepprocedures.Mostofthesetechniquesalsoneedparameterstooperate.Forexample:recursivefiltersuserecursioncoefficients,featuredetectioncanbeimplementedbycorrelationandthresholds,animagedisplaydependsonthebrightnessandcontrastsettings,etc.Algorithmsdescribewhatistobedone,whileparametersprovideabenchmarktojudgethedata.Theproperselectionofparametersisoftenmoreimportantthanthealgorithmitself.Neuralnetworkstakethisideatotheextremebyusingverysimplealgorithms,butmanyhighlyoptimizedparameters.Thisisarevolutionarydeparturefromthetraditionalmai……
-
artificial_neural_network_(mat...,ebook-artificial_neural_network_(matlab_toolbox)_[www……
-
AneuralnetworkmodelingofmicrowavecircuitsonPBGstructures……
-
applicationsofneuralnetworksindesignofdigitalfilters55.1APPLICATIONOFNEURALNETWORKINDESIGNOFDIGITALFILTERSDaliWangandAliZilouchianINTRODUCTIONAnyactiononasignalthatmodifiesthespectralcontentofthesignaliscalledfiltering.Thisincludestheenhancementorsuppressionofcertainfeaturesofthesignalandisusuallyachievedbytheuseoflineartimeinvariantsystems.Therearesituationswherethesystemmaychangewithtimeinaparticularmanner;suchsystemsarecalledadaptivefilters.Inthissection,wedescribefixedfiltersonly.Therearetwobroadclassesofdigitalfilters.Thefirstclassiscalledfiniteimpulseresponse(FIR)filters,sincetheirresponsetoanimpulsediesawayinafinitenumberofsamples.FIRfiltersaredevelopedasnon-recursivestructuresandareinherentlysimplertodes……
