简介:100piecesof26650-typeLithiumironphosphate(LiFePO4)batteriescycledwithafixedchargeanddischargeratearetested,andtheinfluenceofthebatteryinternalresistanceandtheinstantaneousvoltagedropatthestartofdischargeonthestateofhealth(SOH)isdiscussed.Abackpropagation(BP)neuralnetworkmodelusingadditionalmomentumisbuiltuptoestimatethestateofhealthofLi-ionbatteries.Theadditional10piecesareusedtoverifythefeasibilityoftheproposedmethod.Theresultsshowthattheneuralnetworkpredictionmodelhaveahigheraccuracyandcanbeembeddedintobatterymanagementsystem(BMS)toestimateSOHofLiFePO4Li-ionbatteries.
简介:Handwrittensignaturerecognitionispresentedbasedonananglefeaturevectorbyusingtheartificialneuralnetwork(ANN)inthisresearch.Eachsignatureimagewillberepresentedbyananglevector.ThefeaturevectorwillconstitutetheinputtotheANN.Thecollectionofsignatureimagesisdividedintotwosets.OnesetwillbeusedfortrainingtheANNinasupervisedfashion.TheothersetwhichisneverseenbytheANNwillbeusedfortesting.Aftertraining,theANNwillbetestedbyrecognizingthesignatures.Whenasignatureisclassifiedcorrectly,itisconsideredcorrectrecognition,otherwiseitisafailure.Theachievedrecognitionrateofthissystemis94%.
简介:Inthepresentstudy,artificialneuralnetwork(ANN)approachwasusedtopredictthestress–straincurveofnearbetatitaniumalloyasafunctionofvolumefractionsofaandb.Thisapproachistodevelopthebestpossiblecombinationorneuralnetwork(NN)topredictthestress–straincurve.Inordertoachievethis,threedifferentNNarchitectures(feed-forwardback-propagationnetwork,cascade-forwardback-propagationnetwork,andlayerrecurrentnetwork),threedifferenttransferfunctions(purelin,Log-Sigmoid,andTan-Sigmoid),numberofhiddenlayers(1and2),numberofneuronsinthehiddenlayer(s),anddifferenttrainingalgorithmswereemployed.ANNtrainingmodules,theloadintermsofstrain,andvolumefractionofaaretheinputsandthestressasanoutput.ANNsystemwastrainedusingthepreparedtrainingset(a,16%a,40%a,andbstress–straincurves).Aftertrainingprocess,testdatawereusedtochecksystemaccuracy.Itisobservedthatfeed-forwardback-propagationnetworkisthefastest,andLog-Sigmoidtransferfunctionisgivingthebestresults.Finally,layerrecurrentNNwithasinglehiddenlayerconsistsof11neurons,andLog-Sigmoidtransferfunctionusingtrainlmastrainingalgorithmisgivinggoodresult,andaveragerelativeerroris1.27±1.45%.Intwohiddenlayers,layerrecurrentNNconsistsof7neuronsineachhiddenlayerwithtrainrpasthetrainingalgorithmhavingthetransferfunctionofLogSigmoidwhichgivesbetterresults.Asaresult,theNNisfoundedsuccessfulforthepredictionofstress–straincurveofnearbtitaniumalloy.
简介:Inthispaper,weproposetwoweightedlearningmethodsfortheconstructionofsinglehiddenlayerfeedforwardneuralnetworks.Bothmethodsincorporateweightedleastsquares.Ourideaistoallowthetraininginstancesnearertothequerytoofferbiggercontributionstotheestimatedoutput.Byminimizingtheweightedmeansquareerrorfunction,optimalnetworkscanbeobtained.Theresultsofanumberofexperimentsdemonstratetheeffectivenessofourproposedmethods.
简介:Neurodegenerativedisordersaffectmorethan30millionindividualsthroughouttheworldandleadtosignificantdisabilityaswellasdeath.Thesestatisticswillincreasealmostexponentiallyasthelifespanandageofindividualsincreasegloballyandindividualsbecomemoresusceptibletoacutedisorderssuchasstrokeaswellaschronicdiseasesthatinvolvecognitiveloss,Alzheimer’sdisease,andParkinson’sdisease.Currenttherapiesforsuchdisordersareeffectiveonlyforasmallsubsetofindividualsorprovidesymptomaticreliefbutdonotalterdiseaseprogression.Oneexcitingtherapeuticapproachthatmayturnthetideforaddressingneurodegenerativedisordersinvolvesthemammaliantargetofrapamycin(mTOR).mTORisacomponentoftheproteincomplexesmTORComplex1(mTORC1)andmTORComplex2(mTORC2)thatareubiquitousthroughoutthebodyandcontrolmultiplefunctionssuchasgenetranscription,metabolism,cellsurvival,andcellsenescence.mTORthroughitsrelationshipwithphosphoinositide3-kinase(PI3-K)andproteinkinaseB(Akt)andmultipledownstreamsignalingpathwayssuchasp70ribosomalS6kinase(p70S6K)andprolinerichAktsubstrate40kDa(PRAS40)promotesneuronalcellregenerationthroughstemcellrenewalandoverseescriticalpathwayssuchasapoptosis,autophagy,andnecroptosistofosterprotectionagainstneurodegenerativedisorders.TargetingbymTORofspecificpathwaysthatdrivelong-termpotentiation,synapticplasticity,andβ-amyloidtoxicitymayoffernewstrategiesfordisorderssuchasstrokeandAlzheimer’sdisease.Overall,mTORisanessentialneuroprotectivepathwaybutmustbecarefullytargetedtomaximizeclinicalefficacyandeliminateanyclinicaltoxicsideeffects.
简介:Vehielesenlistedwitheomputing,sensingandcommunicatingdevicescancreatevehicularnetworks,asubsetofcooperativesystemsinheterogeneousenvironments,aimingatimprovingsafetyandentertainmentintraffic.Invehicularnetworks,avehicle’sidentityisassociatedtoitsowner’sidentityasauniquelinkage.Therefore,itisofimportancetoprotectprivacyofvehiclesfrombeingpossiblytracked.Obviously,theprivacyprotectionmustbescalablebecauseofthehighmobilityandlargepopulationofvehicles.Inthiswork,wetakeanon-trivialsteptowardsprotectingprivacyofvehicles.Asprivacydrawspublicconcerns,wefirstlypresentprivacyimplicationsofoperationalchallengesfromthepublicpolicyperspective.Additionally,weenvisionvehicularnetworksasgeographicallypartitionedsubnetworks(cells).Eachsubnetworkmaintainsalistofpseudonyms.Eachpseudonymincludesthecell’sgeographicidandarandomnumberashostid.Beforestartingcommunication,vehiclesneedtorequestapseudonymondemandfrompseudonymserver.Inordertoimproveutilizationofpseudonyms,weaddressastochasticmodelwithtime-varyingarrivalanddeparturerates.Ourmaincontributionincludes:1)proposingascalableandeffectivealgorithmtoprotectprivacy;2)providinganalyticalresultsofprobability,varianceandexpectednumberofrequestsonpseudonymservers.Theempiricalresultsconfirmtheaccuracyofouranalyticalpredictions.
简介:Networktrafficclassificationaimsatidentifyingtheapplicationtypesofnetworkpackets.ItisimportantforInternetserviceproviders(ISPs)tomanagebandwidthresourcesandensurethequalityofservicefordifferentnetworkapplications.However,mostclassificationtechniquesusingmachinelearningonlyfocusonhighflowaccuracyandignorebyteaccuracy.TheclassifierwouldobtainlowclassificationperformanceforelephantflowsastheimbalancebetweenelephantflowsandmiceflowsonInternet.Theelephantflows,however,consumemuchmorebandwidththanmiceflows.Whentheclassifierisdeployedfortrafficpolicing,thenetworkmanagementsystemcannotpenalizeelephantflowsandavoidnetworkcongestioneffectively.Thisarticleexploresthefactorsrelatedtolowbyteaccuracy,andsecondly,itpresentsanewtrafficclassificationmethodtoimprovebyteaccuracyattheaidofdatacleaning.Experimentsarecarriedoutonthreegroupsofreal-worldtrafficdatasets,andthemethodiscomparedwithexistingworkontheperformanceofimprovingbyteaccuracy.Experimentshowsthatbyteaccuracyincreasedbyabout22.31%onaverage.Themethodoutperformstheexistingoneinmostcases.
简介:Theadventofthetimeofbigdataalongwithsocialnetworksmakesthevisualizationandanalysisofnetworksinformationbecomeincreasinglyimportantinmanyfields.Basedontheinformationfromsocialnetworks,theideaofinformationvisualizationanddevelopmentoftoolsarepresented.Popularsocialnetworkmicro-blog(‘Weibo’)ischosentorealizetheprocessofusers’interestandcommunicationsdataanalysis.Userinterestvisualizationmethodsarediscussedandchosenandprogramsaredevelopedtocollectusers’interestanddescribeitbygraph.Thevisualizationresultsmaybeusedtoprovidethecommercialrecommendationorsocialinvestigationapplicationfordecisionmakers.
简介:Intheenvironmentofheterogeneouswirelessnetworks,itisvitaltoselectacurrentlyoptimalnetworkforapplicationsandsubscribers.Theuseofmultipleattributedecisionmaking(MADM)forheterogeneousnetworkselectioncanprovidesubscriberswithsatisfactoryservicequality.ConvertingheterogeneousnetworkselectionintoaMADMproblem,theauthorspresentanimprovedalgorithmforMADMbasedongroupdecisiontheory.Thealgorithmcombinesweightvectorsofmultipleattributedecisionmakingtoobtainacombinationalweightvector.Thentheresults’compatibilitywillbeassessed.Iftheydonotmeettherequirementsofcompatibility,thejudgmentmatrixwillbemodifieduntilacomprehensivevectorthatsatisfiescompatibilityrequirementsisproduced.Thevectoriscombinedwithsimpleweightingmethod(SAW)fornetworkselection.Simulationshowsthatthealgorithmcanprovideuserswithsatisfactoryqualityofservice(QoS).
简介:Aseachtypeofsatellitenetworkhasdifferentlinkfeatures,itsdatatransmissionmustbedesignedbasedonitslinkfeaturestoimprovetheefficiencyofdatatransferring.Thetransmissionofnavigationintegratedservicesinformation(NISI)inaglobalnavigationsatellitesystem(GNSS)withinter-satellitelinks(ISLs)isstudiedbytakingtherealsituationofinter-satellitecommunicationlinksintoaccount.Anon-demandcomputingandbufferingcentralizedroutestrategyisproposedbasedondynamicgroupingandthetopologyevolutionlawoftheGNSSnetworkwithinwhichthesatellitenodesareoperatedinthemannerofdynamicgrouping.Dynamicgroupingisbasedonsatellitesspatialrelationshipsandthegrouproleofthesatellitenodechangesbyturnsduetoitsspatialrelationships.Theroutestrategyprovidessignificantadvantagesofhighefficiency,lowcomplexity,andflexibleconfiguration,bywhichtheestablishedGNSScanpossessthefeaturesandcapabilitiesoffeasibledeployment,efficienttransmission,convenientmanagement,structuralinvulnerabilityandflexibleexpansion.
简介:Modulatingboththeclockfrequencyandsupplyvoltageofthenetwork-on-chip(NoC)duringruntimecanreducethepowerconsumptionandheatflux,butwillleadtotheincreaseofthelatencyofNoC.Itisnecessarytofindatradeoffbetweenpowerconsumptionandcommunicationlatency.Soweproposeananalyticallatencymodelwhichcanshowustherelationshipofthem.TheproposedmodeltoanalyzelatencyisbasedontheM/G/1queuingmodel,whichissuitablefordynamicfrequencyscaling.Theexperimentresultsshowthattheaccuracyofthismodelismorethan90%.