简介：Themainfocusisnonlinearmodel-baseddynamicpositioning(DP)controlsystemdesign.Anonlinearuniformglobalexponentialstability(UGES)observerproducesnoise-freeestimatesoftheposition,theslowlyvaryingenvironmentaldisturbancesandthevelocity,whichareusedinaproportional-derivative(PD)+feedforwardcontrollaw.Thestabilityofthisobserver-controllersystemisprovedbyintroducingaspecificnonlinearcascadedsystem.ThesimulationresultshavesuccessfullydemonstratedtheperformanceofdesignedDPcontrolsystem.

简介：Inthispaper,anAntColonySystem(AC)basedheuristicalgorithmispresentedtofindthemulti-constraineddynamicQualityofService(QoS)multicastrouting.WealsoexplorethescalabilityoftheACalgorithmandmulticasttreebyusing"Pull"modeinsteadof"Push"mode,andtheimprovementonthetimecomplexityofACalgorithmbyusinganewdatastructure,I.e.,apointerarrayinsteadoftheprevious"matrix"structure.Ourextensivetestsshowthatthepresentedalgorithmcanfindtheglobaloptimumorsuboptimum,andhasagoodscalabilitywithdynamicadaptationtothechangeofmulticastgroup,andgivesbetterperformanceintermsofthetotalcostthanothertwoalgorithms.

简介：Accordingtothecharacteristicofthesensorinertia,thedynamicpredictiontoimprovethesystemdynamicprecisionispresentedinthispaper.Withtherecurrencecalculationoftimeconstantofthesensor,thesystemdynamicprecisionisgreatlyimproved.Theexampleusingthismethodisgiven.

简介：新思科技日前宣布,Graphcore采用新思科技DesignPlatform成功设计其ColossusTM智能处理单元(IPU),与现有处理器(CPU和GPU)相比,加速了人工智能(AI)计算。新思科技全面的数字和定制设计流程通过关键的人工智能优化技术,使设计者能够在应用于云计算、数据中心、汽车和移动应用的AI芯片上,提供同类最佳的设计实现质量(QoR)和最短的设计收敛时间(TTR)。新思科技Fusion技术TM增强了AI芯片设计能力,包括互连规划、乘法累加(MAC)拓扑优化和完整的AIIP参考流程,以实现最快速度、最小面积、最低功耗及实现三者的最佳平衡。

简介：ＤｙｎａｍｉｃＣｕｒｒｅｎｔＳｈａｒｉｎｇｉｎＰａｒａｌｅｌｉｎｇＰｏｗｅｒＭＯＳＦＥＴ①ＷＡＮＧＢａｏｃｈｅｎｇ，ＷＵＷｅｉｙａｎｇ，ＺＨＡＮＧＣｈｕｎｊｉａｎｇ（ＹａｎｓｈａｎＵｎｉｖｅｒｓｉｔｙ，Ｑｉｎｈｕａｎｇｄａｏ０６６００４，ＣＨＮ）Ａｂ...

简介：遮蔽下载服务，作为脱机下载请求的内容并且在云存储保留它直到用户们检索它的新应用，最近成为了在中国吸引百万个用户的一个趋势。在面临在下载请求和存储资源的限制的生长之间的窘境时，云服务者不得不设计一个有效资源分配计划提高存储的利用以及满足象一短下载时间一样的用户需要。当用户搅拌机行为被看作一个Markov链过程时，下载速度的合适的分配能优化存储资源利用，这被发现。因此，二动态资源分配策划包括切换的速度(SS)增加的计划和速度(SI)策划被建议。理论分析和模拟结果证明我们的计划能有效地减少存储资源的消费并且为好用户经验使下载时间足够短。

简介：Ahybridapproachforfuzzysystemdesignbasedonclusteringandakindofneurofuzzynetworksisproposed.Anunsupervisedclusteringtechniqueisfirstlyusedtodeterminethenumberofif-thenfuzzyrulesandgenerateaninitialfuzzyrulebasefromthegiveninput-outputdata.Then,aclassofneurofuzzynetworksisconstructedanditsweightsaretunedsothattheobtainedfuzzyrulebasehasahighaccuracy.Finally,twoexamplesoffunctionapproximationproblemsaregiventoillustratetheeffectivenessoftheproposedapproach.

简介：ＡｎＩｎｔｅｌｌｉｇｅｎｔＤｅｓｉｇｎｏｆＥｌｅｃｔｒｏｎＬｅｎｓＳｙｓｔｅｍ￥ＬｅｔＷｅｉ；ＴｏｎｇＬｉｎｓｕ＆ＷｕＳｈｅｎｇｌｉ（ＥｌｅｃｔｒｏｎｉｃＥｎｇｉｎｅｅｒｉｎｇＤｅｐａｒｔｍｅｎｔｏｆＳｏｕｔｈｅａｓｔＵｎｉｖｅｒｓｉｔｙ，Ｎａｍｉｎ...

简介：这篇论文为雪崩光电二极管(APD)建议一个控制方法有温度的反向的偏爱赔偿和负担抵抗赔偿。APD察觉电路上的背景光和负载抵抗的影响详细被分析。温度赔偿和负担抵抗赔偿的一个理论模型被建立，它被用于偏导电路设计的APD。这个控制方法对激光范围查找者在车辆上使用了的LD特别合适，这被预言。试验性的结果证实设计在这建议纸能可观地改进范围查找者的表演。CLC数字TN364+.2

简介：Thispaperoptimizestheburiedchannelcharge-coupleddevice(BCCD)structurefabricatedbycomplementarymetaloxidesemiconductor(CMOS)technology.TheoptimizedBCCDhasadvantagesoflownoise,highintegrationandhighimagequality.Thechargetransferprocessshowsthatinterfacetraps,weakfringingfieldsandpotentialwellbetweenadjacentgatesallcausethedecreaseofchargetransferefficiency(CTE).CTEandwellcapacityaresimulatedwithdifferentoperatingvoltagesandgapsizes.CTEcanachieve99.999%andthewellcapacityreachesupto25000electronsforthegapsizeof130nmandthemaximumoperatingvoltageof3V.

简介：Animprovedstructureoflineartransconductorispresentedinthispaper.ItisanalyzedintheoryandsimulatedwithSpectrebasedon0.25μmCMOSprocess.Thesimulationresultsshowthatthedifferentialinputvoltageoftheproposedtransconductoris4.0Vpp(peaktopeak),whereasthedifferentialinputvoltageoftheexistingsourcedegenerationstructureis2.2Vpp,whentheirnonlinearerrorsarerequiredtobelessthan0.15%.

简介：Studiedarethecontrollerdesignandbasicprinciplesofintelligentlightingnetwork.TI’sMSP430F123isusedasamaincontroller.ByusingtheZigBeemodules(Xbee/Xbee-PRO)andtheGSMmodule(SIM300C)forwirelesscommunications,thelightingcontrolisenabledtoaccesswirelessnetwork.Thissystemusesamobilephonetoachievelighton-offdirectly,whichcanaccomplishwirelesscontrolofintelligentlightinginfamilies.

简介：

简介：Tomeetthechallengeofimplementingrapidlyadvanced,time-consumingmedicalimageprocessingalgorithms,itisnecessarytodevelopamedicalimageprocessingtechnologytoprocessa2Dor3Dmedicalimagedynamicallyontheweb.Butinapremiersystem,onlystaticimageprocessingcanbeprovidedwiththelimitationofwebtechnology.ThedevelopmentofJavaandCORBA(commonobjectrequestbrokerarchitecture)overcomestheshortcomingofthewebstaticapplicationandmakesthedynamicprocessingofmedicalimagesonthewebavailable.Todevelopanopensolutionofdistributedcomputing,weintegratetheJava,andwebwiththeCORBAandpresentaweb-basedmedicalimagedynamicprocessingmethed,whichadoptsJavatechnologyasthelanguagetoprogramapplicationandcomponentsofthewebandutiliestheCORBAarchitecturetocopewithheterogeneouspropertyofacomplexdistributedsystem.Themethodalsoprovidesaplatform-independent,transparentprocessingarchitecturetoimplementtheadvancedimageroutinesandenableuserstoaccesslargedatasetandresourcesaccordingtotherequirementsofmedicalapplications.TheexperimentinthispapershowsthatthemedicalimagedynamicprocessingmethodimplementedonthewebbyusingJavaandtheCORBAisfeasible.

简介：Theproblemofrecognizingnaturalscenes,suchaswater,smoke,fire,wind-blownvegetationandaflockofflyingbirds,isconsidered.Thesescenesexhibitthecharacteristicdynamicpattern,buthavestochasticextent.Theyarereferredtoasdynamictexture(DT).Inreality,thediversityofDTsondifferentviewpointsandscalesareverycommon,whichalsobringgreatdifficultytorecognizeDTs.Inthepreviousstudies,duetonoconsideringofthedeformableandtransientnatureofelementsinDT,themotionestimationmethodisbasedonbrightnessconstancyassumption,whichseeminappropriateforaggregateandcomplexmotions.Anovelmotionmodelbasedonrelativemotionintheneighborhoodoftwo-dimensionalmotionfieldsisproposed.Theestimationofnon-rigidmotionofDTsisbasedonthecontinuityequation,andthenthelocalvectordifference(LVD)isproposedtocharacterizeDTlocalrelativemotion.SpatiotemporalstatisticsoftheLVDsisusedastherepresentationofDTsequences.ExcellentperformancesofclassifyingallDTsinUCLAdatabasedemonstratethecapabilityoftheproposedmethodindescribingDT.

简介：Unmannedaerialvehicles(UAVs)mayplayanimportantroleindatacollectionandoffloadinginvastareasdeployingwirelesssensornetworks,andtheUAV’sactionstrategyhasavitalinfluenceonachievingapplicabilityandcomputationalcomplexity.Dynamicprogramming(DP)hasagoodapplicationinthepathplanningofUAV,butthereareproblemsintheapplicabilityofspecialterrainenvironmentandthecomplexityofthealgorithm.BasedontheanalysisofDP,thispaperproposesahierarchicaldirectionalDP(DDP)algorithmbasedondirectiondeterminationandhierarchicalmodel.WecompareourmethodswithQ-learningandDPalgorithmbyexperiments,andtheresultsshowthatourmethodcanimprovetheterrainapplicability,meanwhilegreatlyreducethecomputationalcomplexity.

简介：

简介：AkindofDynamicFullBandwidthUtilized(DFBU)allocationalgorithmisintroduced.Thisalgorithmallowsasinglelinktousebandwidthfarbeyonditsfairsharebandwidthinamulti-servicepackettransportingsystem.Threeimportantparametersastheboundonmaximumandminimumbandwidth,themaximumpacketdelayandtheminimumbandwidthutilizationarediscussedandanalyzed.ResultsofexperimentsshowthattheDFBU-algorithmiscapableofmakingasinglelinkinthesystemtouseallthesparebandwidth(uptofull-bandwidth)whiletheperformanceoffairnessandQoSrequirementisstillguaranteed.

简介：Aimingatthetraditionalpassivedeceptionmodels,thispaperconstructsaDecoyPlatformbasedonIntelligentAgent(DPIA)torealizedynamicdefense.Thepaperexploresanewdynamicdefensemodelbasedonactivedeception,introducesitsarchitecture,andexpatiatesoncommunicationmethodsandsecurityguaranteeininformationtransference.SimulationresultsshowthattheDPIAcanattracthackeragilityandactivity,leadabnormaltrafficintoit,distributealargenumberofattackdata,andensurerealnetworksecurity.

简介：Astate-dependentroutingalgorithmbasedontheneuralnetworkmodel,whichtakesadvantageofotherdynamicroutingalgorithmforcircuit-switchednetwork,isgivenin[1].ButtheAlgorthmin[1]isacentralizedcontrolmodelwithcomplexO(N^7),therefore,isdiffculttorealizebyhardware.Asimplifiedalgorithmisputforwardinthispaper,inwhichroutingcanbecontrolleddecentralizekly,anditscomplexityisreducedtoO(10N^3).Computersimulationsaremadeinafullyconnectedtestnetworkwitheightnodes.TheresultsshowthatthecentralizedcontrolmodelhasveryeffectiveperformancethatcanmatchRTNR,andthecentralizedcontrolmodelisnotasgoodasthecentralizedonebutbetterthanDAR-1.