簡介：JOURNALOFMATERIALSPROCESSINGTECHNOLOGY142200320–28DELTAFERRITEPREDICTIONINSTAINLESSSTEELWELDSUSINGNEURALNETWORKANALYSISANDCOMPARISONWITHOTHERPREDICTIONMETHODSMVASUDEVANA,?,AKBHADURIA,BALDEVRAJA,KPRASADRAOBAMETALLURGYANDMATERIALSGROUP,INDIRAGANDHICENTREFORATOMICRESEARCH,KALPAKKAM,INDIABDEPARTMENTOFMETALLURGY,INDIANINSTITUTEOFTECHNOLOGY,CHENNAI,INDIARECEIVED2MAY2002RECEIVEDINREVISEDFORM11DECEMBER2002ACCEPTED17FEBRUARY2003ABSTRACTTHEABILITYTOPREDICTTHEDELTAFERRITECONTENTINSTAINLESSSTEELWELDSISIMPORTANTFORMANYREASONSDEPENDINGONTHESERVICEREQUIREMENT,MANUFACTURERSANDCONSUMERSOFTENSPECIFYDELTAFERRITECONTENTASANALLOYSPECIFICATIONTOENSURETHATWELDCONTAINSADESIREDMINIMUMORMAXIMUMFERRITELEVELRECENTRESEARCHACTIVITIESHAVEBEENFOCUSEDONSTUDYINGTHEEFFECTOFVARIOUSALLOYINGELEMENTSONTHEDELTAFERRITECONTENTANDCONTROLLINGDELTAFERRITECONTENTBYMODIFYINGTHEWELDMETALCOMPOSITIONSOVERTHEYEARS,ANUMBEROFMETHODSINCLUDINGCONSTITUTIONDIAGRAMS,FUNCTIONFITMODEL,FEEDFORWARDBACKPROPAGATIONNEURALNETWORKMODELHAVEBEENPUTFORWARDFORPREDICTINGTHEDELTAFERRITECONTENTINSTAINLESSSTEELWELDSAMONGALLTHEMETHODS,NEURALNETWORKMETHODWASREPORTEDTOBEMOREACCURATECOMPAREDTOOTHERMETHODSAPOTENTIALRISKASSOCIATEDWITHNEURALNETWORKANALYSISISOVERFITTINGOFTHETRAININGDATATOAVOIDOVERFITTING,MACKAYHASDEVELOPEDABAYESIANFRAMEWORKTOCONTROLTHECOMPLEXITYOFTHENEURALNETWORKMAINADVANTAGESOFTHISMETHODARETHATITPROVIDESMEANINGFULERRORBARSFORTHEMODELPREDICTIONSANDALSOITISPOSSIBLETOIDENTIFYAUTOMATICALLYTHEINPUTVARIABLESWHICHAREIMPORTANTINTHENONLINEARREGRESSIONINTHEPRESENTWORK,BAYESIANNEURALNETWORKBNNMODELFORPREDICTIONOFDELTAFERRITECONTENTINSTAINLESSSTEELWELDHASBEENDEVELOPEDTHEEFFECTOFVARYINGCONCENTRATIONOFTHEELEMENTSONTHEDELTAFERRITECONTENTHASBEENQUANTIFIEDFORTYPE309AUSTENITICSTAINLESSSTEELANDTHEDUPLEXSTAINLESSSTEELALLOY2205THEBNNMODELISFOUNDTOBEMOREACCURATECOMPAREDTOTHATOFTHEOTHERMETHODSFORPREDICTINGDELTAFERRITECONTENTINSTAINLESSSTEELWELDS?2003ELSEVIERSCIENCEBVALLRIGHTSRESERVEDKEYWORDSNEURALNETWORKANALYSISDELTAFERRITECONTENTAUSTENITICSTAINLESSSTEELDUPLEXSTAINLESSSTEEL1INTRODUCTIONTHEABILITYTOESTIMATETHEDELTAFERRITECONTENTACCURATELYHASPROVENVERYUSEFULINPREDICTINGTHEVARIOUSPROPERTIESOFAUSTENITICSSWELDSAMINIMUMDELTAFERRITECONTENTISNECESSARYTOENSUREHOTCRACKINGRESISTANCEINTHESEWELDS1,2,WHILEANUPPERLIMITONTHEDELTAFERRITECONTENTDETERMINESTHEPROPENSITYTOEMBRITTLEMENTDUETOSECONDARYPHASES,EGSIGMAPHASE,ETC,FORMEDDURINGELEVATEDTEMPERATURESERVICE3ATCRYOGENICTEMPERATURES,THETOUGHNESSOFTHEAUSTENITICSSWELDSISSTRONGLYINFLUENCEDBYTHEDELTAFERRITECONTENT4INDUPLEXSTAINLESSSTEELWELDMETALS,ALOWERFERRITELIMITISSPECIFIEDFORSTRESSCORROSIONCRACKINGRESISTANCEWHILETHEUPPERLIMITISSPECIFIEDTOENSUREADEQUATEDUCTILITYANDTOUGHNESS5HENCE,DEPENDINGONTHESERVICEREQUIREMENTALOWERLIMITAND/ORANUPPERLIMITONDELTAFERRITECONTENTISGENERALLYSPECIFIEDDURINGTHESELEC?CORRESPONDINGAUTHORTEL91411480232FAX91411440381EMAILADDRESSDEVIGCARERNETINMVASUDEVANTIONOFFILLERMETALCOMPOSITION,THEMOSTACCURATEDIAGRAMTODATEWRC1992ISUSEDGENERALLYTOESTIMATETHE?FERRITECONTENT6THECREQANDNIEQFORMULAEUSEDFORGENERATINGTHEWRC1992CONSTITUTIONDIAGRAMISGIVENBYCREQCRMO07NBANDNIEQNI35C20N025CUTHELIMITATIONOFTHESEEQUATIONSISTHATVALUESOFTHECOEFFICIENTSFORTHEDIFFERENTELEMENTSREMAINUNCHANGEDIRRESPECTIVEOFTHECHANGEINTHEBASECOMPOSITIONOFTHEWELDHOWEVER,THERELATIVEINFLUENCEOFEACHALLOYINGADDITIONGIVENBYTHEELEMENTALCOEFFICIENTSINTHECREQANDNIEQEXPRESSIONSISLIKELYTOCHANGEOVERTHEFULLCOMPOSITIONRANGEFURTHERMORE,THESEEXPRESSIONSIGNORETHEINTERACTIONBETWEENTHEELEMENTSALSO,THEREAREANUMBEROFALLOYINGELEMENTSTHATHAVENOTBEENCONSIDEREDINTHEWRC1992DIAGRAMELEMENTSLIKESI,TI,WHAVENOTBEENGIVENDUETOCONSIDERATIONS,THOUGHTHEYAREKNOWNTOINFLUENCETHEDELTAFERRITECONTENTHENCE,THEDELTAFERRITECONTENTESTIMATEDUSINGTHEWRC1992DIAGRAMWOULDALWAYSBELESSACCURATEANDMAYNEVERBECLOSETOTHEACTUALMEASUREDVALUEINTHEFUNCTIONFITMODEL7FORESTIMATINGFERRITE,THEDIFFERENCEINFREEENERGYBETWEENTHEFERRITEANDTHEAUSTENITEWASCALCULATED09240136/–SEEFRONTMATTER?2003ELSEVIERSCIENCEBVALLRIGHTSRESERVEDDOI101016/S092401360300430822MVASUDEVANETAL/JOURNALOFMATERIALSPROCESSINGTECHNOLOGY142200320–28FIG1SCHEMATICDIAGRAMOFTHENETWORKSTRUCTURESHOWINGTHEINPUTNODES,HIDDENUNITSANDTHEOUTPUTNODEFUNCTIONSOTHATEACHINPUTCONTRIBUTESTOEVERYHIDDENUNIT,WHERENISTHETOTALNUMBEROFINPUTVARIABLESHITANH??N?JW1IJXJΘ1I??3HERETHEBIASISDESIGNATEDASΘANDISANALOGOUSTOTHECONSTANTINLINEARREGRESSIONTHETRANSFERFROMTHEHIDDENUNITSTOTHEOUTPUTISLINEAR,ANDISGIVENBYYN?IW2IHIΘ24THEOUTPUTYISTHEREFOREANONLINEARFUNCTIONOFXJ,WITHTHEFUNCTIONUSUALLYSELECTEDFORFLEXIBILITYBEINGTHEHYPERBOLICTANGENTTHUS,THENETWORKISCOMPLETELYDESCRIBEDIFTHENUMBEROFINPUTNODES,OUTPUTNODESANDTHEHIDDENUNITSAREKNOWNALONGWITHALLTHEWEIGHTSWIJANDBIASESΘITHESEWEIGHTSWIJAREDETERMINEDBYTRAININGTHENETWORKANDINVOLVESMINIMIZATIONOFAREGULARIZEDSUMOFSQUAREDERRORSTHEBNNANALYSISOFMACKAY10ALLOWSTHECALCULATIONOFERRORBARSWITHTWOCOMPONENTSONEREPRESENTINGTHEPERCEIVEDLEVELOFNOISEΣVINTHEOUTPUTANDTHEOTHERINDICATINGTHEUNCERTAINTYINTHEDATAFITTINGTHISLATTERCOMPONENTOFTHEERRORBARSEMANATINGFROMTHEBAYESIANFRAMEWORKALLOWSTHERELATIVEPROBABILITIESOFTHEMODELSWITHDIFFERENTCOMPLEXITYTOBEASSESSEDTHISENABLESESTIMATIONOFQUANTITATIVEERRORBARS,WHICHVARYWITHTHEPOSITIONINTHEINPUTSPACEDEPENDINGONTHEUNCERTAINTYINFITTINGTHEFUNCTIONINTHATSPACEHENCE,INSTEADOFCALCULATINGAUNIQUESETOFWEIGHTS,APROBABILITYDISTRIBUTIONOFWEIGHTSISUSEDTODEFINETHEUNCERTAINTYINFITTINGTHEREFORE,THESEERRORBARSBECOMELARGEWHENDATAARESPARSEORLOCALLYNOISYINTHISCONTEXT,AVERYUSEFULMEASUREOFTHEERRORISTHELOGARITHMOFTHEPREDICTIVEERRORLPEGIVENBYTHEFOLLOWINGLPE?N12?TN?YN2ΣN2YLOG2ΠΣNY1/2?5WHERETISTHETARGETFORTHESETOFINPUTSX,WHILEYTHECORRESPONDINGNETWORKOUTPUTΣYISRELATEDTOTHEUNCERTAINTYOFFITTINGFORTHESETOFINPUTSXBYUSINGLPE,THEPENALTYFORMAKINGAWILDPREDICTIONISREDUCEDIFTHATPREDICTIONISACCOMPANIEDBYANAPPROPRIATELYLARGEERRORBAR,WITHALARGERVALUEOFTHELPEIMPLYINGABETTERMODELFURTHER,BYTHISMETHODITISALSOPOSSIBLETOAUTOMATICALLYIDENTIFYTHEINPUTVARIABLESTHATARESIGNIFICANTININFLUENCINGTHEOUTPUTVARIABLE,ASTHEINPUTVARIABLESTHATARELESSSIGNIFICANTAREDOWNWEIGHTEDINTHEREGRESSIONANALYSIS31OVERFITTINGPROBLEMINBNNANALYSIS,TWOSOLUTIONSAREIMPLEMENTEDWHICHCONTRIBUTETOAVOIDOVERFITTINGTHEFIRSTISCONTAINEDINTHEALGORITHMDUETOMACKAY12THECOMPLEXITYPARAMETERSΑANDΒAREINFERREDFROMTHEDATA,THEREFOREALLOWINGAUTOMATICCONTROLOFTHEMODELCOMPLEXITYTHESECONDRESIDESINTHETRAININGMETHODTHEDATABASEISEQUALLYDIVIDEDINTOATRAININGSETANDATESTINGSETTOBUILDAMODEL,ABOUT80NETWORKSARETRAINEDWITHDIFFERENTNUMBEROFHIDDENUNITSANDSEEDS,USINGTHETRAININGSETTHEYARETHENUSEDTOMAKEPREDICTIONSONTHEUNSEENTESTINGSETANDARERANKEDBYLPE32COMMITTEEMODELTHENETWORKSWITHDIFFERENTNUMBEROFHIDDENUNITSWILLGIVEDIFFERENTPREDICTIONSBUTPREDICTIONSWILLALSODEPENDONTHEINITIALGUESSMADEFORTHEPROBABILITYDISTRIBUTIONOFTHEWEIGHTSTHEPRIOROPTIMUMPREDICTIONSAREOFTENMADEUSINGMORETHANONEMODEL,BYBUILDINGACOMMITTEETHEPREDICTIONˉYOFACOMMITTEEOFNETWORKSISTHEAVERAGEPREDICTIONOFITSMEMBERS,ANDTHEASSOCIATEDERRORBARISCALCULATEDACCORDINGTOEQ6ˉY1L?LYLΣ21L?LΣL2Y1L?LYL?ˉY26WHERELISTHENUMBEROFNETWORKSINACOMMITTEENOTETHATWENOWCONSIDERTHEPREDICTIONSFORAGIVENSINGLESETOFINPUTSANDTHATEXPONENTLREFERSTOTHEMODELUSEDTOPRODUCETHECORRESPONDINGPREDICTIONYLINPRACTICE,ANINCREASINGNUMBEROFNETWORKSAREINCLUDEDINACOMMITTEEANDTHEIR

簡介：MIXEDDSP/FPGAIMPLEMENTATIONOFANERRORRESILIENTIMAGETRANSMISSIONSYSTEMBASEDONJPEG2000MARCOGRANGETTO,ENRICOMAGLI,MAURIZIOMARTINA,FABRIZIOVACCACERCOMCENTERFORWIRELESSMULTIMEDIACOMMUNICATIONSDIPARTIMENTODIELETTRONICAPOLITECNICODITORINOCORSODUCADEGLIABRUZZI2410129TORINOITALYGRANGETTOMAGLIPOLITOITMARTINAVACCAVLSILABOLPOLITOITPH390115644195FAX39011564099ABSTRACTTHISPAPERDESCRIBESADEMONSTRATOROFANERRORRESILIENTIMAGECOMMUNICATIONSYSTEMOVERWIRELESSPACKETNETWORKS,BASEDONTHENOVELJPEG2000STANDARDINPARTICULAR,THEDECODERIMPLEMENTATIONISADDRESSED,WHICHISTHEMOSTCRITICALTASKINTERMSOFCOMPLEXITYANDPOWERCONSUMPTION,INVIEWOFUSEONAWIRELESSPORTABLETERMINALFORCELLULARAPPLICATIONSTHESYSTEMIMPLEMENTATIONISBASEDONAMIXEDDSP/FPGAARCHITECTURE,WHICHALLOWSTOPARALLELIZESOMECOMPUTATIONALTASKS,THUSLEADINGTOEFICIENTSYSTEMOPERATION1INTRODUCTION’NOWADAYS,THEREISAGROWINGINTERESTINTHEENDTOENDTRANSMISSIONOFIMAGES,ESPECIALLYMOTIVATEDBYTHESHORTTERMDEPLOYMENTOFNEXTGENERATIONMOBILECOMMUNICATIONSERVICESUMTSIMT2000HOWEVER,TRANSMISSIONINANETWORKED,TETHERLESSENVIRONMENTPROVIDESBOTHOPPORTUNITIESANDCHALLENGESTHEWIRELESSCONTEXTIMPLIESTHATTHEDATAMAYUNDERGOBITERRORSANDPACKETLOSSES,MAKINGITNECESSARYTOFORESEEERRORRECOVERYMODALITIESITISTHEREBYNECESSARYTHATIMAGECOMMUNICATIONTECHNIQUESAREPROVIDEDWITHTHEABILITYTORECOVER,ORATLEASTCONCEAL,THEEFFECTOFSUCHLOSSESTHEFORTHCOMINGJPEG2000IMAGECOMPRESSIONSTANDARDHASBEENDESIGNEDTOMATCHTHESEREQUIREMENTS,ANDEMBEDSSOMEERRORDETECTIONANDCONCEALMENTTOOLSTHISPAPERADDRESSESTHEDEVELOPMENTOFADEMONSTRATOROFANERRORRESILIENTJPEG2000111DECODERIMPLEMENTATIONFORIMAGECOMMUNICATIONOVERALOSSYPACKETNETWORKTHEROBUSTNESSTOPACKETERASURESISACHIEVEDBYCOMBININGTHEFLEXIBILITYOFTHEJPEG2000FRAMEWORKWITHTHEPOWERFULNESSOFSOURCECHANNEL078037147X/01/100002001IEEEADAPTIVE,OPTIMIZEDREEDSOLOMONCODESTHEDECODERIMPLEMENTATIONISPARTICULARLYSIGNIFICANTINTHECONTEXTOFWIRELESSPORTABLETERMINALSFORNEXTGENERATIONCELLULARSYSTEMS,WHERETHELIMITEDPOWERBUDGETANDAVAILABLEDIMENSIONSIMPOSESEVERECONSTRAINTSONTHEDESIGNOFAMULTIMEDIAPROCESSINGSYSTEM2SYSTEMOVERVIEWTHEFUNCTIONALUNITSOFTHEIMPLEMENTEDSYSTEM21JPEG2000IMAGECOMPRESSIONJPEG2000ISTHENOVELIS0STANDARDFORSTILLIMAGECODING,ANDISINTENDEDTOPROVIDEINNOVATIVESOLUTIONSACCORDINGTOTHENEWTRENDSINMULTIMEDIATECHNOLOGIESATTHETIMEOFTHISWRITING,THESTANDARDISINADVANCEDPUBLICATIONSTAGETHEFINALCOMMITTEEDRAFTLISTHEMOSTRECENTJPEG2000DESCRIPTIONPUBLICLYAVAILABLE,WHICHOURIMPLEMENTATIONCONFORMSTOJPEG2000NOTONLYYIELDSSUPERIORPERFORMANCEWITHRESPECTTOEXISTINGSTANDARDSINTERMSOFCOMPRESSIONCAPABILITYANDSUBJECTIVEQUALITY,BUTALSONUMEROUSADDITIONALFUNCTIONALITIES,SUCHASIOSSLESSANDLOSSYCOMPRESSION,PROGRESSIVETRANSMISSION,ANDERRORRESILIENCETHEARCHITECTUREOFTHEJPEG2000ISBASEDONTHETRANSFORMCODINGAPPROACHANIMAGEMAYBEDIVIDEDINTOSEVERALSUBIMAGESTILES,TOREDUCEMEMORYANDCOMPUTINGREQUIREMENTSABIORTHOGONALDISCRETEWAVELETTRANSFORMDWTISFIRSTAPPLIEDTOEACHTILE,WHOSEOUTPUTISASERIESOFVERSIONSOFTHETILEATDIFFERENTRESOLUTIONLEVELSSUBBANDSTHEN,THETRANSFORMCOEFFICIENTSAREQUANTIZED,INDEPENDENTLYFOREACHSUBBAND,WITHANEMBEDDEDDEADZONEQUANTIZEREACHSUBBANDOFTHEWAVELETDECOMPOSITIONISDIVIDEDINTORECTANGULARBLOCKSCODEBLOCKS,WHICHAREININTHEFOLLOWINGWEPROVIDEABRIEFDESCRIPTIONOF1330CODETHERECEIVEDBITSTREAMMAKESTHEUSEOFAREEDSOLOMONFPGAIMPLEMENTATIONVERYATTRACTIVETHEJPEG2000DECODERMODULE,ENTIRELYIMPLEMENTEDONDSP,ISCOMPOSEDBYFOURMAINBLOCKSSYNTAXPARSER,ENTROPYDECODEREBCOT,UNIFORMSCALARDEQUANTIZER,ANDINVERSEWAVELETTRANSFORMMOREOVER,TWOADDITIONALTASKS,DEVOTEDTOCOMMUNICATIONMANAGEMENTBETWEENDSP,FPGAANDAPERSONALCOMPUTER,HAVEBEENINTRODUCED31SYNTAXPARSERTHEPARSERISTHEFUNCTIONALBLOCKTHATINTERFACESTHEJPEG2000DECODERWITHTHERSDECODERITRETRIEVESRSDECODEDPACKETS,ANDEXTRACTSFROMTHECOMPRESSEDJPEG2000BITSTREAMALLTHERELEVANTINFORMATIONNEEDEDTOPERFORMIMAGERECONSTRUCTIONFIRSTLY,THEBITSTREAMMAINHEADERISREAD,WHICHCONTAINSINFORMATIONONTHEPARAMETERSUSEDDURINGTHEENCODINGPROCESSEGIMAGESIZE,WAVELETFILTERUSED,NUMBEROFDECOMPOSITIONLEVELS,QUANTIZATIONTHRESHOLDS,ANDSOONAFTERTHAT,TILEHEADERSAREREAD,WHICHPROVIDEINFORMATIONSPECIFICTOEACHIMAGETILEFINALLY,EACHPACKETCONTAINEDINTHEBITSTREAMISREAD,ANDTHEDATAANDPARAMETERSOFEACHCODEBLOCKAREEXTRACTED,ANDFEDASINPUTSTOTHEEBCOTDECODER32EBCOTRIGHTAFTERTHEBITSTREAMSYNTAXPARSER,THESUBSEQUENTSTAGEINTHEJPEG2000DECOMPRESSIONCHAINISTHEENTROPYDECODEREBCOTFROMANALGORITHMICPOINTOFVIEW,EBCOTISABLOCKBASEDBITPLANEENCODERFOLLOWEDBYAREDUCEDCOMPLEXITYARITHMETICCODERMQITSUBDIVIDESEACHWAVELETSUBBANDINTOADISJOINTSETOFRECTANGULARBLOCKS,CALLEDCODEBLOCKSTHENTHECOMPRESSIONALGORITHMISINDEPENDENTLYAPPLIEDTOEVERYCODEBLOCKTHESAMPLESOFEVERYCODEBLOCKAREARRANGEDINTOSOCALLEDBITPLANESTODECODEACODEBLOCK,EBCOTALWAYSSTARTSFROMTHEMOSTSIGNIFICANTBITPLANES,ANDTHENMOVESTOWARDSTHELEASTSIGNIFICANTONESTHECOMPRESSEDINFORMATIONOFEVERYCODEBLOCKISTHENARRANGEDINSEVERALQUALITYLAYERS,TOCREATEASCALABLECOMPRESSEDBITSTREAMCONCEPTUALLY,EACHQUALITYLAYERMONOTONICALLYINCREASESTHEKNOWLEDGEOFSAMPLESMAGNITUDES,IEINCREASESTHEQUALITYOFTHERECONSTRUCTEDIMAGEFORMALLY,EBCOTISMADEOFTHREEMAINSTEPS,NAMELYSIGNIFICANCEPROPAGATIONSP,MAGNITUDEREFINEMENTMR,ANDCLEANUPCLEACHOFTHEABOVESTEPSCANRESORTTOFOURDECODINGPRIMITIVES,NAMELYZEROCODING,SIGNCODING,MAGNITUDEREFINEMENTCODING,ANDRUNLENGTHCODINGTHEBITPLANEVISITINGORDERFOLLOWSTHESEQUENCESPMRCLITISWORTHNOTICINGTHATEVERYSAMPLEOFAGIVENCODEBLOCKISPROCESSEDINJUSTONEOFTHETHREESTEPSASFARASCOMPUTATIONALCOMPLEXITYISCONCERNED,CLDEMANDSTHELARGESTEFFORTDURINGTHEDECODINGOFTHEMOSTSIGNIFICANTBITPLANESASSPSTEPSAREAPPLIED,ANINCREASINGNUMBEROFSAMPLESBECOMESIGNIFICANT,ANDAREINSERTEDINALISTOFMRREADYSAMPLESPROGRESSIVELY,THELOADREQUIREDBYMRSTEPSGROWS,MAKINGTHEDECODEREFFICIENCYDIRECTLYDEPENDENTONTHEMRANDCLOPTIMIZATIONLEVELDURINGTHEDEVELOPMENTOFTHEEBCOTDECODERBLOCK,PARTICULARCAREHASBEENPOSEDONTHEDESIGNOFAGILEDATASTRUCTURES,PARTICULARLYSUITEDTODSPOPTIMIZEDCCODEOFMRANDCLSTEPS33UNIFORMSCALARDEQUANTIZERACCORDINGTOL,THEQUANTIZATIONMETHODSUPPORTEDBYJPEG2000ISCALLEDSCALARUNIFORMUNIFORMSCALARDEQUANTIZATIONCANBESIMPLYACCOMPLISHEDBYMEANSOFASINGLEMULTIPLICATIONFOREACHWAVELETCOEFFICIENT34INVERSEWAVELETTRANSFORMTHEDISCRETEWAVELETTRANSFORMCANBEEVALUATEDBYMEANSOFACONVOLUTIONBASEDKERNEL,ORALIFTINGBASEDKERNEL,THISLATTERBEINGTHEDEFAULTTRANSFORMKERNELEMPLOYEDINJPEG2000ITHASBEENDEMONSTRATED4THATTHELIFTINGSCHEMEMAYRUNUPTOTWICEASFASTASCONVOLUTIONTHEWAVELETTRANSFORMHASTOBEPERFORMEDONBOTHIMAGEROWSANDCOLUMNS,INORDERTOOBTAINASEPARABLETWODIMENSIONALSUBBANDDECOMPOSITIONJPEG2000PERFORMSFIRSTTHECOLUMNWISE,ANDTHENTHEROWWISEFILTERINGTHEDEFAULTFILTERUSEDFORLOSSYCOMPRESSIONISTHEWELLKNOWNDB9,7SINCEITDOESNOTHAVERATIONALCOEFFICIENTS,PARTICULARCAREOUGHTTOBEPOSEDTOTHEEFFECTSOFFINITEPRECISIONREPRESENTATION5DUETOTHEUSEOFAFIXEDPOINTTITMS320C6201DSP,ADETAILEDSTUDYOFINTERNALDATAREPRESENTATIONHASBEENPERFORMEDEXPERIMENTALRESULTSSHOWSTHATEXCELLENTPERCEPTIVEQUALITYCANBEACHIEVEDRECURRINGTO9FRACTIONALBITSFORFILTERCOEFFICIENTSINORDERTOOPTIMIZETHEDYNAMICRANGEAROUNDZERO,ADCSHIFTISFORESEENBYTHESTANDARD,ASTHEDCCOMPONENTCOULDLEADTOANEXCESSIVEGROWTHOFTHEDYNAMICRANGEOFLOWPASSSUBBANDCOEFFICIENTSMOREOVER,THELOWPASSFILTERSCANKEEPTHESAMPLESINAFIXEDRANGE,PROVIDEDTHATAUNITARYDCFILTERGAINISGUARANTEEDTHEJOINTEFFECTOFDCCOMPONENTSUPPRESSIONANDUNITARYGAINENSURESTHATRANGEREQUIREMENTSAREFULFILLEDDURINGTHEWHOLEWAVELETTRANSFORM35ADAPTIVEREEDSOLOMONPACKETDETHEDEINTERLEAVINGRSDECODERHASBEENMAPPEDONTHEFPGADEVICEITISSPLITINTOTWOFUNCTIONALSUBBLOCKSTHEFIRSTISTHEDEINTERLEAVER,THESECONDISTHERSDECODERTHEFORMERCOLLECTSPACKETSRECEIVEDFROMCODING1332

簡介：INSTANTOPENCVFORIOSLEARNHOWTOBUILDREALTIMECOMPUTERVISIONAPPLICATIONSFORTHEIOSPLATFORMUSINGTHEOPENCVLIBRARYKIRILLKORNYAKOVALEXANDERSHISHKOVBIRMINGHAMMUMBAIPREFACEINSTANTOPENCVFORIOSISAPRACTICALGUIDE,SHOWINGEVERYIMPORTANTSTEPFORBUILDINGACOMPUTERVISIONAPPLICATIONFORTHEIOSPLATFORMITWILLHELPYOUTOPORTYOUROPENCVCODE,PROFILEANDOPTIMIZEIT,ANDTHENWRAPINTOAGUIAPPLICATIONTHISBOOKHELPSYOUTOLEARNHOWTOBUILDASIMPLE,BUTPOWERFULCOMPUTERVISIONAPPLICATIONFORTHEIOSDEVICESFROMSCRATCHTHROUGHOUTTHEBOOK,YOULLLEARNDETAILSTHATWILLHELPYOUTOBECOMEAPROFESSIONALATIOSDEVELOPMENTUSINGOPENCVASUSUAL,YOUBEGINWITHTHESIMPLE“HELLOWORLD“APPLICATION,BUTFINALLYYOUWILLBEABLETOCREATECOMPLEXIMAGEPROCESSINGAPPLICATIONSWITHSUPREMEEFFICIENCYEACHRECIPEISACCOMPANIEDWITHASAMPLEPROJECT,HELPINGYOUTOFOCUSONAPARTICULARASPECTOFTHETECHNOLOGYWHATTHISBOOKCOVERSFGETTINGSTARTEDWITHIOSSIMPLE,HELPSYOUTOSETUPYOURDEVELOPMENTENVIRONMENTANDRUNYOURFIRST“HELLOWORLD“IOSAPPLICATIONFDISPLAYINGANIMAGEFROMRESOURCESSIMPLE,INTRODUCESYOUTOBASICGUICONCEPTSONIOS,ANDCOVERSLOADINGOFANIMAGEFROMRESOURCESANDDISPLAYINGITONTHESCREENFLINKINGOPENCVTOANIOSPROJECTSIMPLE,EXPLAINSHOWTOLINKOPENCVLIBRARYANDCALLANYFUNCTIONFROMITFDETECTINGFACESWITHCASCADECLASSIFIERINTERMEDIATE,SHOWSHOWTODETECTFACESUSINGOPENCVFPRINTINGAPOSTCARDINTERMEDIATE,DEMONSTRATESHOWASIMPLEPHOTOEFFECTCANBEIMPLEMENTEDFWORKINGWITHIMAGESINGALLERYINTERMEDIATE,EXPLAINSHOWTOLOADANDSAVEIMAGESFROM/TOGALLERYFAPPLYINGARETROEFFECTINTERMEDIATE,DEMONSTRATESANOTHERINTERESTINGPHOTOEFFECTTHATMAKESPHOTOSLOOKOLDFTAKINGPHOTOSFROMCAMERAINTERMEDIATE,SHOWSHOWTOCAPTURESTATICIMAGESWITHCAMERA

簡介：1AFFECTIVEINTELLIGENTCARINTERFACESWITHEMOTIONRECOGNITIONCHRISTINELLISETTIDEPARTMENTOFMULTIMEDIACOMMUNICATIONS,INSTITUTEURECOMSOPHIAANTIPOLIS,FRANCECHRISTINELISETTIEURECOMFRFATMANASOZSCHOOLOFCOMPUTERSCIENCE,UNIVERSITYOFCENTRALFLORIDAORLANDO,FLFATMACSUCFEDUABSTRACTINTHISPAPER,WEUNCOVERANEWPOTENTIALAPPLICATIONFORMULTIMEDIATECHNOLOGIESAFFECTIVEINTELLIGENTCARINTERFACESFORENHANCEDDRIVINGSAFETYWEALSODESCRIBETHEEXPERIMENTWECONDUCTEDINORDERTOMAPCERTAINPHYSIOLOGICALSIGNALSGALVANICSKINRESPONSE,HEARTBEAT,ANDTEMPERATURETOCERTAINDRIVINGRELATEDEMOTIONSANDSTATESFRUSTRATION/ANGER,PANIC/FEAR,ANDBOREDOM/SLEEPINESSWEDEMONSTRATETHERESULTSWEOBTAINEDANDDESCRIBEHOWWEUSETHESERESULTSTOFACILITATEAMORENATURALHUMANCOMPUTERINTERACTIONINOURMULTIMODALAFFECTIVECARINTERFACEFORTHEDRIVERSOFTHEFUTURECARS1INTRODUCTIONANDMOTIVATIONHUMANSARESOCIALBEINGSTHATEMOTEANDTHEIRCOGNITIONISAFFECTEDBYTHEIREMOTIONSEMOTIONSINFLUENCEVARIOUSCOGNITIVEPROCESSESINHUMANS,INCLUDINGPERCEPTIONANDORGANIZATIONOFMEMORYBOWER,1981,CATEGORIZATIONANDPREFERENCEZAJONC,1984,GOALGENERATION,EVALUATION,ANDDECISIONMAKINGDAMASIO,1994,STRATEGICPLANNINGLEDOUX,1992,FOCUSANDATTENTIONDERRYBERRYEKMANCHOVIL1991,ANDLEARNINGGOLEMAN,1995PREVIOUSSTUDIESALSOSUGGESTTHATPEOPLEEMOTEWHILETHEYAREINTERACTINGWITHCOMPUTERSREEVESGROSSLEVENSON,1997HOWEVER,INTERPRETINGTHEDATAWITHSTATISTICALMETHODSANDALGORITHMSISBENEFICIALINTERMSOFACTUALLYBEINGABLETOMAPTHEMTOSPECIFICEMOTIONSSTUDIESHAVEDEMONSTRATEDTHATALGORITHMSCANBEVERYSUCCESSFULLYIMPLEMENTEDFORRECOGNITIONOFEMOTIONSFROMPHYSIOLOGICALSIGNALSCOLLETETALCOLLET,VERNETMAURY,DELHOMME,DITTMAR,1997SHOWEDNEUTRALANDEMOTIONALLYLOADEDPICTURESTOPARTICIPANTSINORDERTOELICITHAPPINESS,SURPRISE,ANGER,FEAR,SADNESS,ANDDISGUSTTHEPHYSIOLOGICALSIGNALSMEASUREDWERESKINCONDUCTANCESC,SKINPOTENTIALSP,SKINRESISTANCESR,SKINBLOODFLOWSBF,SKINTEMPERATUREST,ANDINSTANTANEOUSRESPIRATORYFREQUENCYIRFSTATISTICALCOMPARISONOFDATASIGNALSWASPERFORMEDPAIRWISE,WHERE6EMOTIONSFORMED15PAIRSOUTOFTHESE15EMOTIONPAIRS,ELECTRODERMALRESPONSESSR,SC,ANDSPDISTINGUISHED13PAIRS,ANDSIMILARLYCOMBINATIONOFTHERMOCIRCULATORYVARIABLESSBFANDSTANDRESPIRATIONCOULDDISTINGUISH14EMOTIONPAIRSSUCCESSFULLYPICARDETALPICARD,HEALEY,VYZAS,2001USEDPERSONALIZEDIMAGERYANDEMOTIONALLYLOADEDPICTURESTOELICITHAPPINESS,SADNESS,ANGER,FEAR,DISGUST,SURPRISE,NEUTRALITY,PLATONICLOVE,ANDROMANTICLOVETHEPHYSIOLOGICALSIGNALSMEASUREDWEREGSR,HEARTBEAT,RESPIRATION,ANDELECTROCARDIOGRAMTHEALGORITHMSUSEDTOANALYZETHEDATAWERESEQUENTIALFORWARDFLOATINGSELECTIONSFFS,FISHERPROJECTION,ANDAHYBRIDOFTHESETWOTHEBESTCLASSIFICATIONACHIEVEMENTWASGAINEDBYTHEHYBRIDMETHOD,WHICHRESULTEDIN81OVERALLACCURACYHEALEY’SRESEARCHHEALEY,2000WASFOCUSEDONRECOGNIZINGSTRESSLEVELSOFDRIVERSBYMEASURINGANDANALYZINGTHEIRPHYSIOLOGICALSIGNALSSKINCONDUCTANCE,HEARTACTIVITY,RESPIRATION,ANDMUSCLEACTIVITYDURINGTHEEXPERIMENTPARTICIPANTSOFTHISSTUDYDROVEINAPARKINGGARAGE,INACITY,ANDONAHIGHWAYRESULTSSHOWEDTHATTHEDRIVERS’STRESSCOULDBERECOGNIZEDASBEINGRESTIERESTINGINTHEPARKINGGARAGE,CITYIEDRIVINGINTHEBOSTONSTREETS,ANDHIGHWAYIETWOLANEMERGEONTHEHIGHWAYWITH96ACCURACY22OURPRELIMINARYEMOTIONELICITATIONANDRECOGNITIONEXPERIMENTSINOUREMOTIONELICITATIONEXPERIMENTWEUSEDMOVIECLIPSANDDIFFICULTMATHEMATICALQUESTIONSTOELICITSIXEMOTIONSSADNESS,ANGER,SURPRISE,FEAR,FRUSTRATION,ANDAMUSEMENTANDANONINVASIVEWIRELESSWEARABLECOMPUTER–BODYMEDIASENSEWEARARMBANDFIGURE2–TOCOLLECTTHEPHYSIOLOGICALSIGNALSOFOURPARTICIPANTSGALVANICSKINRESPONSE,HEARTRATE,ANDTEMPERATUREFIGURE2BODYMEDIASENSEWEARARMBANDMATHEMATICALQUESTIONSWEREUSEDTOELICITFRUSTRATIONANDMOVIECLIPSWEREUSEDTOELICITTHEOTHERFIVEEMOTIONSMOVIECLIPSWERECHOSENBYCONDUCTINGAPILOTSTUDYTHATWASGUIDEDBYTHEPREVIOUSRESEARCHOFGROSSAND

簡介：視頻會議通過視頻會議通過TCP/IP協(xié)議在個人電腦上的實現(xiàn)協(xié)議在個人電腦上的實現(xiàn)BYJOHNFMCGOWAN,PHDDESKTOPVIDEOEXPERTCENTERAPRIL24,1997總體簡介總體簡介目前，視頻會議系統(tǒng)通過TCP/IP協(xié)議的局域網(wǎng)和廣域網(wǎng)在個人電腦上大量的安裝和配置仍具有許多挑戰(zhàn)。除了視頻會議應用程序之外，還有視頻會議系統(tǒng)的使用和在個人計算機之內頻繁地修改如下五個子系統(tǒng)等挑戰(zhàn)。?視頻顯示?視頻采集?音頻輸出?音頻輸入?TCP/IP網(wǎng)絡（包括網(wǎng)卡和TCP/IP的軟件）每個子系統(tǒng)因本身就是較為復雜的。因此，在視頻會議系統(tǒng)的設施和配置期間，問題可能會發(fā)生系統(tǒng)的任何地方。此外，程序又必須在用戶端安裝，而在大多數(shù)個人計算機上，視頻會議系統(tǒng)硬件和軟件并不是標準的。通常來講，不同的個人計算機上的軟硬件導致在每臺個人計算機安裝的硬件和軟件變化極大。即使一個經(jīng)驗豐富的安裝高手在一臺特殊用戶的個人計算機上也可能遇到陌生的情況。所有版本的WINDOWSWINDOWS31,WINDOWSFORWORKGROUPS31,WINDOWS95,WINDOWS的即將發(fā)布的“MEMPHIS”,以及WINDOWSNT都是建立在一個有處理系統(tǒng)部件和VENDORSUPPLIED設備驅動程序的復雜的系統(tǒng)之上的。一些設備驅動程序實際上是通過硬件設備實現(xiàn),如視頻顯示卡,視頻采集板、聲卡、網(wǎng)卡等。其他的設備驅動程序與硬件設備驅動性能有很多共同之處,但除TCP/IP協(xié)議的實施和具有系統(tǒng)特征的軟件之外。設備驅動程序是非常強大的,因為在所有版本的WINDOWS系統(tǒng)中他們在接入硬體和軟件上都處在一個特權水平上。即使是在WINDOWS95,WINDOWS31中,系統(tǒng)仍允許多個應用程序占用操作系統(tǒng)的內存或其他的應用程序的內存。設備驅動操作的特權級別就意味著驅動比一個應用程序造成更為嚴重的損害。例如，難以言喻的奇怪的系統(tǒng)崩潰以及以及設備間沖突事故。此外，大多數(shù)的系統(tǒng)配置和安裝問題都和驅動有關。盡管在此領域中將會遇到很多的問題不可能被預期,本文將會給出通過TCP/IP的網(wǎng)絡實現(xiàn)的個人電腦上的視頻會議系統(tǒng)的有關整個安裝和配置問題的一個概述。視頻顯示視頻顯示視頻會議系統(tǒng)的重點在于個人計算機的視頻顯示視頻適配器和視頻驅動程序。視頻顯示驅動程序可以包含細微的錯誤，這將導致與應用程序發(fā)生沖突其中也包括與視頻會議的應用程序發(fā)生沖突，造成包括一般保護錯誤和在屏幕上更新等問題在內的許多問題。一般來說，要確保視頻卡具有的是最新驅動程序。大部分主流的視頻卡和視頻芯片供應商都會在其網(wǎng)站和FTP站點提供相應產品的驅動程序。對于部分用戶有可能使用特殊顯示卡驅動而不是一般的視頻芯片驅動程序的情況，用戶應向芯片制造商獲取相應的驅動程序。例如，對于鉆石系列的多媒體視頻卡，無論是鉆石顯卡還是S3顯卡，其供應商在生產鉆石系列芯片時都提供了相應的驅動程序。在鉆石系列中，同S3相比大大增強了鉆石顯卡的驅動程序。的ISA和PCI視頻采集卡有時會遇到資源沖突，這是由于即插即型卡設計實現(xiàn)的缺陷所造成的。視頻采集是通過VFW驅動程序來進行處理的。在WINDOWS31和WINDOWSFORWORKGROUPS系統(tǒng)下所提供的最新的16位的VFW程序是VFW11E版本。而在WINDOWS95系統(tǒng)中提供了一個帶有視頻壓縮功能的32位的VFW版本，此版本和VFW11E版還有一些未知的差異和聯(lián)系，WINDOWS95系統(tǒng)中所提供的這個版本的VFW已經(jīng)具有視頻采集的功能了。基于WINDOWS95OEM服務版本2（OSR2）的ACTIVEMOVIE10，可以在WINDOWS95的早期版本中安裝，但它不提供任何對視頻拍攝的支持。視頻采集卡的軟件安裝過程中應安裝VFW視頻采集驅動程序。此驅動是有如下行定義的MSVIDEOXXXXDRV在MICROSOFTWINDOWSSYSTEMINI文件中進行驅動選擇。視頻顯示應處理的問題視頻顯示應處理的問題?重新啟動?檢查IRQ或者其他資源是否沖突?檢查視頻采集攝像頭和網(wǎng)線連接若有電氣連接故障微調網(wǎng)線?檢查視頻采集卡是否在主板上插裝好?用顯卡檢查視頻會議系統(tǒng)的錯誤記錄文檔?確保有正確或者最新的顯卡適配器驅動?用WINDOWS系統(tǒng)的控制面板重裝顯卡驅動?使用顯卡安裝程序重新安裝顯卡驅動?在WINDOWS31下，通過編輯SYSTEMINI中手動安裝驅動程序（請備份原始SYSTEMINI）?安裝新的或不同的顯卡視頻會議的應用程序視頻會議的應用程序視頻會議系統(tǒng)的客戶端視頻會議系統(tǒng)的客戶端安裝視頻會議系統(tǒng)的第一步都來自于應用程序的安裝。這些通常是INSTALLSHIELD或其它商業(yè)公司提供的安裝程序。安裝程序將安裝包括所有應用的軟件組件。安裝程序經(jīng)常也會伴隨著安裝視頻會議系統(tǒng)所需的視頻捕獲卡和聲卡。不幸的是，安裝程序有時會安裝失敗。如果懷疑程序安裝的有問題，請先卸載視頻會議系統(tǒng)的程序，然后重新運行安裝程序進行安裝。在絕大多數(shù)情況下，這樣操作都會有所幫助?？偨Y總結視頻會議系統(tǒng)通過TCP/IP網(wǎng)絡在個人電腦上的實現(xiàn)可能會涉及大量的個人電腦和網(wǎng)絡的安裝和配置。不幸的是，在個人電腦上安裝和配置系統(tǒng)的硬件和軟件是很復雜一件事。但在個人電腦上安裝和配置系統(tǒng)的硬件和軟件也不是像研究火箭科學那樣艱難。它需要有很多的步驟來去實現(xiàn)，特別是對于像視頻會議系統(tǒng)這樣復雜的產品。單獨拿出每一步來說都很簡單，但很多簡單的步驟組合在一起，導致了一個相當復雜的過程。技術支持人員或用戶必須避免被繁雜的步驟和頻頻出現(xiàn)的各種奇怪的問題所嚇倒，哪怕問題是出現(xiàn)在WINDOWS系統(tǒng)上的安裝問題。

簡介：航空攝影測量中的立體模型重建摘要本文描述的是現(xiàn)代航空攝影測量的操作問題和基本的技術需要。當立體模型重建時，利用航空攝影測量中的外方位元素決定攝影測量點的精度和在對應的模型點中的Y視差分析。真正的航空攝影，在圖像的比例，由12500至160000，與DGPS/IMU的數(shù)據(jù)來源于各種地形，在中國由我們的POS支持的大型區(qū)域網(wǎng)平差計劃WUCAPS處理。實證結果證實來源于大型區(qū)域網(wǎng)平差的外方位元素的精度符合地形勘測規(guī)范的要求。然而，通過POS確定的外方位元素的精度不能滿足地形勘測規(guī)范的要求。關鍵詞空中三角測量（AT）；GPS（全球定位系統(tǒng)）；POS（定位和定向系統(tǒng)）；立體模型重建；地面控制點（GCPS）；精度導言航空攝影測量是從空中影像獲得關于地球表面的三維空間信息的科學和技術。攝影點的決定，其中通過使用圖像找出地面對象，是依據(jù)識別物體的遙感。并且問題的關鍵是迅速和準確地確定圖像的位置和行為上的即時影像。通過基于分布式地面控制點的空中三角測量滿足這一目標。隨著空間定位技術的發(fā)展，遙感技術和計算機科學，以及空中三角測量的演變和發(fā)展走向沒有地面控制點的數(shù)字化勘測。早在1950年，攝影科學家就開始研究如何利用各種輔助數(shù)據(jù)，以減少地面控制點的需要。然而，由于技術的局限性，方法沒有變成現(xiàn)實。直到20世紀70年代，出現(xiàn)了美國的全球定位系統(tǒng)（GPS），在航空攝影過程中人們僅得到通過載波相位差分全球定位系統(tǒng)（DGPS）技術來確定曝光駐地的位置即航攝照片的三個線性元素，用于執(zhí)行空中三角測量（簡稱GPS支持AT）可以減少攝影對地面控制點的依賴，縮短測繪周期；并降低生產成本，在攝影測量的領域觸發(fā)革命。然而，GPS支持AT在空中攝影測量的操作是有利的，主要是在浩大和困難的區(qū)域，在中小型的比例尺，而不是帶狀區(qū)域和城市大比例尺測圖。在20世紀90年代，人們開始探討采用GPS/LNS集成系統(tǒng)也稱POS獲取照片的位置和姿態(tài)即利用GPS獲得曝光駐地的位置，由IMU獲得圖像姿態(tài)元素，目的是照片的定向，最終目標是取代區(qū)域空中。三角測量程序?，F(xiàn)代數(shù)字攝影測量學在4D產品DEM，DOM，DLG，DRG的自動化的生產和空間數(shù)據(jù)庫的更新中將扮演一個重要角色。本文將介紹航空攝影測量學和相關的技術需要在當前的操作應用，特別是，攝影信息鏈的幾何定位精度可從計劃的設計應如圖3所示，即不同模式的空中攝影。23數(shù)字映射從理論上說，在得到準確的內外方位元素的圖像之后，可衡量的立體模型可利用模型重建恢復，其中我們可以做地形的測繪以及物體的自動運行。然而，目前的四維產品的生產工藝是單張照片的內定向立體像對的相對定向?單一模型的絕對定向立體模型的測繪。該方法的模型只有通過POS支持??的航空攝影測量直接地理參考恢復。3實驗和分析航攝定位有兩種方法。其中之一被稱作區(qū)域空中三角測量，關于圖像點的坐標，地面控制點的坐標和（或）圖像的外方位元素加權觀測值，并結合大型區(qū)域網(wǎng)平差來解決圖像定向參數(shù)和目標點的空間坐標，來作為方向控制點的立體模型繪圖和做高度精確的幾何定位的應用。為不同尺度和地形類型的航空攝影測量，航攝照片辦公室操作的地形圖規(guī)格定義了各自空中三角測量方法，地面控制計劃，以及傳輸點精度的具體標準。這種方法已被建立并得到了廣泛的應用。另一種是所謂的直接地理參考，假定高精確的圖像外方位元素是可以得到的，在立體像對中通過使用圖像坐標系統(tǒng)的同名像點的坐標，利用空間交會計算出對應的目標點物體的空間坐標。這種方法直接地確定對象的位置，因此4D產品可以被生產。然后本文將主要討論當利用各種方式獲得圖像的外方位元素時，如何定位精度可以完成立體模型的Y視差。

簡介：VIRTUALREALITYMODELOFAIRPOLLUTIONMMUDROVA,APROCHAZKA,ANDMKOLNOVAINSTITUTEOFCHEMICALTECHNOLOGY,DEPARTMENTOFCOMPUTINGANDCONTROLENGINEERINGABSTRACTTHEPAPERISDEVOTEDTOTHEDESCRIPTIONOFVIRTUALREALITYMODELGENERATIONUSINGTHESYSTEMMATLABANDITSVIRTUALREALITYTOOLBOXTHESELECTEDPOSSIBILITIESOFINTERCONNECTIONOFBOTHSYSTEMSMATLABANDVRMLAREDEMONSTRATEDBYMEANSOFVISUALISATIONOFTWODIMENSIONALINTERPOLATIONOFAIRPOLLUTIONDATA1INTRODUCTIONPRESENTATIONOFTHREEDIMENSIONALDATACURRENTLYREPRESENTSAVERYTOPICALTHEMEINMANYAREAS,ANDNOTONLYTECHNICALONESCLEARANDINTELLIGIBLEARRANGEMENTSOFINFORMATIONINTHE3DMODELISNECESSARYFOREXAMPLEINMEDICINEANDITISUSEDMOREANDMOREFREQUENTLYINOTHERAREASRESULTSOFINFORMATIONPROCESSINGUSINGCOMPLEXMATHEMATICALTOOLSCANNATURALLYLEADTOFORMULATIONOFATHREEDIMENSIONALGRAPHICALMODELITHASBEENPOSSIBLETOOBSERVEORPROCESSSPATIALDATAFORALLUSERSOFPERSONALCOMPUTERSSINCETHEVRMLVIRTUALREALITYMODELINGLANGUAGEHADBEENDEVELOPED5,6,7THEMOSTFREQUENTLYUSEDINTERNETBROWSERSMSINTERNETEXPLORERANDNETSCAPENAVIGATORHAVEAVRMLVIEWERBUILTINTHEREAREALSOMANYOTHERFREEWAREPLUGINMODULESAVAILABLEONTHEINTERNETWITHWHICHONECANENTERTHEVIRTUALWORLDANDOBSERVEITANDPOSSIBLYCONTROLITASWELLTHANKSTOTHESEFACTS,ITISEXPECTEDTHATAREASOFAPPLICATIONSOF3DMODELINGWILLGROWTHEGOALOFTHISPAPERISTODESCRIBETHEPROCESSOFFORMULATIONOFA3DMODELOFAIRPOLLUTIONLEVELINTHECZECHREPUBLICUSINGTHESYSTEMMATLAB,ANDITSEXPORTTOVRML2BACKGROUNDTHEVIRTUALREALISTICSCENEESTABLISHESTOGENERALLYHAVETWOKINDSOFPATHSADOPTINGOPENGL,VRML，DIRECT3DISTHE1ST。FORNONCALCULATORPROFESSIONALPERSONNELTOSAY,MAKEUSEOFOPENGLTOWRITEACOMPLICATED3DSATELLITEOBSERVATIONSANDNAMELYCONCENTRATIONSOFDUSTPARTICLESINTHEAIRTHEONEYEARSTIMESERIESREPRESENTINGCURRENTCONCENTRATIONSOFTHEPM10POLLUTIONOBTAINEDFROMTHEAIMSTATIONSHASBEENKINDLYPROVIDEDBYTHECZECHHYDROMETEOROLOGICALINSTITUTEINPRAGUETHEDATAOBSERVEDATLOCATIONSPRESENTEDINFIG1WEREPREPROCESSEDBYMEANSOFCOMPENSATIONFORVALUESOFMEASUREMENTFAILURESANDELIMINATIONOFDISTANTVALUES4FIGURE1THEAIMSTATIONLAYOUTINTHECZECHREPUBLICWITHMARKEDDELAUNAYTRIANGULATION4TWODIMENSIONALINTERPOLATIONMETHODSITISPOSSIBLETOUSEVARIOUSINTERPOLATIONMETHODSFORESTIMATIONOFVALUESOFANOBSERVEDPOLLUTANTINOTHERNONMEASUREDPOINTSSPECIEDBYASELECTEDORTHOGONALNETTHEDESCRIPTIONANDCOMPARISONOFTHESEPROBLEMSHASBEENPRESENTEDIN4ITISPOSSIBLETOUSETHESIMPLESTMETHODOFTHE0THORDERTHENEARESTNEIGHBOURMETHODTHATINTERPOLATESTHESURROUNDINGSOFAGIVENSTATIONBYTHESAMEVALUEMEASUREDINTHESTATIONVORONOIDIAGRAMS1LIMITTHESURROUNDINGBORDERSRELATEDDELAUNAYTRIANGULATIONMETHODPRESENTEDINFIG1ISUSEDFORHIGHERORDERINTERPOLATIONBILINEAR,CUBIC,SPLINEASELECTEDMETHODAPPLIEDTOTHEARRAYOFCALCULATEDCONCENTRATIONSVALUESRESULTSINTHEORTHOGONALNETDE_NINGLONGITUDEANDLATITUDETHETHREEDIMENSIONALMODELREPRESENTSANATURALPRESENTATIONOFTHESERESULTS

簡介：中文中文5640字出處出處STUDIESINPLANTSCIENCE,2001,8185196外文翻譯題目目植物土壤和肥料中硅的分析植物土壤和肥料中硅的分析土壤、植物和肥料中硅的分析方法2就是SIO2的重量。1122光譜光譜現(xiàn)代光譜技術的發(fā)展導致稱重方法被光譜法取代，光譜法更快速，更適合大量樣本的分析。這種方法常用于溶解含有硅的物質。11221為光譜分析溶解硅為光譜分析溶解硅硅，存在很多物質中，可以被強堿溶解像NA2CO3，NAOH，LIBO2，LIB2O3，氫氧化鈉是個好的選擇，因為樣品可以在便宜的鎳坩堝中低溫進行快速反應（KILMER，1965）冷卻后，催化劑被酸溶解，硅可以在光譜儀中進行測定。硅，存在很多物質中，也可以在封閉消化系統(tǒng)中溶解，CSD技術和熱解的方法相比對樣品個體需要更少的關注，因此它適合測試大量樣品。對于典型的CSD溶解技術是將樣品與王水（硝酸和鹽酸）放在一個密閉的消化容器（有時稱做“消化炸彈”）內進行反應，這個容器在干燥的烘箱內100到110攝氏度下加熱2小時（JONES和DREHER，1996）。冷卻之后，加入H3BO3和樣品再次加熱10到15分鐘，以助于產生沉淀。EILIOTT和SNYDER（1991）發(fā)明了高壓誘導消化法（AID）來溶解水稻中硅，這種方法只需要用NAOH和H2O2作為反應物，器材需要聚乙烯筒和高壓鍋。AID可以一次性處理40個或更多的樣本。AID利用高壓鍋產生壓力，而不是在消化容器中。BELL和SIMMONS（1997）發(fā)現(xiàn)了NIST和AID之間的差別。他們發(fā)現(xiàn)了NIST標注不能識別硅，他們用AID方法測定了NIST的樣本確定了硅的含量。NONOZAMSKY（1984）也描述了一種快速的從植物組織中分離硅的方法。用他們的方法把陸生植物在室溫中浸泡在HCL和HF中一晚，過濾殘渣。11222光譜分析溶解態(tài)的硅光譜分析溶解態(tài)的硅盡管硅可以在氮氧火焰下進行原子光譜吸收測定（AAS）或者ICP方法測定，但這經(jīng)常決定于手工或原子顏色和更低的器材花費、更低的設備限制。通常后者是更好的，因為其更容易被觀察。兩種方法相似，只是硅鉬藍的方法減少了一部分溶解的過程。樣本和鉬酸銨進行反應（KILMER，1965；HALLMARK1982等）加入酒石酸減少磷酸根的干擾。在反應后溶液中加入硫酸鈉，1氨基2萘酚4磺酸，在650微米用硅鉬藍發(fā)，可以檢測到002MGL1的硅（BUNTING，1944）。后者用原子色譜儀分析大量的樣本。11223非破環(huán)性光譜測定總硅非破環(huán)性光譜測定總硅一些現(xiàn)代的技術已經(jīng)被用于測定土壤，植物和肥料中總硅的含量，而無需進行預分析。X射線熒光光譜也被稱為X射線發(fā)射光譜或X射線光譜化學分析，通過硅在土壤和植物中沉淀聚集，盡管存在局限性，但今年來開發(fā)了高科技的設備，使其可以快速分析各種樣品。近紅外光譜（NIRS）也是對樣品中的硅不產生破環(huán)。這種方法有一種可靠的基礎測樣品中的水和氮，但其他部分很少。統(tǒng)計協(xié)會和NIRS標準樣品建立一個大的數(shù)據(jù)庫，但標準和未知組成部分的關系可以通過古典的方法分析。由于其快速，分析簡單，低花費和在操作方面的改進，NIRS

簡介：外文文獻外文文獻提升自行車使用的設計策略研究提升自行車使用的設計策略研究1介紹介紹自行車，一種純粹依靠人力運轉的機器，最早是被發(fā)明用來作為轉移和運輸?shù)墓ぞ?它一直持續(xù)不斷地在結構上被改進以至于現(xiàn)在自行車不僅僅是用來作為有效的交通工具更是一種物理鍛煉工具和休閑活動。2騎自行車可以讓騎車者在25KM/H的狀態(tài)下消耗780千卡路里，讓它變成了一種比其他物理鍛煉更有效的有氧運動。除此之外，考慮到國家每年在交通堵塞上付出的總數(shù)達到了228億元的花費，增加1的自行車交通會產生經(jīng)濟的效果遠比分解交通堵塞所引起的效果大，節(jié)省能源，并且也具有環(huán)境效益。并且，一千米自行車騎行可以減少167KG二氧化碳的排放。因此，自行車是一種非常具有環(huán)境效益的交通工具。1然而，現(xiàn)今在南韓的自行車使用率仍然低于發(fā)達國家。道路結構以及政府交通政策將機動車放在了首要位置,讓本來就不利的自行車使用狀況更為糟糕。自行車交通換乘在南韓只有12，遠遠低于日本的14以及荷蘭的27同時，南韓的自行車量產工業(yè)基礎也被削弱了，將第一占有率的市場轉向了中國生產的自行車。結果，貿易逆差大幅惡化以至于現(xiàn)今出口了需求中998的自行車。4這樣一種對自行車使用的漠不關心已經(jīng)導致了自行車使用率的降低和相關產業(yè)的削減。這份研究檢視了自行車設計的功能，作為一種自行車使用流行化的重要因素。根據(jù)調研，這份研究報告建議結構靠攏那些物質基礎,社會基礎,和自行車工業(yè)最終，它檢驗了自行車設計的功能可以在此類的結構內容中有所表現(xiàn)。2理解圍繞自行車使用的背景理解圍繞自行車使用的背景2121用5W1H5W1H法在每一頁你的材料上分析自行車使用的狀況法在每一頁你的材料上分析自行車使用的狀況這一章分析了自行車使用依靠5W1H法實現(xiàn)的條件。激活自行車使用首先需要“什么”這個部分，意思就是說，適當?shù)墓δ?，結構，和價格。“誰”這個部分包括使用者的年齡、性別、地位、生活方式等等。“哪里”的部分涉及到圖233對于國內自行車使用環(huán)境的分析對于國內自行車使用環(huán)境的分析為了對于國內自行車使用環(huán)境的分析，這一章節(jié)調查了自行車使用的六個區(qū)域/形式/內容這些范圍來自臨時增加的自行車使用的內容。31消費者缺乏使用自行車的動機消費者缺乏使用自行車的動機人們不是十分有動機去使用自行車。根據(jù)一份由公共管理與安全部出示的調查，66的應答者選擇了“缺少興趣”作為不使用自行車的原因。除為了鍛煉身體外，人們在騎自行車上找不到其他任何的益處那是因為汽車給了他們更快更舒適更安全的通勤體驗（去工作場所和學校）。盡管由于交通堵塞和在鬧市區(qū)的停車困難帶來了時間的失以及金錢的損失，人們仍然不愿意轉向騎自行車。在這種情況下，值得注意的是，在1970年代，荷蘭政府不得不依靠限制汽車的使用并且建造自行車道來提升自行車使用率。32自行車制造的工業(yè)基礎的減弱自行車制造的工業(yè)基礎的減弱因為南韓現(xiàn)存越來越少的自行車制造的工業(yè)基礎，由中國制造的廉價自行車充斥了大部分的自行車供給。這種情況使得使用者可以以很低的價格購買自行車，但是機械問題以及不可靠的耐用度等問題上升了。給予自行車發(fā)展和制造的工業(yè)力量已經(jīng)被消耗殆盡，看起來對于國內的工業(yè)來說要滿足自行車市場的潛在需求很難。33物質基礎的缺乏物質基礎的缺乏自行車道的數(shù)量在南韓遠遠不夠用。絕大多數(shù)在城市道路上鋪設的自行車

簡介：附錄附錄ALABVIEWBASEDINSTRUMENTCURRENTTRANSFORMERCALIBRATORXINAIHALBAOYHSONG1NORTHCHINAELECTRICPOWERUNIVERSITY,BEIJING,CHINA1072062BRUNELUNIVERSITYUKABSTRACTTHEVIRTUALINSTRUMENTVIMAINLYREFERSTOBUILDALLKINDSOFINSTRUMENTSBYSOFTWARESUCHASLABVIEW,WHICHLIKESAREALINSTRUMENTBUILDINACOMPUTERITSMAINCHARACTERISTICSAREFLEXIBILITY,MULTIFUNCTIONS,MULTIPLEUSESFORONEPCCOMPUTER,GIVINGHIGHPERFORMANCE,ANDISLESSCOSTLYINTHISPAPER,THEVITECHNOLOGYISAPPLIEDTOTHETESTANDMEASUREMENTOFINSTRUMENTCURRENTTRANSFORMERTABYUSINGTHELABVIEW,THETAACCURACYCALIBRATORWASDEVELOPEDTHISVIRTUALT4CALIBRATORCANAUTOMATICALLYMEASURETHEACCURACYOFT4ANDCANINDICATETHERATIOERRORANDPHASEERRORCURVESTHETESTSANDCALIBRATIONFORTHETASHOWTHATTHEVIRTUALTACALIBRATORCANBEUSEDINPLACEOFTHETRADITIONALCALIBRATORANDISMUCHBETTERTHANTHETRADITIONALONEKEYWORDSINSTRUMENTCURRENTTRANSFORMERTA,TACALIBRATOR,VIRTUALINSTRUMENTS,LABVIEWIINTRODUCTIONSINCE1992THEVXIBUSREV14STANDARDWASESTABLISHEDBYTHEUNITEDSTATESANDLABVIEWWASPRESENTEDBYTHENATIONALINSTRUMENTSCONL,THEVIRTUALINSTRUMENTVIHAVELAINTHEFOUNDATIONFORITSCOMMERCIALUSETHEMAINCHARACTERISTICOFVIRTUALINSTRUMENTISTHATITMAKESINSTRUMENTSBYSOFTWAREMOSTOFTHETRADITIONALINSTRUMENTCANBEDEVELOPEDBYVITHEVIISAREALINSTRUMENTMADEBYTHEPERSONALCOMPUTERTHEINSTRUMENTCURRENTTRANSFORMERTAISWIDELYUSEDINALLKINDSOFCURRENTMEASUREMENTANDITHASTHEFUNCTIONSOFPROTECTION,ISOLATIONANDEXTENDINGTHEMEASURINGRANGEWITHTHERAPIDDEVELOPMENTOFCOMPUTERMEASUREMENTANDCONTROLTECHNOLOGY,ANDWITHTHESEQUENTEMERGENCEOFCURRENTTRANSFORMERANDTRANSDUCER,THEREISANINCREASINGNUMBEROFCURRENTTRANSFORMERSWITHHIGHACCURACYANDLOWSECONDARYCURRENTTHESTANDARDTASECONDARYCURRENTISUSUALLY1AOR5ASOMENONSTANDARDTASECONDARYCURRENTMAYBE01AORLOWERALTHOUGHWEHAVETHETECHNIQUEOFTABEINGMEASUREDRESPECTIVELYROANDR,R,ARESECONDARYWINDINGSRESISTANCEOFSTANDARDTA,ERRORCURRENTDETECTINGRESISTANCE,BURDENRESISTANCEOFTABEINGMEASUREDRESPECTIVELYTOANDK,TBT,AREVOLTAGESAMPLINGPOINTSWHICHCANCALCULATETHECURRENTINTHISPAPER,ONLYVOLTAGEBETWEENKANDT,VOLTAGEBETWEENTBANDT,AREBEINGMEASUREDANDTHEYREPRESENTTHEVOLTAGEONR,ANDR,RESPECTIVELYINGENERAL,THETACALIBRATORSPRINCIPLEOFTHESAMPLERESISTANCESHOULDBE1）ITCANNOTAFFECTTHEACCURACYOFTHECOMPARISONCIRCUITINTHEIDEALCONDITIONR,ANDRDSHOULDBE0,BUTITCANNOTBESAMPLEDSOTHEREMUSTBESAMPLERESISTANCE,INTHISPAPER,R,ASSHOWNINFIG,ISUSED1）THEMAGNITUDEOFTHESAMPLERESISTANCESHOULDMAKETHESAMPLEDSTANDARDCURRENTANDERRORCURRENTINPRORATAANDSHOULDNOTHAVETOOMUCHDIFFERENCETHESAMPLEDRESISTANCEISSETBYEXPERIMENTR,ISTHESECONDARYSTANDARDCURRENTSAMPLINGRESISTANCEANDCANBE01050,R,ISTHEERRORCURRENTSAMPLINGRESISTANCEANDCANBE,R,ISTHEBURDENRESISTANCEANDITDEPENDSONTHETABEINGMEASUREDESAMPLINGTHEVOLTAGEUOANDU,ONR,ANDR,RESPECTIVELY,THERATIOERRORANDPHASEERRORARESHOWEDONTHELEDTHROUGHSOMEPROCESSANDCALCULATIONSACCORDINGTOTHETAERRORSPHASEDIAGRAM,WHENIOISMAXIMUM,THEVALUEOFIDISTHERATIOERRORWHENIOCHANGESFROMNEGATIVETOPOSITIVEANDEQUALSTO0,THEVALUEOFIDISTHEPHASEERRORFORTHESAMEPRINCIPLE,THERELATIONSHIPISEQUALTOTHEVOLTAGESIGNALU,ANDUDSHOWEDINFIG3AANDBISREPRESENTTHERATIOERRORANDPHASEERRORSEPARATELYTHETASREALRATIOERRORCANDPHASEERROR6CANBEFOUNDOUTTHROUGHPROPERCALCULATION,

簡介：外文文獻翻譯ADVANCEDCONTROLALGORITHMSEMBEDDEDINAPROGRAMMABLEABSTRACTTHISPAPERPRESENTSANINNOVATIVESELFTUNINGNONLINEARCONTROLLERASPECTADVANCEDCONTROLALGORITHMSFORPROGRAMMABLELOGICCONTROLLERSITISINTENDEDFORTHECONTROLOFHIGHLYNONLINEARPROCESSESWHOSEPROPERTIESCHANGERADICALLYOVERITSRANGEOFOPERATION,ANDINCLUDESTHREEADVANCEDCONTROLALGORITHMSITISDESIGNEDUSINGTHECONCEPTSOFAGENTBASEDSYSTEMS,APPLIEDWITHTHEAIMOFAUTOMATINGSOMEOFTHECONFIGURATIONTASKSTHEPROCESSISREPRESENTEDBYASETOFLOWORDERLOCALLINEARMODELSWHOSEPARAMETERSAREIDENTIFIEDUSINGANONLINELEARNINGPROCEDURETHISPROCEDURECOMBINESMODELIDENTIFICATIONWITHPREANDPOSTIDENTIFICATIONSTEPSTOPROVIDERELIABLEOPERATIONTHECONTROLLERMONITORSANDEVALUATESTHECONTROLPERFORMANCEOFTHECLOSEDLOOPSYSTEMTHECONTROLLERWASIMPLEMENTEDONAPROGRAMMABLELOGICCONTROLLERPLCTHEPERFORMANCEISILLUSTRATEDONAFIELDTESTAPPLICATIONFORCONTROLOFPRESSUREONAHYDRAULICVALVE’S2005ELSEVIERLTDALLRIGHTSRESERVEDKEYWORDSCONTROLENGINEERINGFUZZYMODELLINGINDUSTRIALCONTROLMODELBASEDCONTROLNONLINEARCONTROLPROGRAMMABLELOGICCONTROLLERSSELFTUNINGREGULATORS1INTRODUCTIONMODERNCONTROLTHEORYOFFERSMANYCONTROLMETHODSTOACHIEVEMOREEFFICIENTCONTROLOFNONLINEARPROCESSESTHANPROVIDEDBYCONVENTIONALLINEARMETHODS,TAKINGADVANTAGEOFMOREACCURATEPROCESSMODELSBEQUETTE,1991HENSONMURRAYSMITHSEBORG,1999INDICATETHATWHILETHEREISACONSIDERABLEANDGROWINGMARKETFORADVANCEDCONTROLLERS,RELATIVELYFEWVENDORSOFFERTURNKEYPRODUCTSEXCELLENTRESULTSOFADVANCEDCONTROLCONCEPTS,BASEDONFUZZYPARAMETERSCHEDULINGTAN,HANG,BABUSˇKA,OOSTERHOFF,OUDSHOORN,GUNDALA,HOO,HA¨GLANDSECTION3GIVESABRIEFDESCRIPTIONOFTHECTANDFINALLY,SECTION4DESCRIBESTHEAPPLICATIONOFTHECONTROLLERTOAPILOTPLANTWHEREITISUSEDFORCONTROLOFTHEPRESSUREDIFFERENCEONAHYDRAULICVALVEINAVALVETESTAPPARATUS2RUNTIMEMODULETHERTMOFTHEASPECTCONTROLLERCOMPRISESASETOFMODULES,LINKEDINTHEFORMOFAMULTIAGENTSYSTEMFIG1SHOWSANOVERVIEWOFTHERTMANDITSMAINMODULESTHESIGNALPREPROCESSINGAGENTSPA,THEONLINELEARNINGAGENTOLA,THEMODELINFORMATIONAGENTMIA,THECONTROLALGORITHMAGENTCAA,THECONTROLPERFORMANCEMONITORCPM,ANDTHEOPERATIONSUPERVISOROS21MULTIFACETEDMODELMFMTHEASPECTCONTROLLERISBASEDONTHEMULTIFACETEDMODELCONCEPTPROPOSEDBYSTEPHANOPOULUS,HENNING,ANDLEONE1990ANDINCORPORATESSEVERALMODELFORMSREQUIREDBYTHECAAANDTHEOLASPECIFICALLY,THEMFMINCLUDESASETOFLOCALFIRSTANDSECONDORDERDISCRETETIMELINEARMODELSWITHTIMEDELAYANDOFFSET,WHICHARESPECIFIEDBYAGIVENSCHEDULINGVARIABLESKTHEMODELEQUATIONOFFIRSTORDERLOCALMODELSIS

簡介：PROCEEDINGSOFTHEIEEERASINTERNATIONALCONFERENCEONHUMANOIDROBOTSCOPYRIGHT?2001MECHANICALSYSTEMANDCONTROLSYSTEMOFADEXTEROUSROBOTHANDDIRKOSSWALD,HEINZW?RNUNIVERSITYOFKARLSRUHEDEPARTMENTOFCOMPUTERSCIENCEINSTITUTEFORPROCESSCONTROLANDROBOTICSIPRENGLERBUNTERING8BUILDING4028D76131KARLSRUHEEMAILOSSWALDIRAUKADE,WOERNIRAUKADEABSTRACTINRECENTYEARSNUMEROUSROBOTSYSTEMSWITHMULTIFINGEREDGRIPPERSORHANDSHAVEBEENDEVELOPEDALLAROUNDTHEWORLDMANYDIFFERENTAPPROACHESHAVEBEENTAKEN,ANTHROPOMORPHICANDNONANTHROPOMORPHICONESNOTONLYTHEMECHANICALSTRUCTUREOFSUCHSYSTEMSWASINVESTIGATED,BUTALSOTHENECESSARYCONTROLSYSTEMWITHTHEHUMANHANDASANEXEMPLAR,SUCHROBOTSYSTEMSUSETHEIRHANDSTOGRASPDIVERSEOBJECTSWITHOUTTHENEEDTOCHANGETHEGRIPPERTHESPECIALKINEMATICABILITIESOFSUCHAROBOTHAND,LIKESMALLMASSESANDINERTIA,MAKEEVENCOMPLEXMANIPULATIONSANDVERYFINEMANIPULATIONSOFAGRASPEDOBJECTWITHINTHEOWNWORKSPACEOFTHEHANDPOSSIBLESUCHCOMPLEXMANIPULATIONSAREFOREXAMPLEREGRASPINGOPERATIONSNEEDEDFORTHEROTATIONOFAGRASPEDOBJECTAROUNDARBITRARYANGLESANDAXISWITHOUTDEPOSITINGTHEOBJECTANDPICKINGITUPAGAININTHISPAPERANOVERVIEWONTHEDESIGNOFSUCHROBOTHANDSINGENERALISGIVEN,ASWELLASAPRESENTATIONOFANEXAMPLEOFSUCHAROBOTHAND,THEKARLSRUHEDEXTEROUSHANDIITHEPAPERTHENENDSWITHTHEPRESENTATIONOFSOMENEWIDEASWHICHWILLBEUSEDTOBUILDANENTIRENEWROBOTHANDFORAHUMANOIDROBOTUSINGFLUIDICACTUATORSKEYWORDSMULTIFINGEREDGRIPPER,ROBOTHAND,FINEMANIPULATION,MECHANICALSYSTEM,CONTROLSYSTEM1INTRODUCTIONTHESPECIALRESEARCHAREAHUMANOIDROBOTSFOUNDEDINKARLSRUHE,GERMANYINJULY2001ISAIMEDATTHEDEVELOPMENTOFAROBOTSYSTEMWHICHCOOPERATESANDINTERACTSPHYSICALLYWITHHUMANBEINGSINNORMALENVIRONMENTSLIKEKITCHENORLIVINGROOMSSUCHAROBOTSYSTEMWHICHISDESIGNEDTOSUPPORTHUMANSINNONSPECIALIZED,NONINDUSTRIALSURROUNDINGSLIKETHESEMUST,AMONGMANYOTHERTHINGS,BEABLETOGRASPOBJECTSOFDIFFERENTSIZE,SHAPEANDWEIGHTANDITMUSTALSOBEABLETOFINEMANIPULATEAGRASPEDOBJECTSUCHGREATFLEXIBILITYCANONLYBEREACHEDWITHANADAPTABLEROBOTGRIPPERSYSTEM,ASOCALLEDMULTIFINGEREDGRIPPERORROBOTHANDTHEHUMANOIDROBOT,WHICHWILLBEBUILTINTHEABOVEMENTIONEDRESEARCHPROJECT,WILLBEEQUIPPEDWITHSUCHAROBOTHANDSYSTEMTHISNEWHANDWILLBEBUILTBYTHECOOPERATIONOFTWOINSTITUTES,THEIPRINSTITUTEFORPROCESSCONTROLANDROBOTICSATTHEUNIVERSITYOFKARLSRUHEANDTHEIAIINSTITUTEFORAPPLIEDCOMPUTERSCIENCEATTHEKARLSRUHERESEARCHCENTERBOTHORGANIZATIONSALREADYHAVEEXPERIENCEINBUILDINGSUCHKINDOFSYSTEMS,BUTFROMSLIGHTLYDIFFERENTPOINTSOFVIEWTHEKARLSRUHEDEXTEROUSHANDIISEEFIGURE1BUILTATTHEIPR,WHICHISDESCRIBEDHEREINDETAIL,ISAFOURFINGEREDAUTONOMOUSGRIPPERTHEHANDSBUILTATTHEIAISEEFIGURE17AREBUILTASPROSTHESISFORHANDICAPPEDPEOPLETHEAPPROACHTAKENSOFARWILLBEPRESENTEDANDDISCUSSEDINTHEFOLLOWINGSECTIONS,ASITFOUNDSTHEBASISFORTHENOVELHANDOFTHEHUMANOIDROBOTFIGURE1KARLSRUHEDEXTROUSHANDIIFROMIPRPROCEEDINGSOFTHEIEEERASINTERNATIONALCONFERENCEONHUMANOIDROBOTSCOPYRIGHT?2001GRASPSTATESENSORSPROVIDEINFORMATIONABOUTTHECONTACTSITUATIONBETWEENTHEFINGERANDTHEOBJECTTHISTACTILEINFORMATIONCANBEUSEDTODETERMINETHEPOINTINTIMEOFTHEFIRSTCONTACTWITHTHEOBJECTWHILEGRASPING,ANDTOAVOIDUNDESIREDGRASPS,LIKEGRASPINGATANEDGEORATIPOFTHEOBJECTBUTITCANALSOBEUSEDTODETECTSLIPPAGEOFANALREADYGRASPEDOBJECT,WHICHMIGHTLEADTOALOSSOFTHEOBJECTOBJECTSTATEORPOSESENSORSAREUSEDTODETERMINETHESHAPE,POSITIONANDORIENTATIONOFANOBJECTINTHEWORKSPACEOFTHEGRIPPERTHISISNECESSARYIFTHESEDATAISNOTKNOWNEXACTLY,PRIORTOGRASPINGTHEOBJECTIFTHEOBJECTSTATESENSORSSTILLWORKSONAGRASPEDOBJECTITCANBEUSEDTOCONTROLTHEPOSEPOSITIONANDORIENTATIONOFAGRASPEDOBJECTTOO,EGTODETECTSLIPPAGEDEPENDINGONTHEACTUATORSYSTEMTHEGEOMETRICALINFORMATIONABOUTTHEFINGERJOINTPOSITIONCANBEMEASUREDATTHEMOVEMENTGENERATORORDIRECTLYATTHEJOINTFOREXAMPLEIFTHEREISASTIFFCOUPLINGBETWEENANELECTRICMOTORANDTHEFINGERJOINTTHENTHEJOINTPOSITIONCANBEMEASUREDBYANANGLEENCODERATTHEAXISOFTHEMOTORBEFOREORAFTERTHEGEARTHISISNOTPOSSIBLEIFTHECOUPLINGISLESSSTIFFANDAHIGHPOSITIONPRECISIONISDESIRED34THEMECHANICALSYSTEMOFTHEKARLSRUHEDEXTEROUSHANDIIINORDERTOPERMITMORECOMPLEXMANIPULATIONSLIKEREGRASPINGTHECURRENTKARLSRUHEDEXTEROUSHANDIIKDHIIWASBUILTWITH4FINGERSAND3INDEPENDENTJOINTSPERFINGERITISDESIGNATEDFORAPPLICATIONSININDUSTRIALENVIRONMENTSSEEFIGURE2ANDFORMANIPULATIONOFOBJECTSLIKEBOXES,CYLINDERS,SCREWSORNUTSTHEREFOREASYMMETRIC,NONANTHROPOMORPHICCONFIGURATIONOFFOURIDENTICALFINGERS,EACHROTATEDBY90°WASCHOSENSEEFIGURE3DUETOTHEEXPERIENCESGAINEDWITHTHEFIRSTKARLSRUHEDEXTEROUSHAND,LIKEEGMECHANICALPROBLEMSCAUSEDBYTHEDRIVEBELTSORCONTROLLINGPROBLEMSCAUSEDBYLARGEFRICTIONFACTORS,SOMEDIFFERENTDESIGNDECISIONSWERECHOSENFORTHEKDHIITHEDCMOTORSFORJOINT2AND3OFEACHFINGERAREINTEGRATEDINTOTHEPREVIOUSFINGERLIMBSEEFIGURE4THISPERMITSTHEUSEOFVERYSTIFFBALLSPINDLEGEARSFORTHEFORWARDINGOFTHEMOVEMENTTOTHEFINGERJOINTANGLEENCODERSDIRECTLYONTHEMOTORAXISBEFORETHEGEARAREUSEDASVERYPRECISEPOSITIONSTATESENSORSFIGURE4SIDEVIEWOFTHEKDHII3LASERSENSORSFIXATIONFRAMEONECOMPLETEFINGERCONTROLHARDWAREMICROCONTROLLERFIGURE3TOPVIEWOFTHEKDHIIFIGURE2KDHIIMOUNTEDONANINDUSTRIALROBOT

簡介：BENDABILITYANDFORMINGBEHAVIOUROFHIGHSTRENGTHSTEELINUBENDINGOPERATIONPKAEWTATIP1,NPRASITKHETKHAN1,AKHANTACHAWANA1,VPREMANOND1,RHATO1,BSRESOMREONG1,NKOGA21KINGMONGKUT’SUNIVERSITYOFTECHNOLOGYTHONBURI,THAILAND2NIPPONINSTITUTEOFTECHNOLOGY,JAPANSUMMARYINTHISWORK,THEBENDABILITYANDFORMINGBEHAVIOUROFHIGHSTRENGTHSTEELSINUBENDINGWERESTUDIEDTHESHEETMATERIALSUSEDINTHEEXPERIMENTSAREJISSAPH440THICKNESS2MM,SPFH590THICKNESS2MMANDSPFC780YTHICKNESS12MMFIRSTLY,MINIMUMBENDRADIUSOFEACHMATERIALWASINVESTIGATEDNEXT,THEEXPERIMENTSTOCOMPARETHREEDIFFERENTMETHODSOFMINIMIZINGSPRINGBACKWERECARRIEDOUTTHEFIRSTMETHODIS,SOCALLEDBOTTOMINGTHATISTHEREDUCTIONOFSHEETTHICKNESSATTHEBOTTOMOFTHEWORKPIECEORATBENDEDANGLEBY10THESECONDONEISBOTTOMOVERBENDING,IE,THEMETHODTOOVERBENDTHEWORKPIECEATBOTTOMBY6OFORTHISMETHOD,THEPUNCHHAVINGACONCAVEBOTTOMSURFACEANDTHEBOTTOMPADHAVINGCORRESPONDEDCONVEXUPPERSURFACEHAVEBEENUSEDTHELASTONEISFLANGEOVERBENDING,IE,THEMETHODTOOVERBENDTHEFLANGEOFWORKPIECEBY6OTHISCANBEDONEBYUSINGTHEPUNCHHAVINGSIDERELIEFANGLETAPERPUNCHTHEPUNCHESANDDIESEDGERADIIARE5MMTHECLEARANCEBETWEENPUNCHANDDIEAREDETERMINEDTOBETHESAMEASMATERIALTHICKNESSTWOKINDSOFSHEETLAYOUTSWEREEXPERIMENTED,IE,THESHEETSWEREPLACEDINORDERTHATTHEIRROLLINGDIRECTIONSWERE1PARALLELAND2PERPENDICULARTOTHEBENDLINETHERESULTSREVEALEDTHATTHESPRINGBACKANGLEINCREASEDWITHTHESTRENGTHOFSHEETMATERIALTHEBOTTOMANDFLANGEOVERBENDINGMETHODSAREMOREEFFECTIVETOREDUCESPRINGBACKTHANBOTTOMINGMETHODINADDITION,FORBOTTOMINGMETHOD,THEFORCEREQUIREDWASABOUT8TIMESHIGHERTHANCONVENTIONALBENDINGFORCE1INTRODUCTIONNOWADAYS,HIGHSTRENGTHSTEELSHEETSHAVEBEENWIDELYUSEDINAUTOMOBILEINDUSTRYINORDERTOREDUCEWEIGHTOFTHEVEHICLESWHICHISSTRONGLYRELATEDTOTHEIRFUELCONSUMPTIONRATE14HOWEVER,ITISGENERALLYKNOWNTHATTHESTRENGTHOFTHESHEETS,WHICHISRELATIVELYHIGHERTHANTHATOFTHECONVENTIONALCARBONSTEELSHEETS,LEADSTOTHEIRLOWFORMABILITYANDHIGHSPRINGBACKOFTHEDEFORMEDPARTSMANYWORKSPROPOSEDTOREDUCESPRINGBACKOFTHEHIGHSTRENGTHSTEELFOREXAMPLES,MORI2PROPOSEDTOCONTROLTHESPRINGBACKOFTHEVBENDEDPARTBYUTILIZINGCNCSERVOPRESSTOREDUCETHESHEETTHICKNESSATBENDANGLENEXT,YAMANO3HASBEENSTUDIEDTOREDUCESIDEWALLCURLOFTHEDRAWBENDEDUSHAPEPARTBYUSINGSOCALLEDOVERRUNINDUCINGPUNCHYOSHIDA4STUDIEDACRASHFORMINGMETHODTOREDUCESPRINGBACKOFTHEPARTMADEOFHIGHSTRENGTHSTEELSHEETYANAGIMOTO5,6SHOWEDTHATSPRINGBACKFREEFORMINGOFHIGHSTRENGTHSTEELSHEETSCOULDBEACHIEVEDBYFORMINGTHESHEETATELEVATEDTEMPERATUREINTHERANGEOFWARMWORKINGTEMPERATUREHIGHERTHAN750KBUTCONSIDERABLYLOWERTHANHOTWORKINGTEMPERATURETHEMETHODSTOREDUCESPRINGBACKUSEDINTHEPREVIOUSWORKSAREMOSTLYBASEDONBOTTOMINGANDOVERBENDINGPRINCIPLESINTHISWORK,THEEXPERIMENTSTOCOMPARETHERESULTSOFTHREEDIFFERENTMETHODSOFMINIMIZINGSPRINGBACKWERECARRIEDOUTINORDERTOVERIFYTHEIREFFECTIVENESSINELIMINATIONTHESPRINGBACKOFHIGHSTRENGTHSTEELSHEETTHOSEMETHODSAREBOTTOMING,FLANGEOVERBENDINGANDBOTTOMOVERBENDING,RESPECTIVELYINADDITION,THEBENDABILITYWHICHISREPRESENTEDBYMINIMUMBENDRADIUSWASALSOINVESTIGATEDFORTHESHEETHAVINGTHESTRENGTHRANGEDFROM440TO780MPA2EXPERIMENTALSETUPANDMETHODOLOGYTHREEKINDSOFSHEETMATERIALS,ASSHOWNINTABLE1,WEREUSEDINTHEEXPERIMENTTHESHEETSWERECUTINTORECTANGULARSHAPEWITHDIMENSIONOF120X50MMTHEWORKPIECESHAVEBEENDEFORMED,BYTHETOOLSINFIG1A,INTOUSHAPEHAVINGDIMENSIONASSHOWNINFIG1BTHEDIESETISSHOWNINFIGTA12ICTP2008THE9THINTERNATIONALCONFERENCEONTECHNOLOGYOFPLASTICITY295ABOTTOMINGBFLANGEOVERBENDINGCBOTTOMOVERBENDINGFIG3THREEDIFFERENTMETHODSOFELIMINATIONSPRINGBACKUSEDINTHEEXPERIMENTS3RESULTSANDDISCUSSIONS31BENDABILITIESMINIMUMBENDRADIUSTHERATIOOFBENDINGFORCEREQUIREDANDSHEETTHICKNESSFOREACHMATERIALARESHOWNINFIG4ITISCLEARLYSHOWNTHATLARGERFORCESAREREQUIREDFORTHEMATERIALHAVINGHIGHERSTRENGTHANDFORTHEPUNCHHAVINGSMALLEREDGERADIUSINCASESOFUSINGTHEPUNCHWITHSHARPEDGERP0,THEREQUIREDFORCESARELARGESTWHICHARE187190TIMESOFTHOSEWHENUSINGTHEPUNCHHAVINGRP5MMTHERESULTSOFMINIMUMBENDRADIUSARESHOWNINTABLE2THEDEFORMEDPARTSWEREOBSERVEDBYBOTHVISUALMETHODANDOPTICALMICROSCOPEFOURDIFFERENTSYMBOLSWEREUSEDTODISTINGUISHTHEQUALITYOFPARTSTHEDEFINITIONOFEACHSYMBOLISINDICATEDBELOWTHESAMETABLETHESAMPLEPICTURES,INTHECASESOFUSINGSHARPEDGEPUNCH,CORRESPONDEDTOEACHSYMBOLARESHOWNINTABLE3ASTHERESULTS,FORALLTHREEKINDSOFSHEETMATERIALS,BENDINGPERPENDICULARTOROLLINGDIRECTIONISEASIERTHANBENDINGPARALLELTOROLLINGDIRECTION,ASGENERALLYKNOWNBENDABILITIES,WHICHAREREPRESENTEDBYMINIMUMBENDRADIUS,BECAMEWORSEWITHINCREASINGOFSTRENGTHOFMATERIALSFORSAPH440,THEWORKPIECESWITHOUTFRACTURECOULDBEOBTAINEDALTHOUGHUSINGTHEPUNCHWITHSHARPEDGERP0WHENBENDINGPERPENDICULARTOROLLINGDIRECTIONONTHEOTHERHAND,SPFH590COULDBESUCCESSFULLYBENDEDIFTHERATIOOFPUNCHRADIUSANDSHEETTHICKNESSWASLARGERTHAN050MOREOVER,THESAMERATIOSHOULDBELARGERTHAN083INTHECASEOFSPFC780YSHEETTHESEMIGHTBEEXPLAINEDBYTHEDIFFERENTVALUESOFTHEDUCTILITYOFTHESHEETMATERIALSTHEMINIMUMBENDRADIUSISSMALLERFORTHEMATERIALHAVINGHIGHERDUCTILITYELONGATIONATBREAKASSHOWNINTABLE132COMPARISONOFDIFFERENTMETHODSOFSPRINGBACKELIMINATIONTHEFORCETRAVELDIAGRAMSOFFORMINGSAPH440WORKPIECESBYCONVENTIONALUBENDING,BOTTOMING,FLANGEOVERBENDINGANDBOTTOMOVERBENDINGARESHOWNINFIG5THEMAXIMUMFORCESREQUIREDFORFIG4FORCESREQUIREDFORUBENDINGOFEACHMATERIALR56OR501T6OR5137108957318514911297244198163129051015202530R0R1R2R5SAPH440SPFH590SPFC780YPUNCHEDGERADIUSRPMMBENDINGFORCE/SHEETTHICKNESSKN/MMICTP2008THE9THINTERNATIONALCONFERENCEONTECHNOLOGYOFPLASTICITY297

簡介：ORIGINALARTICLEGENERATINGTOOLPATHWITHSMOOTHPOSTURECHANGEFORFIVEAXISSCULPTUREDSURFACEMACHININGBASEDONCUTTER’SACCESSIBILITYMAPLLLIYFZHANGHYLILGENGRECEIVED26MAY2009/ACCEPTED16JULY2010/PUBLISHEDONLINE4AUGUST2010SPRINGERVERLAGLONDONLIMITED2010ABSTRACTINFIVEAXISHIGHSPEEDMILLING,ONEOFTHEKEYREQUIREMENTSTOENSURETHEQUALITYOFTHEMACHINEDSURFACEISTHATTHETOOLPATHMUSTBESMOOTH,IE,THECUTTERPOSTURECHANGEFROMONECUTTERCONTACTPOINTTOTHENEXTNEEDSTOBEMINIMIZEDTHISPAPERPRESENTSANEWMETHODFORGENERATINGFIVEAXISTOOLPATHSWITHSMOOTHTOOLMOTIONANDHIGHEFFICIENCYBASEDONTHEACCESSIBILITYMAPAMAPOFTHECUTTERATAPOINTONTHEPARTSURFACETHECUTTER’SAMAPATAPOINTREFERSTOITSPOSTURERANGEINTERMSOFTHETWOROTATIONALANGLES,WITHINWHICHTHECUTTERDOESNOTHAVEANYINTERFERENCEWITHTHEPARTANDTHESURROUNDINGOBJECTSBYUSINGTHEAMAPATAPOINT,THEPOSTURECHANGERATESALONGTHEPOSSIBLECUTTINGDIRECTIONSCALLEDTHESMOOTHNESSMAPORSMAPATTHEPOINTAREESTIMATEDBASEDONTHEAMAPSANDSMAPSOFALLTHESAMPLEDPOINTSOFTHEPARTSURFACE,THEINITIALTOOLPATHWITHTHESMOOTHESTPOSTURECHANGEISGENERATEDFIRSTSUBSEQUENTLY,THEADJACENTTOOLPATHSAREGENERATEDONEATATIMEBYCONSIDERINGBOTHPATHSMOOTHNESSANDMACHININGEFFICIENCYCOMPAREDWITHTRADITIONALTOOLPATHGENERATIONMETHODS,EG,ISOPLANAR,THEPROPOSEDMETHODCANGENERATETOOLPATHSWITHSMALLERPOSTURECHANGERATEANDYETSHORTEROVERALLPATHLENGTHTHEDEVELOPEDTECHNIQUESCANBEUSEDTOAUTOMATEFIVEAXISTOOLPATHGENERATION,INPARTICULARFORHIGHSPEEDMACHININGFINISHCUTKEYWORDSFIVEAXISMILLINGTOOLPATHGENERATIONCUTTERPOSTURECHANGEMACHININGSTRIPWIDTH1INTRODUCTIONASTHENEEDFORCOMPLEXCOMPONENTSSUCHASTHREEDIMENSIONAL3DMOULDSANDDIESHASRISEN,SCULPTUREDSURFACEMACHININGHASASSUMEDAMOREANDMOREIMPORTANTROLEINMANUFACTURINGFORTHELASTFEWDECADESTHEEMPLOYMENTOFFIVEAXISNUMERICALCONTROLNCMACHINESINSCULPTUREDSURFACEMACHININGOFFERSNUMEROUSADVANTAGESOVERTHREEAXISMODESUCHASSETUPREDUCTION,FASTMATERIALREMOVALRATES,ANDIMPROVEDSURFACEQUALITYTOMAKETHEBESTUSEOFFIVEAXISMACHINING,HOWEVER,PROBLEMSRELATEDTOCOMPLICATIONANDCOMPLEXITYCAUSEDBYTHETWOADDITIONALROTARYAXESHAVETOBESOLVEDONEOFTHECHALLENGINGTASKSISTOAUTOMATICALLYGENERATEERRORFREETOOLPATHWITHOUTUSERINTERACTIONFORMACHININGSCULPTUREDSURFACESINTHEPROCESSPLANNINGOFFIVEAXISFINISHCUT,THETOOLPATHGENERATIONTASKISTOSELECTATOOLPATHPATTERN,GENERATETHECUTTERCONTACTCCPOINTSTHATSATISFYTHEACCURACYREQUIREMENT,ANDDETERMINETHECUTTER’SPOSTUREORIENTATIONATEVERYCCPOINTWITHOUTCAUSINGANYINTERFERENCEDURINGTHISPROCESS,TOENSURETHEQUALITYOFTHEMACHINEDSURFACE,THESMOOTHDYNAMICSOFCUTTERMOTIONISAMUST,IE,THEPOSTURECHANGEFROMONEPOINTTOTHENEXTMUSTBEMINIMIZEDEXTREMECHANGEINCUTTERPOSTURE,WHICHISNECESSARYFORINTERFERENCEAVOIDANCE,ISAMAJORCAUSEFORTHEUNNATURALMOVEMENTOFTHECUTTERANDWILLLEADTOOVERANDUNDERCUTTINGINFIVEAXISFINISHANDUNDESIRABLEIRREGULARITYOFTHESURFACEAPPEARANCE1,2SOFAR,THEREISLIMITEDREPORTEDWORKONOBTAININGTHECUTTERLOCATIONCLDATAWITHSMOOTHCONTINUOUSCHANGEINCUTTERPOSTURESALONGAPRESETPATHANDCUTTINGDIRECTION1–3,ANDTHEREISNOREPORTEDMETHODTHATCANGENERATECLDATAWITHGLOBALOPTIMIZATIONOFCUTTERMOTIONDYNAMICSWITHRESPECTTOACUTTINGDIRECTIONINFIVEAXISFINISHCUTLLLIYFZHANGHYLILGENGDEPARTMENTOFMECHANICALENGINEERING,NATIONALUNIVERSITYOFSINGAPORE,10KENTRIDGECRESCENT,SINGAPORE119260,SINGAPOREEMAILMPEZYFNUSEDUSGINTJADVMANUFTECHNOL201153699–709DOI101007/S0017001028492EFFICIENCY,ANDINPARTICULAR,SMOOTHPOSTURECHANGESUSINGTHEAMAPSANDSMAPSATALLTHESAMPLEDPOINTSONTHESURFACE,THEOPTIMALTOOLPATHSAREGENERATEDSUCHTHATTHECHANGEINCUTTERPOSTUREISMINIMIZEDWHENPASSINGTHROUGHTHEGENERATEDCCPOINTSBESIDES,MACHININGSTRIPWIDTHBETWEENADJACENTPATHSISALSOCONSIDEREDTOACHIEVEHIGHMACHININGEFFICIENCY3THEACCESSIBILITYMAPOFACUTTERATAPOINTTHEACCESSIBILITYMAPAMAPISDEFINEDINRESPECTTOACUTTERPOSITIONEDATAPOINTONTHEPARTSURFACEITREFERSTOTHEPOSTURERANGEINTERMSOFTHETWOROTATIONALANGLES,ANDWITHINTHISRANGE,THECUTTERDOESNOTHAVEANYINTERFERENCEWITHTHEPARTANDTHESURROUNDINGOBJECTSTHEAMAPEFFECTIVELYCHARACTERIZESTHEACCESSIBILITYOFACUTTERTOAPOINT,WHICHPROVIDESIMPORTANTGEOMETRICINFORMATIONFORCUTTERSELECTIONANDINTERFERENCEFREETOOLPATHGENERATIONABRIEFINTRODUCTIONOFAMAPISGIVENHEREASSHOWNINFIG1A,THELOCALFRAMEXL?YL?ZLORIGINATESATTHEPOINTOFINTERESTPCWITHZLAXISALONGTHESURFACENORMALVECTOR,XLAXISALONGTHESURFACEMAXIMUMPRINCIPALDIRECTION,ANDYLAXISALONGTHESURFACEMINIMUMPRINCIPALDIRECTION21ACUTTERPOSTUREL,ΘMEANSTHATTHECUTTER’SAXISINCLINESCOUNTERCLOCKWISEWITHLABOUTYLAXISANDROTATESAΘABOUTZLAXISTHEAMAPOFTHECUTTERATTHISPOINTISREPRESENTEDINL,ΘDOMAININTHELOCALFRAMEINORDERTOFINDTHEAMAPATTHEPOINT,THEFOURACCESSIBLEPOSTURERANGESBASEDONTHEIRRESPECTIVEINTERFERENCEFREEATTRIBUTES,IE,MACHINEAXISLIMITSML,LOCALGOUGINGLG,REARGOUGINGRG,ANDGLOBALCOLLISIONGC,AREFOUNDFIRSTFORIMPLEMENTATION,THEPARTSURFACETOBEMACHINEDISFIRSTLYSAMPLEDINTOMPOINTSATASPECIFICPOINT,THEFEASIBLERANGEOFΘ–LBASEDONMLISFIRSTCALCULATEDΘISTHENUNIFORMLYSAMPLEDINTOKANGLESATEACHDISCRETEΘ,THEMINIMUMLNEEDEDTOELIMINATELGISFOUNDFORRGANDGC,THERANGEOFLATEACHDISCRETEΘTHATISINTERFERENCEFREEFROMALLTHEREMAININGM?1POINTSISIDENTIFIEDTHEAMAPATTHISPOINTISSIMPLYTHEINTERSECTIONOFTHESEFOURACCESSIBLEPOSTURERANGESSEEFIG1BTHEOVERALLALGORITHMFORFINDINGTHEAMAPFORACUTTERATAPOINTONTHEPARTSURFACEISCALLEDTHECUTTERACCESSIBILITYCAALGORITHMOBVIOUSLY,THECAALGORITHMISNUMERICALINNATUREWITHACOMPUTATIONALCOMPLEXITYOFΟKMFORMOREDETAILSABOUTTHEEVALUATIONOFTHEAMAP,READERSCANREFERTO22ADIRECTAPPLICATIONOFTHISAMAPCONCEPTISFORTHEOPTIMALCUTTERSELECTIONTOFINISHAGIVENSCULPTUREDSURFACE22BYAPPLYINGTHECAALGORITHMTOALLTHESAMPLEDPOINTSONAPARTSURFACE,ONECANJUDGEWHETHERACUTTERCANTRAVERSETHEWHOLESURFACEWITHOUTANYINTERFERENCETHEOPTIMALCUTTERCANTHEREFOREBETHELARGESTAVAILABLECUTTERWITHNONEMPTYAMAPSATALLSAMPLEDSURFACEPOINTS4THESMOOTHNESSMAPOFACUTTERATAPOINTSINCETHEAMAPONLYCHARACTERIZESTHEGEOMETRICPROPERTYOFTHECUTTER’SPOTENTIALCONFIGURATIONATAPOINT,ITISNECESSARYTOADDTHEDYNAMICPROPERTYOFTHECUTTERATTHEPOINTTOCOMPLETETHEINFORMATIONSETTHEDYNAMICPROPERTYOFCUTTERISACOMPLEXISSUEINVOLVINGMANYFACTORS,EG,FEEDRATE,CUTTINGLOAD,ANDPATHSMOOTHNESSSINCETHISWORKFOCUSESONFINISHINGTOOLPATHPLANNING,ONLYPATHSMOOTHNESSISCONSIDERED,WHICHISMEASUREDBYTHEPOSTURECHANGERATEPCROFTHECUTTERATTHEPOINTGIVENACCPOINTPIANDNEXTCCPOINTPI1ONAPATH,PCRIISDEFINEDASPCRI?JVIT1?VIJJPIT1?PIJD1TWHEREVIISTHEUNITVECTOROFCUTTERAXISALONGITSPOSTUREΘI,LIATPIINTHEGLOBALFRAMEBEFOREACUTTINGDIRECTIONISSELECTED,ITISNECESSARYTOOBTAINTHEPCRSALONGALLPOSSIBLECUTTINGDIRECTIONS,WHICHISCALLEDTHESMOOTHNESSATHECUTTERINTHELOCALFRAMEBTHEAMAPATPCXLYLPCZLΛΘFIG1THECUTTERAMAPATAPOINTONTHEPARTSURFACEATHECUTTERINTHELOCALFRAMEBTHEAMAPATPCINTJADVMANUFTECHNOL201153699–709701

簡介：ELSEVIERJOURNALOFMATERIALSPROCESSINGTECHNOLOGY681199725726RJOURNALOFMATERIALSPOROCESS1NGTECHNOBGYADVANCESINTHEPREDSIONMACHININGOFSMALLDEEPHOLESZMWANGEOEZUGWU“,,DSUB“SCHOOLOFENGKWERINGVSTEMSANDDEIGN,SOUTHBANKUNILCRSIO,,LOMTMTSEI04,4,UKHDEPARTMENT/AHCHANICALENGINEERING,VOLTINGHAMTREML_TIRCRSITVVOTLHTGHAMNGI4BU,UKRECEIVED20OCTOBER1995ABSTRACTBASEDONTHEANALYSISOFTHEFORCESACTINGONADEEPHOLEDRILLINGHEADANDTHEUNDERSTANDINGOFTHEBURNISHINGACTIONOFTHECARBIDEGUIDEPADS,ANINVESTIGATIONOFTHEPRECISIONDRILLINGMETHODOFSMALLDEEPHOLESINDIFFICULTTOCUTMATERIALSWASCARRIEDOUTTHEINFLUENCEOFTOOLGEOMETRYANDCUTTINGPARAMETERSCUTTINGSPEEDSANDFEEDRATESONTHESURFACEQUALITYOFTHEDRILLEDHOLESWASSTUDIED,THERESULTSOFWHICHINDICATINGTHAT1020MMSMALLDEEPHOLESOFHIGHTOLERANCEGRADESIT7TOIT9ANDLOWERSURFACEROUGHNESSVALUESRDINTHERANGEOF0216TJMCANBEACHIEVEDANDTHEPROBLEMOFAXIALDEVIATIONOFHOLESDRILEDMINIMISEDUNDEROPTIMISEDCUTTINGPARAMETERSBYMEANSOFTHENEW/IMPROVEDBTADRILL“,,1997ELSEVIERSCIENCESAKEYWORDSDIFFICULTTOCUTMATERIALSSMALLDEEPHOLESPRECISIONDRILLINGSTABILITYBURNISHINGACTION1INTRODUCTIONDEEPHOLEMACHININGISRECOGNISEDASADIFFICULTPROCESSDUETOTHECONFINEDCUTTINGSPACEANDTHEPOORCUTTINGCONDITIONS,DIFFICULTCHIPBREAKINGANDCHIPREMOVAL,ANDCANTILEVEREDCUTTINGACTION,ASWELLASPOORMACHININGSYSTEMSTIFFNESSFRAZAOETAL1POINTEDOUTTHAT“THEMACHININGOFHOLESOFHIGHLENGTHTODIAMETERRATIOTOHIGHSTANDARDSOFSIZE,PARALLELISM,STRAIGHTNESSANDSURFACEFINISHHASALWAYSPRESENTEDPROBLEMS“LOWMACHINABILITYOFDIFFICULTTOCUTMATERIALSMAKESITMOREDIFFICULTTOACHIEVETHEPROCESSEFFICIENTLYANDECONOMICALLY,ESPECIALLYINTHESOLIDDRILLINGOFSMALLDEEPHOLESCURRENTLY,THEREARETHREEMAINTECHNIQUESAVAILABLEFORDEEPHOLEDRILLING,IETHEGUN,THEBTAFROMBORINGTREPANNINGASSOCIATIONANDTHEEJECTORDRILLINGSYSTEMS2,3THEBASICPRINCIPLESOFEACHPROCESSARETHESAME,BUTTHEBTAPROCESSINITIALLYDEVELOPEDINGERMANYDURINGWORLDWARTWOREPRESENTSTHEMOSTECONOMICALMETHODFORDEEPHOLEDRILLINGWITHHIGHLENGTHTODIAMETERRATIO4THEBTAPROCESSCANCOVERALARGERANGEOFBOREDIAMETERS6750RAMANDHIGHLENGTHTODIAMETERRATIOSUPTO1005SMALLBTATOOLSUNDER20RAMNORMALLYCONSISTOFASINGLECUTTINGTIPWITHTWOCORRESPONDINGAUTHORFAX44171815769909240136/97/1700?1997ELSEVIERSCIENCESAALLRIGHTSRESERVEDPIIS0924013696000295SELLPILOTINGCARBIDEPADSFIG1WHENDRILLINGWITHTHESELFPILOTINGTOOLINTEGRATEDWITHAHIGHPRESSURECOOLANTSYSTEMWHICHFLUSHESTHECHIPBACKTHROUGHTHEINTERIOROFTHEBORINGBAR,THECUTTINGFORCESGENERATEDATTHECUTTINGEDGESAREBALANCEDBYGUIDEPADSRUBBINGAGAINSTTHEBOREWALLTHISMEANSTHATTHEREISBURNISHINGACTIONOFTHEGUIDEPADSONTHEWALLOFTHEMACHINEDHOLEITISDUETOTHEBURNISHINGANDFIXEDSIZEACTIONOFTHEGUIDEPADSCOMBINEDWITHINTERIORCHIPREMOVALMETHODOFTHEBTAPROCESSANDTHEUSEOFANEXTREMELYRIGIDCYLINDRICALBORINGBARTHATHIGHPRECISIONDEEPHOLESCANISACHIEVEDINONEPASS6,7THISWORKISAIMEDATDEVELOPINGARELIABLEMETHODFORTHEPRECISIONDRILLINGOFSMALLDEEPHOLESINDIFFICULOCUTMATERIALSGUDEPADSE±RDLNGRMCUT±NGLIP/FIGTHEBTADRILLINGHEADZMWANGETAL/JOURNAL/“MATERIALSPROCESSINGTECHNOLOGY680997257261259350030002500200010005003000250020001500ANDBTORQUEAXIALFORCEATTHEOUTEREDGEISOBTAINEDPERIODAWHENTHEMIDDLEEDGEENTERS,THEREALAXIALFORCEINCREASESTOBPOINTWHENTHEINNEREDGEMAKESCONTACTWITHTHEWORKPIECE,ARESULTANTAXIALTBRCEKTWITHAPOINTCISOBSERVEDTHEAXIALFORCEINCREASESSUDDENLYASSOONASTHECOREISCUTANDTHEGUIDEPADSENTERTHEMACHINEDHOLEAFTERASHORTPERIOD,ASTABLEAXIALFORCEFADEVELOPSTHECUTTINGTORQUETANDTHERESULTANTTORQUETAWEREMEASUREDINTHESAMEWAYFIG4BSHOWINGATYPICALTORQUETIMECURVEASSHOWNINFIG2,THESUPPORTINGFORCEACTINGONEACHOFTHEGUIDESDEPENDSMAINLYONBOTHTHEMAGNITUDEANDTHEDIRECTIONOFFANDTHEPOSITIONALANGLESQHANDQ2OFTHEGUIDEPADSATQH87°ANDQ2183THESUPPORTINGFORCEONGNEARLYEQUALSTHECUTTINGFORCEF,WHILSTTHEFORCEONG2ISABOUT20OFTHECUTTINGFORCEF10THEREFOREITCANBECONSIDEREDTHATTHEFIRSTPADGBEARSTHECUTTINGIRCEF,WHILETHESECONDPADG2DETERMINESTHEDIAMETEROFTHEDRILLEDHOLE2EXPERIMENTALPROCEDURESBASEDONTHECAREFULINVESTIGATIONOFTHEBTADEEPHOLEDRILLINGPROCESS,INCLUDINGTHEANALYSISOFTHEFORCESACTINGONTHEDEEPHOLEDRILLINGHEAD,THEOBSERVATIONOFTHEDRILLINGSTABILITYANDTHEUNDERSTANDINGOFTHEBURNISHINGACTIONOFTHEGUIDEPADS,ANIMPROVED16THESTUDYINDICATESTHAT1020MMSMALLDEEPHOLESOFHIGHTOLERANCEGRADESIT7TOIT9FIG6ANDLOWERSURFACEROUGHNESSVALUESRAINTHERANGEOF02TO16PMFIG7CANBEACHIEVEDUNDERTHEOPTIMISEDCUTTINGPARAMETERSBYMEANSOFTHENEW/IMPROVEDTYPEOFBTADRILL31EFFECTOFTOOLGEOMETRYEXPERIMENTALINVESTIGATIONOFDEEPHOLEDRILLINGINDIFFICULTTOCUTMATERIALCRNI3MOVINDICATESTHATTHEGOODSURFACEFINISH,HIGHDIMENSIONACCURACYANDTHEIMPROVEDRUNOUTOFTHEMACHINEDDEEPHOLESCANBEATTRIBUTEDPARTLYTOTHEHIGHSTABILITYOFTHEDRILLINGHEADANDTHEIMPROVEDBURNISHINGACTIONOFTHEGUIDEPADSTHEHIGHSTABILITYOFTHEDRILLINGHEADISACHIEVEDBYARRANGINGTHETWOCARBIDEPADSATPOSITIONALANGLESOF8095°AND180190°ANDINSERTINGANASSISTANTPILOTINGCARBIDEPADAT270275°CLOCKWISEFROMTHECUTTIFFGEDGESTHEVIBRATIONRESISTINGPADSATTHEBACKOFDRILLINGHEADFURTHERPROMOTETHESTABILITYOFTHEDRILLINGHEADDUETOTHEIRDAMPINGACTIONANOTHERSIGNIFICANTCONTRIBUTIONOFTHESOFTPADSISTOSUPPRESSTHEAXIALDEVIATIONOFHOLE,WHICHLATTERISONEOFTHEMAINPROBLEMSFACEDINDEEPHOLEDRILLING,ESPECIALLYINDIFFICULTTOCUTMATERIALSWHENASLENDERBORINGBARBEARSAHIGHAXIALFORCE,BENDINGOFTHEBAROCCURSANDTHEDRILLINGHEADINTHESUPPORTPILOTINGBUSHINCLINESINTHISCASE,THEMISGUIDANCEOFTHEDRILLINGHEADWILLENCOURAGETHERUNOUTOFTHEHOLETHEINTRODUCTIONOFTHEVIBRATIONRESISTINGPADSRESULTSINASTABLECANTILEVEREDGUIDINGPROCESSANDELIMINATESORRAIN