簡介：ANEMBEDDEDSOFTWARERECONFIGURABLECOLORSEGMENTATIONARCHITECTUREFORIMAGEPROCESSINGSYSTEMSGRIGORIOSCHRYSOSA,?,APOSTOLOSDOLLASA,NIKOLAOSBOURBAKISA,BATECHNICALUNIVERSITYOFCRETE,ECEDEPT,CHANIA,CRETE,GREECEBWRIGHTSTATEUNIVERSITY,ENGRCOLLEGEATRCENTER,DAYTON,OH45435,USAARTICLEINFOARTICLEHISTORYAVAILABLEONLINE17DECEMBER2011KEYWORDSRECONFIGURABLEARCHITECTURESIMAGESEGMENTATIONEMBEDDEDSYSTEMSABSTRACTIMAGESEGMENTATIONISONEOFTHEFIRSTIMPORTANTANDDIFFICULTSTEPSOFIMAGEANALYSISANDCOMPUTERVISIONANDITISCONSIDEREDASONEOFTHEOLDESTPROBLEMSINMACHINEVISIONLATELY,SEVERALSEGMENTATIONALGORITHMSHAVEBEENDEVELOPEDWITHFEATURESRELATEDTOTHRESHOLDING,EDGELOCATIONANDREGIONGROWINGTOOFFERANOPPORTUNITYFORTHEDEVELOPMENTOFFASTERIMAGE/VIDEOANALYSISANDRECOGNITIONSYSTEMSINADDITION,FUZZYBASEDSEGMENTATIONALGORITHMSHAVEESSENTIALLYCONTRIBUTEDTOSYNTHESISOFREGIONSFORBETTERREPRESENTATIONOFOBJECTSTHESEALGORITHMSHAVEMINORDIFFERENCESINTHEIRPERFORMANCEANDTHEYALLPERFORMWELLTHUS,THESELECTIONOFONEALGORITHMVSANOTHERWILLBEBASEDONSUBJECTIVECRITERIA,OR,DRIVENBYTHEAPPLICATIONITSELFHERE,ALOWCOSTEMBEDDEDRECONFIGURABLEARCHITECTUREFORTHEFUZZYLIKEREASONINGSEGMENTATIONFRSMETHODISPRESENTEDTHEFRSMETHODHASTHREESTAGESSMOOTHING,EDGEDETECTIONANDTHEACTUALSEGMENTATIONTHEINITIALSMOOTHINGOPERATIONISINTENDEDTOREMOVENOISETHESMOOTHERANDEDGEDETECTORALGORITHMSAREALSOINCLUDEDINTHISPROCESSINGSTEPTHESEGMENTATIONALGORITHMUSESEDGEINFORMATIONANDTHESMOOTHEDIMAGETOFINDSEGMENTSPRESENTWITHINTHEIMAGEINTHISWORKTHEFRSSEGMENTATIONALGORITHMWASSELECTEDDUETOITSPROVENGOODPERFORMANCEONAVARIETYOFAPPLICATIONSFACEDETECTION,MOTIONDETECTION,AUTOMATICTARGETRECOGNITIONATRANDHASBEENDEVELOPEDINALOWCOST,RECONFIGURABLECOMPUTINGPLATFORM,AIMINGATLOWCOSTAPPLICATIONSINPARTICULAR,THISPAPERPRESENTSTHEIMPLEMENTATIONOFTHESMOOTHING,EDGEDETECTIONANDCOLORSEGMENTATIONALGORITHMSUSINGSTRETCHS5000PROCESSORSANDCOMPARESTHEMWITHASOFTWAREIMPLEMENTATIONUSINGTHEMATLABTHENEWARCHITECTUREISPRESENTEDINDETAILINTHISWORK,TOGETHERWITHRESULTSFROMSTANDARDBENCHMARKSANDCOMPARISONSTOALTERNATIVETECHNOLOGIESTHISISTHEFIRSTSUCHIMPLEMENTATIONTHATWEKNOWOF,HAVINGATTHESAMETIMEHIGHTHROUGHPUT,EXCELLENTPERFORMANCEATLEASTINSTANDARDBENCHMARKSANDLOWCOST?2011ELSEVIERBVALLRIGHTSRESERVED1INTRODUCTION11SEGMENTATIONMANYCOMPUTERVISION,PATTERNRECOGNITION,IMAGEANALYSISANDOBJECTEXTRACTIONSYSTEMSHAVEBEENDEVELOPEDDURINGTHELASTTHIRTYYEARSATTHESAMETIME,FUZZYANDSEMIFUZZYCLUSTERINGALGORITHMSHAVEBEENALSOPRESENTEDFORTHEEXTRACTIONANDRECOGNITIONOFANOBJECT’SFEATURESINORDERFORTHESESYSTEMSANDALGORITHMSTOBESUCCESSFULTHEYGENERALLYHAVETOSTARTWITHAROBUSTSMOOTHINGAND/ORSEGMENTATIONTECHNIQUETHUS,IMAGESEGMENTATIONISANIMPORTANTSTARTINGSTEPFORALMOSTALLVISIONANDPATTERNRECOGNITIONMETHODOLOGIESSEVERALSTUDIESHAVEBEENDONETOCATEGORIZESEGMENTATIONINTOCLASSESBASEDONCHARACTERISTICS,SUCHASTHRESHOLDINGORCLUSTERING,EDGEDETECTION,REGIONGROWING/MERGINGANDOTHERS1–3INPARTICULAR,LEEANDCHUNG4SHOWEDTHATTHRESHOLDINGWOULDUSUALLYPRODUCEGOODRESULTSINBIMODALIMAGESONLY,WHERETHEIMAGESCOMPRISEOFONLYONEOBJECTANDITSBACKGROUNDHOWEVER,WHENTHEOBJECTAREAISSMALLCOMPAREDTOTHEBACKGROUNDAREA,ORWHENBOTHTHEOBJECTANDBACKGROUNDHAVEABROADRANGEOFGRAYLEVELS,SELECTINGAGOODTHRESHOLDISDIFFICULTANOTHERWEAKNESSOFTHISTECHNIQUEOCCURSWHENMULTIPLEOBJECTSAREPRESENTWITHINTHEIMAGEINSUCHCASES,FINDINGSHARPVALLEYSWITHINTHEHISTOGRAMISFURTHERCOMPLICATED,ANDSEGMENTATIONRESULTSMAYBEVERYPOOREDGEDETECTIONISANOTHERAPPROACHASSOCIATEDTOIMAGESEGMENTATION5ANEDGEISDEFINEDASALOCATIONWHEREASHARPCHANGEINGRAYLEVELORCOLORISDETECTEDHOWEVER,INTHISMETHODITISDIFFICULTTOMAINTAINTHECONTINUITYOFTHEDETECTEDEDGESASEGMENTMUSTALWAYSBEENCLOSEDBYACONTINUOUSEDGEREGIONGROWINGORMERGINGISATHIRDAPPROACHFORIMAGESEGMENTATION6INTHISCASE,LARGE,EASYTOFINDCONTINUOUSREGIONSORSEGMENTSAREDETECTEDFIRSTAFTERWARDS,SMALLREGIONSMAYBEMERGEDBYUSINGHOMOGENEITYCRITERIA7,8ONEDISADVANTAGEOFREGIONGROWINGANDMERGINGISTHEINHERENTLYSEQUENTIAL01419331/SEEFRONTMATTER?2011ELSEVIERBVALLRIGHTSRESERVEDDOI101016/JMICPRO201112004?CORRESPONDINGAUTHOREMAILADDRESSESCHRYSOSMHLTUCGRGCHRYSOS,DOLLASMHLTUCGRADOLLAS,NIKOLAOSBOURBAKISWRIGHTEDUNBOURBAKISMICROPROCESSORSANDMICROSYSTEMS362012215–231CONTENTSLISTSAVAILABLEATSCIVERSESCIENCEDIRECTMICROPROCESSORSANDMICROSYSTEMSJOURNALHOMEPAGEWWWELSEVIERCOM/LOCATE/MICPRO2THEFRSSEGMENTATIONMETHODOLOGYSEGMENTATIONISAPROCESSUSEDTOFACILITATETHEEXTRACTIONOFOBJECTSTHATFORMANIMAGETHEFRSMETHODOLOGY,WHICHISSTUDIEDINTHISPAPER,CONSISTSOFTHREESTEPSPRIORTOTHERECOGNITIONITSELFSMOOTHING,EDGEDETECTIONANDCOLORSEGMENTATIONTHEDATAFLOWOFSEGMENTATIONPROCESSISDESCRIBEDINFIG1INTHISWORK,ASWILLBESHOWNBELOW,THEHISHUE,INTENSITY,SATURATIONMODELISUSED,FROMORIGINALRGBIMAGES,ANAPPROACHWHICHISQUITETYPICALANDHASBEENSHOWNINLITERATURESEESECTION1TOWORKWELL21SMOOTHINGALGORITHMTHEIMAGESCONTAINNOISEINTRODUCEDEITHERBYTHECAMERAORBECAUSEOFTHEIMAGE’STRANSMISSIONOVERANOISYMEDIUMINEITHERCASE,THENOISEMUSTBEREMOVEDBEFOREANYFURTHERIMAGEPROCESSINGISAPPLIEDTHEMOSTCOMMONWAYOFNOISEREMOVALISTHEUSEOFFILTERSANIMPORTANTCONCEPTFORASMOOTHINGALGORITHMISTHENEIGHBORHOODBETWEENTWOPIXELSTHISALGORITHMALLOWSFORAFUZZYDEGREEOFNEIGHBORHOOD,INWHICHFOREACHNEIGHBORINGPIXELTHEREISTHECORRESPONDINGDEGREEOFNEIGHBORHOOD,ASSHOWNINFIG2EACHPIXEL’SCOLORISCOMPAREDWITHTHECOLOROFEACHOFITSNEIGHBORINGBLOCKS,ASSHOWNINFIG3THESIZEOFBLOCKSFOROURIMPLEMENTATIONWAS3?3,WHICHRESULTSTOASTRONGSMOOTHINGOFTHEIMAGETHEAVERAGECOLORFOREACHOFTHENEIGHBORINGBLOCKSWASCALCULATEDTAKINGINTOACCOUNTTHENEIGHBORHOODMEMBERSHIPFUNCTIONASSHOWNINTHEEQ1FORSMOOTHING,THECOLORCONTRASTBETWEENTHECENTERPIXELANDALLOFTHESURROUNDINGBLOCKSMUSTBEMEASUREDTHECOLORCONTRASTBETWEENTHEPIXELI,JANDTHEBLOCKBISTHEEUCLIDEANDISTANCEINTHERGBDOMAINASSHOWNINTHEFOLLOWINGEQUATIONCIJB?PPT3Q?PPKT3S?KLSQ?CSQPPT3Q?PPKT3S?KLSQD1TCONTRASTIJB?DR2?R1T2TDG2?G1T2TDB2?B1T2D2TTHESTEPSOFTHESMOOTHINGALGORITHMTHATWEREIMPLEMENTEDINTHISWORKARESHOWNINFIG4ANDTHEYAREPRESENTEDANALYTICALLYIN2322EDGEDETECTIONEDGEDETECTIONISTHEPROCESSOFTHELIMITSPECIFICATIONOFTHEOBJECTSANIMAGECONSISTSOFHUE,INTENSITYANDSATURATIONREPRESENTEDASH,I,ANDSRESPECTIVELYAREONESETOFPARAMETERSTHATAREUSEDTOEVALUATEPIXELS’EDGESTRENGTHWITHINIMAGESTHESEPARAMETERSARECOMPUTEDFROMTHEORIGINALIMAGERGBVALUESBYTHEEQUATIONSBELOWX?049RT031GT02BY?0177RT0812GT0011BZ?001GT099BL?116YY0??13A?500XX0??13?YY0??13“B?200YY0??13?ZZ0??13“I?LS?FFIFFIFFIFFIFFIFFIFFIFFIFFIFFIFFIFFIFFIFFIFFIFFIA2TB2QH?TAN?1BA??SMOOTHINGEDGEDETECTIONCOLORSEGMENTATIONORIGINALIMAGESMOOTHEDIMAGEEDGEIMAGECOLORSEGMENTEDIMAGEFIG1THEDATAFLOWOFFSRALGORITHMFIG2TABLEOFNEIGHBORHOODDEGREELSQGCHRYSOSETAL/MICROPROCESSORSANDMICROSYSTEMS362012215–231217

簡介：中文中文5000字文獻出處文獻出處NEUMANND,GRUENDINGERA,JOHAMM,ETALPILOTCOORDINATIONFORLARGESCALEMULTICELLTDDSYSTEMSC//INTERNATIONALITGWORKSHOPONSMARTANTENNASVDE,201416畢業(yè)設計論文外文資料翻譯學院專業(yè)通信工程（嵌入式系統(tǒng)開發(fā)）學生姓名班級學號戶服務。每個基站中基站天線的數(shù)量是遠大于同時服務的用戶數(shù)，也就是說。我們進一步假設TDD模式中的通信系統(tǒng)和信道是相互支撐的。我們考慮一個堵塞的衰落信道模型。用表示相干塊中用戶K在單元J到基站I的所有天線合成信道增益的矢量。這些矢量成對統(tǒng)計獨立并且每個矢量信道都是滿足零均值的高斯分布和協(xié)方差矩陣。對于每個符號，我們采集單位J到基站I中所有K個用戶的信道矢量作為矩陣HIJ的列。用表示SNR的有效性，用TTR表示可用導頻符號的數(shù)量，也就是可用的正交導頻序列。在假設中，訓練同時發(fā)生在所有蜂窩中，感應也是同步的，在基站中獲取的訓練信號是由下式給出的（1）式中正交列包括單位J中所有K個用戶的導頻序列，而且的入口都是假設獨立分布和零均值和單位方差的高斯分布。Ⅲ信道估計信道估計如果我們在所有蜂窩中重新使用相同的導頻，比如，根據(jù)公式2估計，關聯(lián)獲得的訓練導頻和信號。在基站I中有著最小方形估計的一致性，在基站中的是白噪聲。由于中的正交行向量，所以轉(zhuǎn)移噪聲矩陣還是滿足零均值和單位方差的獨立同分布。我們注意到及即使在每個蜂窩中重新使用了相同的導頻，用戶的導頻分配還是影響了信道估計。分配的模型由公式公式建模得到，其中是一個包含著一堆標準正交導頻序列的單一矩陣，是一個描述符合的蜂窩J的導頻分配。

簡介：PILOTCOORDINATIONFORLARGESCALEMULTICELLTDDSYSTEMSDAVIDNEUMANN,ANDREASGR¨UNDINGER,MICHAELJOHAM,ANDWOLFGANGUTSCHICKASSOCIATEINSTITUTEFORSIGNALPROCESSING,TECHNISCHEUNIVERSIT¨ATM¨UNCHEN,80290MUNICH,GERMANY{DNEUMANN,GRUENDINGER,JOHAM,UTSCHICK}TUMDEABSTRACTPILOTCONTAMINATIONLIMITSTHEPERFORMANCEOFAMULTICELLTIMEDIVISIONDUPLEXSYSTEMWITHALARGENUMBEROFBASESTATIONANTENNASWESTUDYTHEPOTENTIALBENEFITSOFCOORDINATIONDURINGTHETRAININGPHASEANDWEPROPOSEEFFICIENTALGORITHMSFORPRACTICALSYSTEMSOURDERIVATIONSAREBASEDONRESULTSFROMASYMPTOTICANALYSISANDTHEPRACTICALRELEVANCEISDEMONSTRATEDBYSIMULATIONSWITHREALISTICSYSTEMPARAMETERSIINTRODUCTIONRECENTLY,THEREHASBEENANINCREASINGINTERESTINCELLULARNETWORKSWITHALARGENUMBEROFBASESTATIONANTENNASTHISSOCALLEDMASSIVEMIMOCONCEPTPROMISESHIGHGAINSWITHVERYSIMPLESIGNALPROCESSINGMETHODS1,2THEHIGHNUMBEROFANTENNASMAKESCHANNELESTIMATIONANDFEEDBACKVERYCOSTLYINAFREQUENCYDIVISIONDUPLEXFDDSYSTEMTHUS,MOSTWORKSONTHISTOPICASSUMETIMEDIVISIONDUPLEXTDDSYSTEMS,WHERETHEESTIMATIONOFTHECHANNELTAKESPLACEINANUPLINKTRAININGPHASE1,3–5THATIS,THERESOURCESSPENTONPILOTSDEPENDONTHENUMBEROFSERVEDUSERS,BUTNOTONTHENUMBEROFANTENNASATTHEBASESTATIONFORTHEVERYHIGHANTENNAGAINSINTHESESYSTEMS,THEPERFORMANCEISSEVERELYDEGRADEDBYCHANNELESTIMATIONERRORSDUETOINTERCELLINTERFERENCEINTHETRAININGPHASE,SOCALLEDPILOTCONTAMINATIONITCANBESHOWNTHATTHISINTERFERENCEULTIMATELYLIMITSTHEPERFORMANCEFORUNCOORDINATEDBASESTATIONSWITHAVERYHIGHNUMBEROFANTENNASANDWITHFAVORABLEPROPAGATIONCONDITIONS,IE,INDEPENDENTLYDISTRIBUTEDCHANNELCOEFFICIENTSFOREACHANTENNA1,3,6–9AFEWMETHODSHAVEBEENPROPOSEDTOTACKLETHECONTAMINATIONISSUEINTHEUNCOORDINATEDCASE4,9–11INTHISWORK,WESTUDYTHECOORDINATIONOFPILOTSINTHEUPLINKTRAININGPHASEPREVIOUSWORKONTHISSUBJECTHASBEENDONEIN12,WHEREASTRAIGHTFORWARDGREEDYALGORITHMISPROPOSED,BASEDONTHECHANNELESTIMATIONERRORASPERFORMANCEMETRICTHISAPPROACHISBASEDONASPECIFICSPATIALCHANNELMODELTHATLEADSTOLOWRANKCHANNELCOVARIANCEMATRICESINCONTRASTTOTHISWORK,WEUSERESULTSFROMTHEASYMPTOTICANALYSESIN3AND5TOFORMULATEACOMBINATORIALNETWORKUTILITYMAXIMIZATIONNUMPROBLEMWITHRESPECTTOTHECOORDINATIONSTRATEGYTHUS,OURAPPROACHCANHANDLEARBITRARYCOVARIANCEMATRICESANDWECANSHOWANIMPROVEDPERFORMANCEEVENWHENTHECOVARIANCEMATRICESARESCALEDIDENTITIESWEANALYZEPOSSIBLEBENEFITSFROMPILOTCOORDINATIONBYANOPTIMALALGORITHMBASEDONEXHAUSTIVEENUMERATIONANDPROVIDEEFFICIENTALGORITHMSFORTRAININGCOORDINATIONINPRACTICALSYSTEMSIISYSTEMMODELWECONSIDERACELLULARNETWORKWITHLBASESTATIONS,WHEREEACHBASESTATIONHASMTRANSMITANTENNASANDSERVESKSINGLEANTENNAUSERSTHENUMBEROFBASESTATIONANTENNASISSIGNIFICANTLYLARGERTHANTHENUMBEROFSIMULTANEOUSLYSERVEDUSERSPERBASESTATION,IE,K?MWEFURTHERASSUMETHATTHECOMMUNICATIONSYSTEMISINTDDMODEANDTHATCHANNELRECIPROCITYHOLDSWECONSIDERABLOCKFADINGCHANNELMODELLETHIJK∈CMDENOTETHEVECTOROFCOMPLEXCHANNELGAINSFROMUSERKINCELLJTOALLANTENNASOFBASESTATIONIINONECOHERENCEBLOCKTHESEVECTORSAREPAIRWISESTATISTICALLYINDEPENDENTANDEACHVECTORCHANNELISGAUSSIANDISTRIBUTEDWITHZEROMEANANDCOVARIANCEMATRIXRIJK∈CMMFOREASEOFNOTATION,WECOLLECTTHECHANNELVECTORSOFALLKUSERSINCELLJTOTHEBASESTATIONIASCOLUMNSOFTHEMATRIXHIJLETΡTRDENOTETHEEFFECTIVETRAININGSNRANDTTRTHENUMBEROFAVAILABLEPILOTSYMBOLS,IE,THEAVAILABLENUMBEROFORTHOGONALPILOTSEQUENCESUNDERTHEASSUMPTIONSTHATTHETRAININGTAKESPLACESIMULTANEOUSLYINALLCELLSANDTHERECEPTIONISSYNCHRONIZED,THERECEIVEDTRAININGSIGNALSATBASESTATIONIAREGIVENBYWI√ΡTRL?J1HIJDJNI∈CMTTR1WHERETHEORTHONORMALROWSOFDJ∈CKTTRCONTAINTHEPILOTSEQUENCESFORALLKUSERSINCELLJANDTHEENTRIESOFNIAREASSUMEDTOBEIIDCOMPLEXGAUSSIANDISTRIBUTEDWITHZEROMEANANDUNITVARIANCEIIICHANNELESTIMATIONIFWEREUSETHESAMEPILOTSEQUENCESINALLCELLS,IE,DJˉD?J,ANDCORRELATETHERECEIVEDTRAININGSIGNALSWITHTHEPILOTS,WEOBTAINTHEESTIMATEDUETOˉDHˉDI,YIWI1√ΡTRˉDHL?J1HIJ1√ΡTR?NI2ATBASESTATIONI,THATCOINCIDESWITHTHELEASTSQUARESLSESTIMATEOFTHECHANNELSHII,SINCETHENOISEATTHEBASESTATIONANTENNASISWHITEBECAUSEOFTHEORTHONORMALROWSINˉD,THETRANSFORMEDNOISEMATRIX?NINˉDHSTILLHASIIDENTRIESWITHZEROMEANANDUNITVARIANCEWENOTETHAT,EVENIFWEREUSETHESAMEPILOTSEQUENCESINEACHCELL,THEASSIGNMENTOFTHEPILOTSTOTHEUSERSINFLUENCESΓULIK?1MTRΦIK,IK?21ΡULM1MTRΦIK,IK1M?J,M1MTRRIJMΦIK,IK?J,M∈KΜI,KJ,M?I,K|1MTRΦIK,JM|28ΓDLIKΡIK1MTRΦIK,IK1ΡDLM1M?J,MΡJMTRRJIKΦJM,JM/TRΦJM,JM?J?IMΜI,KΜJ,MΡJM1M|TRΦJM,IK|2/TRΦJM,JM9VIALGORITHMSAEXHAUSTIVEENUMERATIONTOGETANIDEAOFTHEPOTENTIALBENEFITSOFCOORDINATION,WESOLVETHENUMPROBLEMIN10OPTIMALLYBYEXHAUSTIVEENUMERATIONOFALLPOSSIBLEPILOTASSIGNMENTSFORONECELL,THENUMBEROFPOSSIBLEASSIGNMENTSISK?1?K0TTR?KTTRTTR?K11NOTETHATWECANFIXTHEASSIGNMENTOFONECELLWITHOUTAFFECTINGTHEPERFORMANCETHETOTALNUMBEROFPOSSIBLEASSIGNMENTSISTHUS?TTRTTR?K?L?112FORLARGERSYSTEMS,THEENUMERATIONOFALLPOSSIBLEASSIGNMENTSQUICKLYBECOMESCOMPUTATIONALLYINTRACTABLETHUS,WENEEDEFFICIENTALGORITHMSTOMANAGETHETRAININGCOORDINATIONBDEGRADATIONBASEDGREEDYASSIGNMENTTHEFIRSTGREEDYALGORITHMWEINTRODUCEISBASEDONADEGRADATIONMEASUREASPROPOSEDIN13ATEACHITERATIONOFTHEALGORITHM,WEHAVEASETOFUSERSWHICHAREALREADYASSIGNEDTOPILOTSANDASETOFFREEUSERSWHICHSTILLHAVETOBEASSIGNEDINITIALLY,THEUSERSINONECELLAREASSIGNEDRANDOMLY,WHILEALLOTHERUSERSAREFREETHEFIRSTSTEPINEACHITERATIONISTOCALCULATETHEUTILITIESTHATRESULTFROMADDINGEACHOFTHEFREEUSERSTOTHESETOFASSIGNEDUSERSFOREACHPOSSIBLEPILOTTHENFOREACHUSERCALCULATETHEDEGRADATION,IE,AMOUNTOFUTILITYTHATISLOST,WHENTHEUSERONLYGETSTHESECONDBESTPILOTTHEUSERWHICHHASTHEHIGHESTDEGRADATION,IE,THEUSERWHICHISMOSTSENSITIVETOTHECURRENTASSIGNMENT,ISTHENASSIGNEDTOITSBESTPILOTTOCALCULATETHEUTILITIESFORTHEASSIGNEDUSERSTHEUTILITYFUNCTIONHASTOBESEPARABLE,IE,UR?I,KUIKRIK13EACHOFTHEASSIGNEDUSERSISINONEOFTHESETSK1,,KTTRANDTHEPARTIALUTILITYISGIVENBY?UK1,,KTTR?I,K∈?PKPUIKRIK14WHERETHERATESRIKARECALCULATEDUSINGTHEASSIGNMENTSKPFORMALLYWEHAVETHEFOLLOWINGSTEPSLETFDENOTETHESETOFUNASSIGNEDUSERSFOREACHUNASSIGNEDUSERI,K∈FCALCULATETHEOPTIMALPILOTP?IKARGMAXP∈PI?UK1,,KP∪{I,K},,KTTR15ANDDEGRADATIONMEASUREDIK?UK1,,KP?IK∪{I,K},,KTTR?ARGMAXP∈PI,P?P?IK?UK1,,KP∪{I,K},,KTTR16WHEREPIDENOTESTHESETOFSTILLAVAILABLEPILOTSINCELLITHESELECTEDUSERISTHENGIVENBYI?,K?ARGMAXI,K∈FDIK17ANDISASSIGNEDTOITSOPTIMALPILOTKP?I?,K?←KP?I?,K?∪{I?,K?}18PI?←PI?\{P?I?,K?}19F←F\I?,K?20THESESTEPSAREREPEATEDUNTILALLUSERSAREASSIGNED,IE,F?CVARIANCEBASEDGREEDYASSIGNMENTTHEDEGRADATIONBASEDGREEDYALGORITHMSTILLNEEDSALOTOFSINREVALUATIONSFOREACHASSIGNMENTTOFURTHERREDUCECOMPLEXITY,WEPROPOSEANOTHERGREEDYALGORITHM,WHEREWEUSEAHEURISTICTOSELECTTHEMOSTSENSITIVEUSERINAGIVENITERATIONNAMELY,WESELECTTHEUNASSIGNEDUSERWITHTHEWORSTAVERAGECHANNELCONDITIONSINCETHISUSERISMOSTLIKELYTOBEAFFECTEDBYINTERCELLINTERFERENCETHUS,WEAVOIDTHECOSTLYSELECTIONPROCESSOFTHEDEGRADATIONBASEDALGORITHMANDONLYHAVETOSEARCHFORTHEOPTIMALPILOTFORTHESELECTEDUSERDPOSITIONBASEDASSIGNMENTANOTHERPOSSIBLECOORDINATIONSTRATEGYISBASEDONTHEOBSERVATIONTHAT,WITHASIMPLEGEOMETRICPATHLOSSMODEL,WEAKUSERSGENERATEALARGEAMOUNTOFINTERFERENCEINNEIGHBORINGCELLS,WHILESTRONGUSERSGENERATEASMALLAMOUNTOFINTERFERENCETHISMOTIVATESACOORDINATIONSTRATEGYWHICHISONLYBASEDONTHEPOSITIONSOFTHEUSERSANDTHATCANBEAPPLIEDINEACHCELLSEPARATELYLETUSFIRSTASSUMEWEHAVETPK,AONEDIMENSIONALWYNERNETWORKWHERETHECELLSARESEQUENTIALLYINDEXEDWITHI1,,LANDASUFFICIENTLYLARGENUMBEROFUNIFORMLYDISTRIBUTEDUSERSTHECOVARIANCEMATRICESARESCALEDIDENTITIES,

簡介：點擊查看更多通信工程（嵌入式系統(tǒng)開發(fā)）外文翻譯--駕駛員駕駛基礎上的眼睛分析精彩內(nèi)容。

簡介：點擊查看更多電子專業(yè)畢業(yè)設計外文翻譯--基于arm的嵌入式系統(tǒng)的速成樣機平臺設計精彩內(nèi)容。

簡介：2500英文單詞，英文單詞，13萬英文字符萬英文字符,中文中文3900字文獻出處文獻出處BUSONOP,ISWAHYUDIA,RAHMANMAA,ETALDESIGNOFEMBEDDEDMICROCONTROLLERFORCONTROLLINGANDMONITORINGBLOODPUMPJPROCEDIACOMPUTERSCIENCE,2015,72217224DESIGNOFEMBEDDEDMICROCONTROLLERFORCONTROLLINGANDMONITORINGBLOODPUMPPRATONDOBUSONO,ANDIISWAHYUDI,MAKBARAULIARAHMAN,ARIOFITRIANTOABSTRACTTHEPERISTALTICPUMPHASAVITALROLEFORTRANSPORTINGTHEBLOODINTHEEXTRACORPOREALCIRCUITOFHEMODIALYSISMACHINESUCHAPUMPCONSITSOFASTEPPERMOTOR,GEARBOXANDROTORSHAFTWITHTWOROLLERATTACHEDONITTHETRANSPORTEDVOLUMEOFTHEPUMPDEPENDSONTHETUBESEGMENTPRESSEDBYROLLERTHEREFORE,THEDEVIATIONINTHEAMOUNTOFDELIVEREDFLUIDMAYOCCURSITISAPROBLEMWHENITISUSEDINTHEMEDICALDEVICEWITHHIGHACCURACYREQUIREMENTSSUCHASHEMODIALYSISMACHINETOOVERCOMESUCHAPROBLEM,THETRANSPORTOFFLUIDNEEDTOBECONTROLLEDTHEAIMOFTHISWORKWASTODESIGNANDIMPLEMENTOFANEMBEDDEDMICROCONTROLLERFORCONTROLLINGANDMONITORINGFLUIDTRANSPORTINTHEPERISTALTICPUMPANELECTRONICHARDWAREWHICHCONSISTSOFSTM32F4DEVELOPMENTBOARD,STEPPERMOTORDRIVERCIRCUIT,CURRENTSENSOR,MAGNETICSENSORANDARTERIALPRESURESENSORCIRCUITSWASDEVELOPEDINPARALLELWITHEMBEDDEDSOFTWAREDEVELOPMENTTHEEMBEDDEDSOFWARECONSISTSOFROUTINESFORMONITORINGCURRENTSENSOR,PRESSURESENSOR,ANDMAGNETICSENSOR,ASWELLASROUTINESFORPUMPSPEEDCONTROLLERTHESTM32F4BOARDWHICHCONSITSOFARMCORTEXM4FMICROCONTROLLERWASCHOSENBECAUSEOFTHETRADEOFFBETWEENPRICE,PERFORMANCE,ANDLOWPOWERCONSUMPTIONTHEPERISTALTICPUMPSYSTEMWASTESTEDWITHTHEMIXTUREOFWATERANDGLYCERINWITHCERTAINCOMPOSISITIONANDHAVINGVISCOSITYSIMULARTOHUMANBLOODTHEVALIDATIONWASALSOBEPERFORMEDBYCOMPARINGTHERESULTSWITHTHEMEASUREMENTCONDUCTEDUSINGTURBINEFLOWMETERTHERESULTSSHOWSTHATTHEPERFORMANCEOFTHECONTROLLERISWORKINGASEXPECTEDKEYWORDBLOODPUMPCONTROLLER,STEPPERMOTOR,MICROCONTROLLER,EMBEDDEDSYSTEM1INTRODUCTIONEMBEDDEDSYSTEMISPLAYINGANIMPORTANTROLEINOURLIFEITISATECHNOLOGYWHICHMAKESOURLIFEMORECOMFORTABLEANDSAFEMOSTEQUIPMENTSUSEDINOURDAILYACTIVITY,RANGINGFROMMEDICALTOCOMMUNICATIONSYSTEMSAREBUILTSUPONEMBEDDEDSYSTEMTHEREFORE,TODAYECONOMYANDSOCIETYWOULDHEAVILYDEPENDONTHEMEMBEDDEDSYSTEMSHAVEDRIVENTHEPRODUCTINNOVATIONANDDIVERSIFICATIONINTERMSOFFUNCTIONALITY,EFFICIENCY,ANDQUALITY1UNLIKEGENERALPURPOSECOMPUTER,EMBEDDEDSYSTEMSAREBUILTFORSPECIFICAPPLICATIONSTHEAPPLICATIONOFEMBEDDEDSYSTEMSINMEDICALAREAFOREXAMPLEISBECOMINGGREATINTERESTSINCEITALLOWSTHECONSTRUCTIONOFCOMPACTDEVICESTHATHELPTHEPHYSICIANINTREATINGTHEPATIENTSWITHTHEINCREASINGPOWEROFPROCESSINGDEVICEASWELLASTHECOMPLEXITYOFCONTROLSYSTEM,WITHTHECURRENTPROCESSINGSYSTEM,ITISPOSSIBLETOINTEGRATENOTONLYTHECONTROLANDMONITORINGACTIVITIES,BUTALSOTOTAKETHEDECISIONTASKSINTOEMBEDDEDSYSTEMSINADDITION,WITHTHEAVAIBILITYOFAGENERALPURPOSERECONFIGURABLEPLATFORM,ITISPOSSIBLETODEVELOPSUCHACOMPLEXSYSTEMWITHREDUCEDTIMEDEVELOPMENTANDGIVERESULTSHAVINGSIMILARCHARACTERISTICSTOOTHERSIMILAREMBEDDEDSYSTEMSTHISSYSTEMISALSOROBUSTANDFULLYRELIABLE,CAPABLEOFWORKINGFOREXTENDEDPERIODSOFTIMEWITHOUTFAILURETHEPERFORMANCEOFEMBEDDEDCONTROLSYSTEMSCANAFFECTEDBYMANYFACTORSITDEPENDSNOTPERISTALTICPUMPINGSYSTEMTHESYSTEMASSHOWNINFIG1CONSISTSOFPERISTALTICPUMPHEAD,STEPPERMOTORWITHGEARBOX,MOTORDRIVERANDCONTROLLERTHEPUMPHEADWASCONSTRUCTEDBYHOUSINGBODYANDAROTORWITHTWOROLLERSATTACHEDONITANELASTICTUBEISFITTEDTOASEMICIRCULARCHAMBERTHATPARTLYSURROUNDSTHEROTORANDTHETRANSFERREDFLUIDGETSINCONTACTONLYWITHTHEINSIDEOFTHETUBINGANDHENCELOWERINGTHERISKOFCONTAMINATION2THETUBESETISCHANGEDBETWEENEACHPATIENTWHENTHEPUMPISINOPERATION,ONEROLLERWILLPRESSTHEELASTICTUBESEGMENTTOTHEHOUSINGANDFORCETHEBLOODTOMOVEFORWARDBEFORETHEFIRSTROLLERREACHESTHEENDOFTHEHOUSINGANDRELEASETHEMANIFOLD,THESECONDROLLERWILLCLOSETUBESEGMENTPREVENTINGTHEBACKFLOW3,4THECOMPRESSIONOFTHEPLASTICTUBEINDUCESARESISTANCEONTHEPUMPMOTORWHICHWILLCREATEARELATIVELYLARGEPRESSUREDISTURBANCEINTHEFLUIDFLOWINORDERTOGETTHEACCURATEBLOODVOLUMETOBETRANSFERRED,CONTROLLINGTHEPUMPSPEEDISIMPORTANTINHEMODIALYSISTHEPUMPSPEEDCANBESETBYCONTROLLINGTHEAMOUNTOFCURRENTSINJECTEDBYTHEELECTRONICMOTORDRIVERSUCHCURRENTSARECONVERTEDTOPWMDUTYCYCLEFORPOWERINGTHEPHASECOILS,ANDHENCECHANGINGTHEMOTORSPEEDSINTHECASEOFPERISTALTICBLOODPUMP,THEPWMDUTYCYCLESARENOTCONSTANTSINCETHELOADISDYNAMICS,DEPENDINGONTHEAMOUNTTRANSFEREDFLUIDVOLUMEFIGURE2SCHEMATICDIAGRAMOFSTEPPERMOTORDRIVER521ELECTRICALMODELSCHEMATICDIAGRAMOFSTEPPERMOTORDRIVERISSHOWNINFIG2IFTWOPHASESTEPPERMOTORUSEDFORDRIVINGTHEPUMP,EACHOFTHETWOELECTRICALPHASESOFTHESTEPPERMOTORCANBEMODELLEDASRLCIRCUITANDBACKELECTROMOTIVEFORCEASFOLLOW5,THEPHASECURRENTIJCANBEMEASUREBYCURRENTSENSORPLACEDCLOSETOTHECOIL22MECHANICALMODELTHEMECHANICALPARTOFTHEMOTORCANBEMODELLEDASARIGIDBODYSUBJECTTOVARIOUSTORQUES5,

簡介：PROCEDIACOMPUTERSCIENCE722015217–224AVAILABLEONLINEATWWWSCIENCEDIRECTCOM18770509?2015PUBLISHEDBYELSEVIERBVTHISISANOPENACCESSARTICLEUNDERTHECCBYNCNDLICENSEHTTP//CREATIVECOMMONSORG/LICENSES/BYNCND/40/PEERREVIEWUNDERRESPONSIBILITYOFORGANIZINGCOMMITTEEOFINFORMATIONSYSTEMSINTERNATIONALCONFERENCEISICO2015DOI101016/JPROCS201512134SCIENCEDIRECTTHETHIRDINFORMATIONSYSTEMSINTERNATIONALCONFERENCEDESIGNOFEMBEDDEDMICROCONTROLLERFORCONTROLLINGANDMONITORINGBLOODPUMPPRATONDOBUSONOA,ANDIISWAHYUDIB,MAKBARAULIARAHMANB,ARIOFITRIANTOAACENTERFORINFORMATIONANDCOMMUNICATIONTECHNOLOGY,BPPT,KAWASANPUSPIPTEKSERPONG,TANGERANG,INDONESIABDEPARTMENTOFELECTRICALENGINNERING,UNIVERSITYOFALAZHARINDONESIA,JAKARTA,INDONESIAABSTRACTTHEPERISTALTICPUMPHASAVITALROLEFORTRANSPORTINGTHEBLOODINTHEEXTRACORPOREALCIRCUITOFHEMODIALYSISMACHINESUCHAPUMPCONSITSOFASTEPPERMOTOR,GEARBOXANDROTORSHAFTWITHTWOROLLERATTACHEDONITTHETRANSPORTEDVOLUMEOFTHEPUMPDEPENDSONTHETUBESEGMENTPRESSEDBYROLLERTHEREFORE,THEDEVIATIONINTHEAMOUNTOFDELIVEREDFLUIDMAYOCCURSITISAPROBLEMWHENITISUSEDINTHEMEDICALDEVICEWITHHIGHACCURACYREQUIREMENTSSUCHASHEMODIALYSISMACHINETOOVERCOMESUCHAPROBLEM,THETRANSPORTOFFLUIDNEEDTOBECONTROLLEDTHEAIMOFTHISWORKWASTODESIGNANDIMPLEMENTOFANEMBEDDEDMICROCONTROLLERFORCONTROLLINGANDMONITORINGFLUIDTRANSPORTINTHEPERISTALTICPUMPANELECTRONICHARDWAREWHICHCONSISTSOFSTM32F4DEVELOPMENTBOARD,STEPPERMOTORDRIVERCIRCUIT,CURRENTSENSOR,MAGNETICSENSORANDARTERIALPRESURESENSORCIRCUITSWASDEVELOPEDINPARALLELWITHEMBEDDEDSOFTWAREDEVELOPMENTTHEEMBEDDEDSOFWARECONSISTSOFROUTINESFORMONITORINGCURRENTSENSOR,PRESSURESENSOR,ANDMAGNETICSENSOR,ASWELLASROUTINESFORPUMPSPEEDCONTROLLERTHESTM32F4BOARDWHICHCONSITSOFARMCORTEXM4FMICROCONTROLLERWASCHOSENBECAUSEOFTHETRADEOFFBETWEENPRICE,PERFORMANCE,ANDLOWPOWERCONSUMPTIONTHEPERISTALTICPUMPSYSTEMWASTESTEDWITHTHEMIXTUREOFWATERANDGLYCERINWITHCERTAINCOMPOSISITIONANDHAVINGVISCOSITYSIMULARTOHUMANBLOODTHEVALIDATIONWASALSOBEPERFORMEDBYCOMPARINGTHERESULTSWITHTHEMEASUREMENTCONDUCTEDUSINGTURBINEFLOWMETERTHERESULTSSHOWSTHATTHEPERFORMANCEOFTHECONTROLLERISWORKINGASEXPECTED?2015PUBLISHEDBYELSEVIERLTDSELECTIONAND/ORPEERREVIEWUNDERRESPONSIBILITYOFTHESCIENTIFICCOMMITTEEOFTHETHIRDINFORMATIONSYSTEMSINTERNATIONALCONFERENCEISICO2015KEYWORDBLOODPUMPCONTROLLER,STEPPERMOTOR,MICROCONTROLLER,EMBEDDEDSYSTEM?2015PUBLISHEDBYELSEVIERBVTHISISANOPENACCESSARTICLEUNDERTHECCBYNCNDLICENSEHTTP//CREATIVECOMMONSORG/LICENSES/BYNCND/40/PEERREVIEWUNDERRESPONSIBILITYOFORGANIZINGCOMMITTEEOFINFORMATIONSYSTEMSINTERNATIONALCONFERENCEISICO2015219PRATONDOBUSONOETAL/PROCEDIACOMPUTERSCIENCE722015217–2242PERISTALTICBLOODPUMPSYSTEMINTHEEXTRACORPOREALCIRCUITOFHEMODIALYSISMACHINES,THEBLOODFLOWISTRANSPORTEDBYAPERISTALTICPUMPINGSYSTEMTHESYSTEMASSHOWNINFIG1CONSISTSOFPERISTALTICPUMPHEAD,STEPPERMOTORWITHGEARBOX,MOTORDRIVERANDCONTROLLERTHEPUMPHEADWASCONSTRUCTEDBYHOUSINGBODYANDAROTORWITHTWOROLLERSATTACHEDONITANELASTICTUBEISFITTEDTOASEMICIRCULARCHAMBERTHATPARTLYSURROUNDSTHEROTORANDTHETRANSFERREDFLUIDGETSINCONTACTONLYWITHTHEINSIDEOFTHETUBINGANDHENCELOWERINGTHERISKOFCONTAMINATION2THETUBESETISCHANGEDBETWEENEACHPATIENTWHENTHEPUMPISINOPERATION,ONEROLLERWILLPRESSTHEELASTICTUBESEGMENTTOTHEHOUSINGANDFORCETHEBLOODTOMOVEFORWARDBEFORETHEFIRSTROLLERREACHESTHEENDOFTHEHOUSINGANDRELEASETHEMANIFOLD,THESECONDROLLERWILLCLOSETUBESEGMENTPREVENTINGTHEBACKFLOW3,4THECOMPRESSIONOFTHEPLASTICTUBEINDUCESARESISTANCEONTHEPUMPMOTORWHICHWILLCREATEARELATIVELYLARGEPRESSUREDISTURBANCEINTHEFLUIDFLOWFIGURE1PERISTALTICBLOODPUMPSYSTEMINORDERTOGETTHEACCURATEBLOODVOLUMETOBETRANSFERRED,CONTROLLINGTHEPUMPSPEEDISIMPORTANTINHEMODIALYSISTHEPUMPSPEEDCANBESETBYCONTROLLINGTHEAMOUNTOFCURRENTSINJECTEDBYTHEELECTRONICMOTORDRIVERSUCHCURRENTSARECONVERTEDTOPWMDUTYCYCLEFORPOWERINGTHEPHASECOILS,ANDHENCECHANGINGTHEMOTORSPEEDSINTHECASEOFPERISTALTICBLOODPUMP,THEPWMDUTYCYCLESARENOTCONSTANTSINCETHELOADISDYNAMICS,DEPENDINGONTHEAMOUNTTRANSFEREDFLUIDVOLUME21ELECTRICALMODELSCHEMATICDIAGRAMOFSTEPPERMOTORDRIVERISSHOWNINFIG2IFTWOPHASESTEPPERMOTORUSEDFORDRIVINGTHEPUMP,EACHOFTHETWOELECTRICALPHASESOFTHESTEPPERMOTORCANBEMODELLEDASRLCIRCUITANDBACKELECTROMOTIVEFORCEASFOLLOW5,TUTETIRDTTDILJJJWJW????FORJA,B1WHEREMMMAPKTE??SIN??2MMMBPKTE??SIN?3CONTROLLERPERISTALTICPUMPSTEPPERMOTORMOTORDRIVER

簡介：點擊查看更多外文翻譯--為嵌入式系統(tǒng)開發(fā)人員建立虛擬的arm平臺精彩內(nèi)容。

簡介：點擊查看更多外文翻譯---基于高性能嵌入式多核mpsoc的多線程系統(tǒng)設計精彩內(nèi)容。

簡介：點擊查看更多外文翻譯---嵌入式單片機溫度控制器設計精彩內(nèi)容。

簡介：點擊查看更多外文翻譯---關于混泥土破裂有限元分析的嵌入式模型（含word版）精彩內(nèi)容。

簡介：點擊查看更多基于fpga的嵌入式系統(tǒng)設計外文翻譯精彩內(nèi)容。

簡介：點擊查看更多圖片分割外文翻譯--一種可重構(gòu)的嵌入式彩色圖像分割處理軟件系統(tǒng)精彩內(nèi)容。

簡介：點擊查看更多[雙語翻譯]嵌入式系統(tǒng)開發(fā)外文翻譯--大規(guī)模多小區(qū)tdd系統(tǒng)的導頻協(xié)調(diào)精彩內(nèi)容。

簡介：點擊查看更多[雙語翻譯]單片機外文翻譯--嵌入式血泵監(jiān)控微控制器的設計中英全精彩內(nèi)容。