简介:Theobjectiveofthispaperistodevelopaneuralnetwork-basedresidualgeneratortodetectthefaultintheactuatorsforaspecificcommunicationsatelliteinitsattitudecontrolsystem(ACS).First,adynamicmultilayerperceptronnetworkwithdynamicneuronsisused,theseneuronscorrespondtoasecondorderlinearInfiniteImpulseResponse(IIR)filterandanonlinearactivationfunctionwithadjustableparameters.Second,theparametersfromthenetworkareadjustedtominimizeaperformanceindexspecifiedbytheoutputestimatederror,withthegiveninput-outputdatacollectedfromthespecificACS.Then,theproposeddynamicneuralnetworkistrainedandappliedfordetectingthefaultsinjectedtothewheel,whichisthemainactuatorinthenormalmodeforthecommunicationsatellite.Thentheperformanceandcapabilitiesoftheproposednetworkweretestedandcomparedwithaconventionalmodel-basedobserverresidual,showingthedifferencesbetweenthesetwomethods,andindicatingthebenefitoftheproposedalgorithmtoknowtherealstatusofthemomentumwheel.Finally,theapplicationofthemethodsinasatellitegroundstationisdiscussed.
简介:Anewmethodforpoint-polygonclassificationispresentedinthepaper.Thebasicideaofthemethodisfirstadvancedandthentwopropertiesofpolygonareintroduced.Thecriterionsforpoint-polygonclassificationaregivenexplicitly.Theanalysisshowsthatthepossessesperfectfunction,extensivesuitabilityandhighefficiency.
简介:Inthispaper,twoapproachesareusedtosolvethePerspectiveThree-PointProblem(P3P):thesymboliccomputationapproachandthegeometricapproach.Inthesymboliccomputationapproach,weuseWu-Ritt'szerodecompositionalgorithmtogiveacompletetriangulardecompositionfortheP3Pequationsystem.ThisdecompositionprovidesthefirstcompleteanalyticalsolutiontotheP3Pproblem.Inthegeometricapproach,wegivesomepuregeometriccriteriaforthenumberofrealphysicalsolutions.Thecompletesolutionclassificationfortwospecialcaseswiththreeandfourparametersisalsogiven.
简介:AnewmethodforrecognizingChinesecharactersisproposed.Itisbasedontheso-calledfeaturepointsofChinesecharacters.Thefeaturepointsweuseincludethoseonthestrokeofacharacter.i.e.,endpoints.turningpoints,forkpointsandcrosspoints.andthekeypointsonthebackgroundofcharacter.ThismethoddiffersfromthepreviousonesforitcombinesthefeaturepointsonstrokewiththoseonbackgroundanditusesfeaturepointstorecognizeChinesecharactersdirectly.AChinesecharacterrecognitionsystembasedtotop-downdynamicalmatchingoffeaturepointisdeveloped.Thesystemcanrecognizenotonly6763printedsampleSongfontChinesecharactersofsize5.6×5.6mm^2withhighrecognitionrate,butalsothegeneralprintedbooks,magazinesanddocumentswithasatisfactoryrecognitionrateandspeed.
简介:我们在场统一的一个多水平分区为能被设计或在一种含蓄的形式代表的表面重建的代数学的集合表面(MPU-APSS)。代数学的点集合表面(APSS)用本地移动从一套未组织起来的点定义光滑的表面最少平方(MLS)代数学的范围适合。由于本地性质,然而,APSS不为几何学编辑并且当模特儿工作很好。相反,我们的方法基于统一途径的分区为散布的点集合造一个含蓄的近似函数。由使用octree分策略,我们适应地首先为点集合构造本地代数学的范围,然后使用weighting功能一起混合这些本地形状功能。最后,我们从表面计算签署的距离功能的控制错误的近似。另外,我们在场为点使我们的表示合适的一个有效设计操作员设定过滤并且动态点采样。我们为表面重建并且几何学当模特儿例如表面结束表明我们的统一途径的有效性。
简介:Inthispaper,ageometry-basedpointcloudreductionmethodisproposed,andareal-timemobileaugmentedrealitysystemisexploredforapplicationsinurbanenvironments.Weformulateanewobjectivefunctionwhichcombinesthepointreconstructionerrorsandconstraintsonspatialpointdistribution.Basedonthisformulation,amixedintegerprogrammingschemeisutilizedtosolvethepointsreductionproblem.Themobileaugmentedrealitysystemexploredinthispaperiscomposedoftheofttineandonlinestages.Attheofflinestage,webuildupthelocalizationdatabaseusingstructurefrommotionandcompressthepointcloudbytheproposedpointcloudreductionmethod.Whileattheonlinestage,wecomputethecameraposeinrealtimebycombininganimage-basedlocalizationalgorithmandacontinuousposetrackingalgorithm.Experimentalresultsonbenchmarkandrealdatashowthatcomparedwiththeexistingmethods,thisgeometry-basedpointcloudreductionmethodselectsapointcloudsubsetwhichhelpstheimage-basedlocalizationmethodtoachievehighersuccessrate.Also,theexperimentsconductedonamobileplatformshowthatthereducedpointcloudnotonlyreducesthetimeconsumingforinitializationandre-initialization,butalsomakesthememoryfootprintsmall,resultingascalableandreal-timemobileaugmentedrealitysystem.
简介:Itisaresearchsubjectincomputervisionto3Dreconstructionofanobjectrepresentedbyasingle2Dlinedrawing.Previousworkson3Dreconstructionfrom2Dlinedrawingsfocusonobjectswithlines,plane,view,andsoon.Thispapermainlystudiesthe3Dreconstructionfrom2Dlinedrawings.Besides,anewapproachisproposed:itisthatfortheresearchofthepointcoordinatesof2Dlinedrawings,soastoachievetheobjectreconstructionbythereconstructionofpointcoordinates.Thereconstructionprocessincludes:①thecollectionofpointcoordinates(X,Y)of2Dlinedrawings;②thederivationofmathematicalformulaaboutthereconstructionofthepointof2Dlinedrawings,andcalculatingthecorrespondingpointofthe3Dcoordinates;③theregenerationof3Dgraphicswith3Dpoints;④analyzeerrorbytheproportionalofparallelofaxonometricprojection,inordertoprovetheaccuracyofthemethod.
简介:PI控制策略为一个水力的起重机吊杆系统(HCBS)的活跃抑制控制基于模糊集合点weighting列在后面被建议。控制Twovalve的PI控制器,基于列在后面的fuzzyset点weighting和一个有限半综合者(大规模集成电路)包括一个比例的前馈控制控制器,分别地被设计。大规模集成电路习惯于限制输出信号并且在光谱的低频率结束阻止。由使用一个范围照相机和电子反馈控制,在HCBS上抑制的尖端能人工地被调整。HOPSAN和MATLAB/SIMULINK的一种合作控制模拟技术被用于控制器设计。模拟结果证明建议PI控制系统让更少象更快的反应一样射。在操作期间在HCBS上抑制的尖端被改进。