简介:Recentlytherehavebeenresearchesaboutnewefficientnonlinearfilteringtechniques[1]~[3]inwhichthenonlinearfiltersgeneralizeelegantlytononlinearsystemswithouttheburdensomelinearizationsteps.Thus,truncationerrorsduetolinearizationcanbecompensated.ThesefiltersincludetheunscentedKalmanfilter(UKF),thecentraldifferencefilter(CDF)andthedivideddifferencefilter(DDF),andtheyarealsocalledSigmaPointFilters(SPFs)inaunifiedway[4].Forhigherorderapproximationofthenonlinearfunction.ItoandXiong[6]introducedanalgorithmcalledtheGaussHermiteFilter,whichisrevisitedin[5].TheGaussHermiteFiltergivesbetterapproximationattheexpenseofhighercomputationburden,althoughit’slessthantheparticlefilter.TheGaussHermiteFilterisusedasintroducedin[5]withadditionalpruningstepbyaddingthresholdfortheweightstoreducethequadraturepoints.
简介:AnewmethodofunscentedextendedKalmanfilter(UEKF)fornonlinearsystemispresented.ThisnewmethodisacombinationoftheunscentedtransformationandtheextendedKalmanfilter(EKF).TheextendedKalmanfilterissimilartothatinaconventionalEKF.However,ineveryrunningstepoftheEKFtheunscentedtransformationisrunning,thedeterministicsampleiscaughtbyunscentedtransformation,thenposteriormeanofnonlinearityiscaughtbypropagating,buttheposteriorcovarianceofnonlinearityiscaughtbylinearizing.TheaccuracyofnewmethodisalittlebetterthanthatoftheunscentedKalmanfilter(UKF),however,thecomputationaltimeoftheUEKFismuchlessthanthatoftheUKF.
简介:SeveralfiltertechniqueswereavailablefortheGPSpositionestimationproblemofmaneuveringvehiclerangingfromusingdifferentprocessnoisestoInteractiveMultipleModel(IMM).ThelimitationofusingstandardKalmanfiltersislisted.Theperformanceofproposedadaptivefilteriscomparedwiththatofthestandardones,twotypesofdynamicmodelingofthemaneuveringvehicleareused.ThesimulationisbasedonthealmanacdataoftheGPSsatellitestocomputeitsfeasibilityduringthesimulationtimeandpositiononshape8trackwithcontinuousvehiclemaneuvering.Thegoalistoobtaincomputationallyefficientfilterwithreasonableaccuracyforvehicleinmaneuveringsituation.ThefilterproposedisanalternativetothefilterproposedinRef.[1]withlowcomputationalburden.
简介:ThispapermakesaprobeintotheapplicationoftheKalmanfilteringmethodtothedataprocessingofacross-faultmeasurements.Onthebasisofstatisticalregression,themathematicandstochasticmodelsoffiltrationareestablishedbycombiningtheregressionmethodwithKalmanfiltering.Inthefilteringcomputation,notonlytherandomnessoffaultmovementsbutalsothetime-dependentvariationofenvironmentaleffectshavebeentakenintoconsideration.Byuseoftheadaptivefilteringmethod,anestimationofthedynamicnoisevariancematrixisobtainedthroughiteration.Modelsforonemeasuringline(levelinglineorbaseline),twomeasuringlines(bothlevelinglinesorbothbaselines)andfourmeasuringlines(twolevelinglinesandtwobaselines)arederivedandestablishedsystematically.Bymeansofthesemodels,thedataofacross-faultmeasurementscanbeprocesseddynamicallyinreal-timetoprovidethefilteredvaluesofheightdifferencebetweenbenchmarksorbaselinelengthatdifferenttimein
简介:我们核心之一与使用眼睛的人的计算机相互作用(HCI)系统有关发出的工作地址作为输入凝视。这个问题是在任何眼睛存在的传感器,传播和另外的延期基于追踪者的系统,减少它的表演。延期效果能被眼睛运动轨道的精确预言补偿。这篇论文介绍使用人的视觉系统的解剖性质预言眼睛运动轨道的人的眼睛的一个数学模型。眼睛数学模型被转变成一种Kalman过滤器形式在所有眼睛运动类型期间提供连续眼睛位置信号预言。在这篇论文介绍的模型使用在转变在之间期间采用的大脑茎控制性质快(急束勒马)并且慢(固定,追求)眼睛运动。在这篇论文介绍的结果显示在一种Kalman过滤器形式的建议眼睛模型改进眼睛运动预言的精确性并且能够即时表演。除了HCI,有直接眼睛的系统凝视输入,建议眼睛模型能立即被申请在即时凝视偶然的系统的bit-rate/computational减小。
简介:以SINSiGPS组合导航系统为背景,在对Kalman滤波原理和工程应用进行深入分析的基础上,总结了该方法的不足,提出了应用神经网络和模糊推理技术对系统噪声、观测噪声和其相关阵进行直接调控的方法。该方法根据新息和新息方差的变化,实时调整自适应因子,间接改变Kalman滤波器的当前观测量和过去信息的比例关系。仿真结果表明,该算法对模型和噪声干扰有较强的自适应性,能够有效抑制滤波发散,在不损失原有精度的前提下,提高了系统的鲁棒性。
简介:Formultisensorsystems,whenthemodelparametersandthenoisevariancesareunknown,theconsistentfusedestimatorsofthemodelparametersandnoisevariancesareobtained,basedonthesystemidentificationalgorithm,correlationmethodandleastsquaresfusioncriterion.SubstitutingtheseconsistentestimatorsintotheoptimalweightedmeasurementfusionKalmanfilter,aself-tuningweightedmeasurementfusionKalmanfilterispresented.Usingthedynamicerrorsystemanalysis(DESA)method,theconvergenceoftheself-tuningweightedmeasurementfusionKalmanfilterisproved,i.e.,theself-tuningKalmanfilterconvergestothecorrespondingoptimalKalmanfilterinarealization.Therefore,theself-tuningweightedmeasurementfusionKalmanfilterhasasymptoticglobaloptimality.Onesimulationexamplefora4-sensortargettrackingsystemverifiesitseffectiveness.
简介:Targetdynamicsareassumedtobeknowninmeasuringdigitalspeckledisplacement.Useismadeofasimplemeasurementequation,wheremeasurementnoiserepresentstheeffectofdisturbancesintroducedinmeasurementprocess.Fromtheseassumptions,Kalmanfiltercanbedesignedtoreducevarianceofmeasurementnoise.Anopticalandanalysissystemwassetup,bywhichobjectmotionwithconstantdisplacementandconstantvelocityisexperimentedwithtoverifyvalidityofKalmanfilteringtechniquesforreductionofmeasurementnoisevariance.
简介:这研究检验与一个整体Kalman过滤器(EnKF)联合确定的四维的变化吸收系统(4DVAR)为数据吸收生产一条优异混合途径的性能。当在阻止过滤器分叉利用4DVAR时,从使用州依赖者的不确定性的联合吸收计划(E4DVAR)好处由EnKF提供了:4DVAR分析通过费用的最小化生产以后的最大的可能性答案整体不安关于被转变的功能,和产生整体分析能为下一个吸收周期并且作为整体预报的一个基础向前被宣传。这条联合途径的可行性和有效性与模仿的观察在一个理想化的模型被表明。E4DVAR能够在完美模型、有瑕疵模型的情形下面超过4DVAR和EnKF,这被发现。联合计划的性能比为标准EnKF或4DVAR实现的那些对整体尺寸或吸收窗口长度也不太敏感。
简介:Anewhybridwavelet-Kalmanfiltermethodfortheestimationofdynamicsystemisdeveloped,Usingthismethod,thereal-timemultiscaleestimationofthedynamicsystemisimplemented,andtheobservationequationestablishedisfortheobjectsignalitselfratherthanitswaveletdecomposition.Thesimulationresultsshowthatthismethodcanbeusedtoestimatetheobject'sstateofthestackedsystem,whichisthefoundationofmultiscaledatafusion;besidestheperformanceofthenewalgorithmdevelopedintheletterisalmostoptimal.
简介:在地球中的错误方程修理了的IMU(惯性的测量单位)协调第一被介绍。过滤的褪色的Kalman简单地被介绍,它的缺点被分析,然后,适应过滤在IMU/GPS综合航行系统,适应因素被褪色的因素在代替被使用。一个实际例子被给。当在IMU/GPS综合航行系统适用时,结果证明与褪色的因素相结合的适应过滤器有效、可靠。
简介:Basedonthemulti-sensoroptimalinformationfusioncriterionweightedbymatricesinthelinearminimumvariancesense,usingwhitenoiseestimators,anoptimalfusiondistributedKalmansmootherisgivenfordiscretemulti-channelARMA(autoregressivemovingaverage)signals.Thesmoothingerrorcross-covariancematricesbetweenanytwosensorsaregivenformeasurementnoises.Furthermore,thefusionsmoothergiveshigherprecisionthananylocalsmootherdoes.
简介:车道线检测是智能驾驶系统的重要组成部分,它提供了车辆与车道位置关系的信息.针对智能车辆驾驶系统在视觉导航过程中车道线检测的精确性和鲁棒性的问题,提出一种有效的车道线检测方法.首先对原始RGB图像分别进行感兴趣区域设定、逆透视变换、灰度化和阈值处理;然后进行霍夫变换处理,利用斜率和中心点位置筛选检测结果;最后利用卡尔曼滤波对检测到的线段进行跟踪,预测当前车道线位置.实验结果表明,该算法能够有效解决图像中车道线不清晰以及一些干扰遮挡的问题,车道线检测准确率可达94%,具有较好的准确性、鲁棒性和较低的计算复杂度,有利于实时性检测系统的构建.
简介:TheunscentedKalmanfilterisadevelopedwell-knownmethodfornonlinearmotionestimationandtracking.However,thestandardunscentedKalmanfilterhastheinherentdrawbacks,suchasnumericalinstabilityandmuchmoretimespentoncalculationinpracticalapplications.Inthispaper,wepresentanovelsamplingstrongtrackingnonlinearunscentedKalmanfilter,aimingtoovercomethedifficultyinnonlineareyetracking.Intheaboveproposedfilter,thesimplifiedunscentedtransformsamplingstrategywithn+2sigmapointsleadstothecomputationalefficiency,andsuboptimalfadingfactorofstrongtrackingfilteringisintroducedtoimproverobustnessandaccuracyofeyetracking.ComparedwiththerelatedunscentedKalmanfilterforeyetracking,theproposedfilterhaspotentialadvantagesinrobustness,convergencespeed,andtrackingaccuracy.Thefinalexperimentalresultsshowthevalidityofourmethodforeyetrackingunderrealisticconditions.
简介:Channelfrictionisanimportantparameterinhydraulicanalysis.AchannelfrictionparameterinversionmethodbasedonKalmanFilterwithunknownparametervectorisproposed.Numericalsimulationsindicatethatwhenthenumberofmonitoringstationsexceedsacriticalvalue,thesolutionishardlyaffected.Inaddition,KalmanFilterwithunknownparametervectoriseffectiveonlyatunsteadystate.Forthenonlinearequations,computationsofsensitivitymatricesaretime-costly.Twosimplifiedmeasurescanreducecomputingtime,butnotinfluencetheresults.Oneistoreducesensitivitymatrixanalysistime,theotheristosubstituteforsensitivitymatrix.