简介：

简介：Inthispaper,weproposeafacerecognitionapproach-StructedSparseRepresentation-basedclassificationwhenthemeasurementofthetestsampleislessthanthenumbertrainingsamplesofeachsubject.Whenthisconditionisnotsatisfied,weexploitNearestSubspaceapproachtoclassifythetestsample.Inordertoadaptallthecases,wecombinethetwoapproachestoanadaptiveclassificationmethod-Adaptiveapproach.TheadaptiveapproachyieldsgreaterrecognitionaccuracythantheSRCapproachandCRC_RLSapproachwithlowsamplerateontheExtendYaleBdataset.Anditismoreefficientthanothertwoapproaches.

简介：Anoveladaptiveswitchingfilter(ASF)basedondirectionaldetectionisproposedfordenoisingtheimagesthatarehighlycorruptedbyimpulsenoise.Theproposedalgorithmemploysanefficientnoisedetectionmechanism.Itfirstemploysanefficientmethodtoestimatethedifferencesbetweenthecurrentpixelanditsneighborsalignedwith28directions.Thecurrentnoisepixelisreplacedbyamedianorameanvaluewithinanadaptivefilterwindowwithrespecttodifferentnoisedensities.Experimentalresultsshowthattheproposedapproachcannotonlyachieveverylowmiss-detectionratioandfalse-alarmratioevenuptohighnoisecorruption,butalsopreservethedetailedinformationofanimageverywell.

简介：这份报纸为调遣目标追踪学习适应格子和模糊交往的多重模型(AGFIMM)的算法，当集中于固定结构的问题时，不高的多重模型(FSMM)算法费用效率比率和Markov转移交往的概率是困难的确切决定的多重模型(IMM)算法。这个算法认识到适应模型由适应格子调整设定，并且获得由模糊逻辑推理在模型集合匹配度的每个模型。模拟结果证明AGFIMM算法能有效地改进精确性和多重模型算法的费用效率比率，并且对设计应用作为结果合适。

简介：Membranealgorithms(MAs),whichinheritfromPsystems,constituteanewparallelanddistributeframeworkforapproximatecomputation.Inthepaper,amembranealgorithmisproposedwiththeimprovementthattheinvolvedparameterscanbeadaptivelychosen.Inthealgorithm,somemembranescanevolvedynamicallyduringthecomputingprocesstospecifythevaluesoftherequestedparameters.Thenewalgorithmistestedonawell-knowncombinatorialoptimizationproblem,thetravellingsalesmanproblem.Theempiricalevidencesuggeststhattheproposedapproachisefficientandreliablewhendealingwith11benchmarkinstances,particularlyobtainingthebestoftheknownsolutionsineightinstances.Comparedwiththegeneticalgorithm,simulatedannealingalgorithm,neuralnetworkandafine-tunednon-adaptivemembranealgorithm,ouralgorithmperformsbetterthanthem.Inpractice,todesigntheairlinenetworkthatminimizethetotalroutingcostontheCABdatawithtwenty-fiveUScities,wecanquicklyobtainhighqualitysolutionsusingouralgorithm.

简介：Abasicassumptionofmostrecentlyproposedwaveformdesignalgorithmsisthatthetargetimpulseresponseisaknowndeterministicfunctionorastochasticprocesswithaknownpowerspectraldensity(PSD).However,itiswell-knownthatatargetimpulseresponseisneithereasilynoraccuratelyobtained;besidesitchangessharplywithattitudeangles.Bothoftheaforementionedcasescomplicatethewaveformdesignprocess.Inthispaper,anadaptiverobustwaveformselectionmethodforunknowntargetdetectioninclutterisproposed.Thetargetimpulseresponseisconsideredtobeunknownbutbelongstoaknownuncertaintyset.Anadaptivewaveformlibraryisdevisedbyusingasignal-to-clutter-plus-noiseratio(SCNR)-basedoptimalwaveformdesignmethod.Byapplyingtheminimaxrobustwaveformselectionmethod,theoptimalrobustwaveformisselectedtoensurethelowestperformanceboundoftheunknowntargetdetectioninclutter.Resultsshowthattheadaptivewaveformlibraryoutperformsthepredefinedlinearfrequencymodulation(LFM)waveformlibraryontheSCNRbound.

简介：Anadaptiveactuatorfailurecompensationschemeisproposedforattitudetrackingcontrolofspacecraftwithunknowndisturbancesanduncertainactuatorfailures.Anewfeatureofthisadaptivecontrolschemeistheadaptationofthefailurepatternparameterestimates,aswellasthefailuresignalparameterestimates,fordirectadaptiveactuatorfailurecompensation.Basedonanadaptivebacksteppingcontroldesign,theestimatesofthedisturbanceparametersareusedtosolvethedisturbancerejectionproblem.Theunknowndisturbancesarecompensatedcompletelywiththestabilityofthewholeclosed-loopsystem.Theschemeisnotonlyabletoaccommodateuncertainactuatorfailures,butalsorobustagainstunknownexternaldisturbances.Simulationresultsverifythedesiredadaptiveactuatorfailurecompensationperformance.

简介：Theattempttorepresentasignalsimultaneouslyintimeandfrequencydomainsisfullofchallenges.TherecentlyproposedadaptiveFourierdecomposition(AFD)offersapracticalapproachtosolvethisproblem.ThispaperpresentstheprinciplesoftheAFDbasedtime-frequencyanalysisinthreeaspects:instantaneousfrequencyanalysis,frequencyspectrumanalysis,andthespectrogramanalysis.AnexperimentisconductedandcomparedwiththeFouriertransforminconvergencerateandshort-timeFouriertransformintime-frequencydistribution.TheproposedapproachperformsbetterthanboththeFouriertransformandshort-timeFouriertransform.

简介：Inthisarticle,weconsideraclassofkernelquantileestimatorswhichisthelinearcombinationoforderstatistics.Thisclassofkernelquantileestimatorscanberegardedasanextensionofsomeexistingestimators.Theexactmeansquareerrorexpressionforthisclassofestimatorswillbeprovidedwhendataareuniformlydistributed.Theimplementationoftheseestimatorsdependsmostlyonthebandwidthselection.Wethendevelopanadaptivemethodforbandwidthselectionbasedontheintersectionconfidenceintervals(ICI)principle.MonteCarlostudiesdemonstratethatourproposedapproachiscomparativelyremarkable.Weillustrateourmethodwitharealdataset.

简介：Considertheproblemsoffrequency-invariantbeampatternoptimizationandrobustnessinbroadbandbeamforming.Firstly,aglobaloptimizationalgorithm,whichisbasedonphasecompensationofthearraymanifolds,isusedtoconstructthefrequency-invariantbeampattern.Comparedwithsomemethodspresentedrecently,theproposedalgorithmisnotonlyavailabletogettheglobaloptimalsolution,butalsosimpleforphysicalrealization.Meanwhile,arobustadaptivebroadbandbeamformingalgorithmisalsoderivedbyreconstructingthecovariancematrix.Theessenceoftheproposedalgorithmistoestimatethespace-frequencyspectrumusingCaponestimatorfirstly,thenintegrateoveraregionseparatedfromthedesiredsignaldirectiontoreconstructtheinterference-plus-noisecovariancematrix,andfinallycaleulatetheadaptivebeamformerweightswiththereconstructedmatrix.Thedesignofbeamformerisformulatedasaconvexoptimizationproblemtobesolved.Simulationresultsshowthattheperformanceoftheproposedalgorithmisalmostalwaysclosetotheoptimalvalueacrossawiderangeofsignaltonoiseratios.

简介：

简介：Thecoherencecubetechnologyhasbecomeanimportanttechnologyfortheseismicattributeinterpretation,whichextractsthediscontinuitiesoftheeventsthroughanalyzingthesimilaritiesofadjacentseismicchannelstoidentifythefaultform.Thecoherencecubetechnologywhichusesconstanttimewindowlengthscannotbalancetheshallowlayersandthedeeplayers,becausethefrequencybandofseismicdatavarieswithtime.Whenanalyzingtheshallowlayers,thetimewindowwillcrossoveralotofevents,whichwillleadtoweakfocusingabilityandfailuretodelineatethedetails.Whilethetimewindowwillnotbelongenoughforanalyzingdeeplayers,whichwillleadtolowaccuracybecausethecoherencesnearthezeropointsoftheeventsareheavilyinfluencedbynoise.Forsolvingtheproblem,weshouldmakearesearchonthecoherencecubetechnologywithself-adaptivetimewindow.Thispaperdeterminesthesamplepoints’timewindowlengthsinrealtimebycomputingtheinstantaneousfrequencybandswithWaveletTransformation,whichgivesacoherencecomputingmethodwiththeself-adaptivetimewindowlengths.Theresultshowsthatthecoherencecubetechnologywithself-adaptivetimewindowbasedonWaveletTransformationimprovestheaccuracyoffaultidentification,andsupressesthenoiseeffectively.Themethodcombinestheadvantagesoflongtimewindowmethodandshorttimewindowmethod.

简介：Underwatersensornetwork(UWSN)adoptsacousticcommunicationwithmoreenergy-consumptionandlongerpropagation-delay,whichbringgreatchallengestoprotocoldesign.Inthispaper,weproposedlevel-basedadaptivegeo-routing(LB-AGR)protocol.LB-AGRdividestrafficsintofourcategories,androutesdifferenttypesoftrafficinaccordancewithdifferentdecisions.Packetsupstreamtothesinkareforwardedunicasttothebestnext-hopinsteadofbroadcastingtoallneighbornodesasinpresentUWSNroutingprotocols.LB-AGRdefinesanintegratedforwardingfactorforeachcandidatenodebasedonavailableenergy,density,location,andlevel-differencebetweenneighbornodes,whichisusedtodeterminethebestnext-hopamongmultiplequalifiedcandidates.Throughsimulationexperiments,weshowthepromisingperformanceofLB-AGR.

简介：Ahybridsystemofthefuzzyc-means(FCM)clusteringalgorithmandadaptive-two-stagelinearapproximationwaspresentedfornonlineardistortioncancellationofradiofrequency(RF)poweramplifier(PA).Thismechanismcaneffectivelyeliminatenoise,adaptivelymodelPA’sinstantaneouschange,andefficientlycorrectnonlineardistortion.ThisarticleputsforwardtheFCMclusteringalgorithmforclusteringreceivedsignalstoeliminatewhitenoise,andthenusestheadaptive-two-stagelinearapproximationtofittheinversefunctionoftheamplitude’sandphase’snonlinearmappingduringthetrainingphase.Parametersofthelinearfunctionandsimilarityfunctionaretrainedusingthegradient-descentandminimummean-squareerrorcriteria.Theproposedapproach’strainingresultsisdirectlyemployedtoeliminatesamplingsignal’snonlineardistortion.Thishybridmethodisrealizedeasierthanthemulti-segmentlinearapproximationandcouldreducethereceivedsignal’sbiterrorrate(BER)moreefficiently.

简介：在这份报纸，适应backstepping模糊cerebellar-model-articulation-control为活动操纵者的机器人(MMR)的运动/力量控制的神经网络的控制(ABFCNC)系统被建议。由在轨道位置控制器使用ABFCNC，MMR控制系统的未知动力学和参数变化问题被放松。另外，一个适应柔韧的赔偿者被建议消除由近似错误组成的无常，不明确的骚乱。基于追踪的position-ABFCNC设计，适应柔韧的控制策略也为MMR的nonholonomicconstraint力量被开发。适应联机的学习算法的设计被使用Lyapunov稳定性定理获得。因此，建议方法证明那是它能不仅保证稳定性和坚韧性而且MMR的追踪的表演控制系统。建议控制系统的有效性和坚韧性被比较模拟结果验证。