简介:Inthispaper,aformationcontrolalgorithmandanobstacleavoidancecontrolalgorithmformobilerobotsaredevelopedbasedonarelativemotionsensorysystemsuchasapan/tiltcameravisionsystem,withouttheneedforglobalsensingandcommunicationbetweenrobots.Thisisachievedbyemployingthevelocityvariation,insteadofactualvelocities,asthecontrolinputs.Simulationandexperimentalresultshavedemonstratedtheeffectivenessoftheproposedcontrolmethods.
简介:Thispaperpresentsavacuumgripper(asanactuatorofanintelligentmicromanipulator)formicroobjects(withadiameterof100~300μm)assemblytasks.Thegripperiscomposedofavacuumunitandacontrolunit.Thevacuumunitwithaproportionalvalveandapressuresensor,andthecontrolunitwithaPC+MCUtwo-layeredcontrolarchitecturearedesigned.Themechanicalstructure,workflowandmajorprogramsofthemicro-gripperarepresented.
简介:Thispaperpresentsasymmetriccooperationstrategyforwirelesssensornetworks,aimingtoimprovethetransmissionefficiencyofthenetwork.Thecooperationstrategyisimplementedbypartitioningthenodesintoseveralcooperativegroups.Then,ineachgroup,theoptimalcooperativebandwidthallocationisobtainedbasedonRaiffa-KalaiSmorodinskybargainingsolution(RBS).Numericalresultsshowthatthesymmetriccooperationstrategycanimprovethesensornode'stransmissionefficiencydramatically.
简介:Adissipative-basedadaptiveneuralcontrolschemewasdevelopedforaclassofnonlinearuncertainsystemswithunknownnonlinearitiesthatmightnotbelinearlyparameterized.Themajoradvantageofthepresentworkwastorelaxtherequirementofmatchingcondition,I.e.,theunknownnonlinearitiesappearonthesameequationasthecontrolinputinastate-spacerepresentation,whichwasrequiredinmostoftheavailableneuralnetworkcontrollers.Bysynthesizingastate-feedbackneuralcontrollertonaketheclosed-loopsystemdissipativewithrespecttoaquadraticsupplyrate,thedevelopedcontrolschemeguaranteesthattheL2-gainofcontrolledsystemwaslessthanorequaltoaprescribedlevel.Andthen,itisshownthattheoutputtrackingerrorisuniformlyultimatebounded.Thedesignschemeisillustratedusinganumericalsimulation.
简介:Afuzzyneuralnetworkcontrollerforunderwatervehicleshasmanyparametersdifficulttotunemanually.Toreducethenumerousworkandsubjectiveuncertaintiesinmanualadjustments,ahybridparticleswarmoptimization(HPSO)algorithmbasedonimmunetheoryandnonlineardecreasinginertiaweight(NDIW)strategyisproposed.OwingtotherestraintfactorandNDIWstrategy,anHPSOalgorithmcaneffectivelypreventprematureconvergenceandkeepbalancebetweenglobalandlocalsearchingabilities.Meanwhile,thealgorithmmaintainstheabilityofhandlingmultimodalandmultidimensionalproblems.TheHPSOalgorithmhasthefastestconvergencevelocityandfindsthebestsolutionscomparedtoGA,IGA,andbasicPSOalgorithminsimulationexperiments.ExperimentalresultsontheAUVsimulationplatformshowthatHPSO-basedcontrollersperformwellandhavestrongabilitiesagainstcurrentdisturbance.ItcanthusbeconcludedthattheproposedalgorithmisfeasibleforapplicationtoAUVs.
简介:Finstabilizerswithfin-liftfeedbackcontrolcanshieldthemappingerrorofcalculationbetweenthefinangleandfinliftforce,whichisinthefinstabilizerwithfin-anglefeedbackcontrol.Inpractice,therearesometechnicaldifficultiesinliftfinstabilizers,suchasliftforcedetectionandliftforcesensorinstallation,soitcannotachievethegoodantirollingperformance.Therefore,afinstabilizersystemwithfin-lift/fin-angleintegratedcontrolisbroughtforward.Datafusionbasedonwaveletdenoisingtechnologyisemployedinthesystem,whichcombinesliftwithfinanglelocalinformationfromtwosensorswithdifferentfrequencyrangesinordertoeliminateredundantandcontradictoryinformation,andusingcomplementaryinformationtoobtaintherelativeintegrityoftheliftforcesignal.Thesystemmodelisestablishedinthispaper,andthefusionsignalandtheantirollingperformanceofthismodelaresimulatedrespectively.Theresultshowsthatthecontrolsystemcanmeettheantirollingneedindifferentseasituations.
简介:Adaptivecritic(AC)basedcontrollersaretypicallydiscreteand/oryieldauniformlyultimatelyboundedstabilityresultbecauseofthepresenceofdisturbancesandunknownapproximationerrors.Acontinuous-timeACcontrollerisdevelopedthatyieldsasymptotictrackingofaclassofuncertainnonlinearsystemswithboundeddisturbances.TheproposedAC-basedcontrollerconsistsoftwoneuralnetworks(NNs)-anactionNN,alsocalledtheactor,whichapproximatestheplantdynamicsandgeneratesappropriatecon...
简介:Arecursiveidentificationmethodisproposedtoobtaincontinuous-timestate-spacemodelsinsystemswithnonuniformlysampled(NUS)data.Duetothenonuniformsamplingfeature,thetimeintervalfromonerecursionsteptothenextvariesandtheparameterisalwaysupdatedpartiallyateachstep.Furthermore,thisidentificationmethodisappliedtoformacombineddatacompressionmethodinNUSprocesses.Thedatatobecompressedarefirstclassifiedwithrespecttoaseriesofpotentiallyexisting(possiblytime-varying)models,andthenmodeledbytheNUSidentificationmethod.Themodelparametersarestoredinsteadoftheidentificationoutputdata,whichmakesthefirstcompression.Subsequently,asthesecondstep,theconventionalswingingdoortrendingmethodiscarriedoutonthedatafromthefirststep.Numericresultsfromsimulationaswellaspracticaldataaregiven,showingtheeffectivenessoftheproposedidentificationmethodandfoldincreaseofcompressionratioachievedbythecombineddatacompressionmethod.
简介:Aneuro-fuzzysystemmodelbasedonautomaticfuzzyclusteringisproposed.Ahybridmodelidentificationalgorithmisalsodevelopedtodecidethemodelstructureandmodelparameters.Thealgorithmmainlyincludesthreeparts:1)AutomaticfuzzyC-means(AFCM),whichisappliedtogeneratefuzzyrulesautomatically,andthenfixonthesizeoftheneuro-fuzzynetwork,bywhichthecomplexityofsystemdesignisreducesdgreatlyatthepriceofthefittingcapability;2)Recursiveleastsquareestimation(RLSE).ItisusedtoupdatetheparametersofTakagi-Sugenomodel,whichisemployedtodescribethebehaviorofthesystem;3)Gradientdescentalgorithmisalsoproposedforthefuzzyvaluesaccordingtothebackpropagationalgorithmofneuralnetwork.Finally,modelingthedynamicalequationofthetwo-linkmanipulatorwiththeproposedapproachisillustratedtovalidatethefeasibilityofthemethod.
简介:AnewandintelligentdesignmethodforPIDcontrollerwithincompletederivationisproposedbasedontheantsystemalgorithm(ASA).ForagivencontrolsystemwiththiskindofPIDcontroller,agroupofoptimalPIDcontrollerparametersK*p,T*i,andT*dcanbeobtainedbytakingtheovershoot,settlingtime,andsteady-stateerrorofthesystem'sunitstepresponseastheperformanceindexesandbyuseofourimprovedantsystemalgorithm.K*p,T*i,andT*dcanbeusedinreal-timecontrol.ThiskindofcontrolleriscalledtheASA-PIDcontrollerwithincompletederivation.ToverifytheperformanceoftheASA-PIDcontroller,threedifferenttypicaltransferfunctionsweretested,andthreeexistingtypicaltuningmethodsofPIDcontrollerparameters,includingtheZiegler-Nicholsmethod(ZN),thegeneticalgorithm(GA),andthesimulatedannealing(SA),wereadoptedforcomparison.ThesimulationresultsshowedthattheASA-PIDcontrollercanbeusedtocontroldifferentobjectsandhasbetterperformancecomparedwiththeZN-PIDandGA-PIDcontrollers,andcomparableperformancecomparedwiththeSA-PIDcontroller.
简介:当察觉到环境时,一个类人动物机器人被察觉到的信息总是充满,;计算通常需要重要时间;处理察觉到的信息。在这篇论文,一条选择基于注意的上下文的感觉途径被建议让类人动物机器人与高效率察觉到环境。首先,注意窗户(AW)的涵义被扩大做一个更多的将军;AW的抽象定义,;它的四种操作;声明转变也被讨论。第二,注意控制政策被描述,它集成指导目的的感性的对象选择;错误答案抑制,;能处理突现的问题。错误答案抑制被用来过滤无关的信息。最后,注意政策被看作机器人的感性的模式,罐头控制它;调整感觉效率。试验性的结果证明介绍途径能显著地支持感性的效率,;感性的费用能有效地通过采用不同注意政策被控制。
简介:流动控制器(UPFC)控制与力量系统stabilizer(PSS)的同时的协调设计一起输入的更好优化的统一力量的选择是在力量系统的现在的情形的挑战。在这份报纸,因此,执行的实验的四个集合被介绍。实验的第一个集合没有切换用线性二次的管理者(LQR)被做的骚乱情形。第二个集合为有用线性二次的gaussian(LQG)的骚乱的力量系统。这里介绍的切换的控制算法在单个机器上被测试力量系统使用的无限的总线(SMIB)linearised飞利浦Heffron模型MATLAB/SIMULINK?站台。