简介:Thispaperpresentsthedesignandprototypeofasmallquadrupedrobotwhosewalkingmotionisrealizedbytwopie-zocompositeactuators.Inthedesign,biomimeticideasareemployedtoobtaintheagilityofmotionsandsustainabilityofaheavyload.Thedesignoftherobotlegsisinspiredbythelegconfigurationofinsects,twojoints(hipandknee)ofthelegenabletwobasicmotions,liftingandstepping.Therobotframeisdesignedtohaveasloperelativetothehorizontalplane,whichmakestherobotmoveforward.Inaddition,theboundinglocomotionofquadrupedanimalsisimplementedintherobot.Experimentsshowthattherobotcancarryanadditionalloadofabout100gandrunwithafairlyhighvelocity.Thequadrupedprototypecanbeanimportantsteptowardsthegoalofbuildinganautonomousmobilerobotactuatedbypiezocompositeactuators.
简介:Thispaperthoroughlyinvestigatestheproblemofrobotself-locationbylinecorrespondences.Theoriginalcontributionsarethree-fold:(1)Obtainthenecessaryandsufficientconditiontodeterminelinearlytherobot'sposebytwolinecorrespondences.(2)Showthatifthespacelinesareverticalones,itisimpossibletodeterminelinearlytherobot'sposenomatterhowmanylinecorrespondenceswehave,andtheminimumnumberoflinecorrespondencesis3todetermineuniquely(butnon-linearly)therobot'spose.(3)Showthatifthespacelinesarehorizontalones,theminimumnumberoflinecorrespondencesis3forlineardeterminationand2fornon-lineardeterminationoftherobot'spose.
简介:Forthepurposeofsolvingoptimalcontrolproblemofawall-crawlingmobilerobotworkingonsphericalcontainers,weproposetheHamel'sformalismforPontryaginMaximumPrinciple,whichgivesageneralframeworkfortheoptimalcontrolofamechanicalsystemwithvelocityconstraints,especiallynonholonomicconstraints.Theeffectivenessoftheproposedframeworkisshownbythesimulationsfortheaboveproblem.
简介:Thegeneralizedregressionneuralnetwork-onekindofRBFneuralnetwork,ischosentoconstructtheinverse-kinematicsmodelfortheshotcreterobotwhichhasredundantdegree-of–freedom.Theinverse-kinematicsmodeloftheobjectistrainedbythegenerallearningmethod.Inconstructingmodelprocess,differentpartitionmethodsistriedtodividethejointspaceanddifferentdiffusioncoefficientvaluetotraintheneuralnetwork.Theinfluenceofthespreadcoefficienttotheapproachabilityisalsostudied.Thesimulationmethodisadoptedtotesttheperformanceoftheneuralnetwork.Thesimulationresultturnsouttobesatisfactory.
简介:Acertainnumberofconsiderationsshouldbetakenintoaccountinthedynamiccontrolofrobotmanipulatorsashighlycomplexnon-linearsystems.Inthisarticle,weprovideadetailedpresentationofthemechanicalandelectricalimpli-cationsofrobotsequippedwithDCmotoractuators.Thismodeltakesintoaccountallnon-linearaspectsofthesystem.Then,wedevelopcomputationalalgorithmsforoptimalcontrolbasedondynamicprogramming.Therobot’strajectorymustbepredefined,butperformancecriteriaandconstraintsapplyingtothesystemarenotlimitedandwemayadaptthemfreelytotherobotandthetaskbeingstudied.Asanexample,amanipulatorarmwith3degreesoffreedomisanalyzed.
简介:用身体部分名词填空:Hello!Everyone.Iamarobot.Ithinkwithmyh__.Iwalkwithmyf__.Ineedn't(不必)washmyf__everyday.Ihavetwobrighte__.Ispeakwithmym__.Ismellwithmyn__.Ilistenwithmye__.Idon'thaveany
简介:Inthispaper,itintroducedthemechanicalstructuredesignofthemultifunctionalwalkingrehabilitationrobot.WeusedtheSolidWorkstodesignandassembletherehabilitationrobot,andoptimizedtherehabilitationrobotbasedonergonomics.Theresultshowsthatthemultifunctionalwalkingrehabilitationrobothasthecharacteristicsofcompletefunctions,convenientoperation,compactstructureandsoon.Itcanmeettherequirementsofmedicalcareequipment,andeffectivelycompletethenursingworkofpatients.
简介:Theresearchonunderwatership-hullcleaningrobotwasconductedonthepurposeofrealizingtheautomationofcleaningunderwatershiphullsothatservicelifeofshipwillbeprolongedandshipspeedwillraised.Moreover,fuelconsumptionandtheworkintensityofdiverswillbereduced.Inthispaper,thecurrentsituationandthelatesttechnologyinChinaandabroadwereanalyzed;meanwhile,thetypicalcharacteristicsoftheunderwatercleaningrobotwereintroduced.Accordingtotheworkprincipleoftheunderwatercleaningrobot,theemphasiswasputontheanalysisandstudyofpermanent-magneticabsorption,magneticwheel,airproofandanticorrosion,underwatercleaningequipmentandcontrolsystem.Therobotiseasyinrotationandsimpleincontrol.
简介:Awheeledmobilemechanismwithapassiveand/oractivelinkagemechanismforroughterrainenvironmentisdevelopedandevaluated.Thewheeledmobilemechanismwhichhashighmobilityinroughterrainneedssophisticatedsystemtoadaptvariousenvironments.Wefocusonthedevelopmentofaswitchingcontrollersystemforwheeledmobilerobotsinroughterrain.Thissystemconsistsoftwosub-systems:anenvironmentrecognitionsystemusinglinkanglesandanadaptivecontrolsystem.Intheen-vironmentrecognitionsystem,weintroduceaSelf-OrganizingMap(SOM)forclusteringlinkangles.Intheadaptivecontrol-lers,weintroduceneuralnetworkstocalculatetheinversemodelofthewheeledmobilerobot.Theenvironmentrecognitionsystemcanrecognizetheenvironmentinwhichtherobottravels,andtheadjustablecon-trollersaretunedbyexperimentalresultsforeachenvironment.Thedualsub-systemswitchingcontrollersystemisexperi-mentallyevaluated.Thesystemrecognizesitsenvironmentandadaptsbyswitchingtheadjustablecontrollers.Thissystemdemonstratessuperiorperformancetoawell-tunedsinglePIDcontroller.
简介:Whilemuchattentionhasbeengiventobio-roboticsinrecentyears,notmuchofthishasbeengiventothechallengingsubjectoflocomotioninslipperyconditions.Thisstudybeginstorectifythisbyproposingabiomimeticapproachtogeneratingthefrictionrequiredtogivesufficientpropulsiveforceonaslipperysubstrate.Wetookinspirationfromasuccessfulbiologicalsolution-thatofapplyinghair-likestructurestothepropulsiveappendages,similartothesetaefoundinnereidpolychaeteslivinginmuddyhabitats.Webeganbyexaminingthemorphologyandthemeanlocomotionparametersofoneofthemostcommonnereids:Nereisdiversicolor.Followingthisstudy,wedesignedandfabricatedaroboticsystemwithappendagesimitatingthebiologicalshapefoundintheworm.Aflexiblecontrolsystemwasdevelopedtoallowmostofthelocomotionparametersobservedintherealwormtobeappliedtotherobot.Experimentsonthreedifferentnaturalsubstratesrangingfromfinesandtogravelshowedthat,whereasaplateattachedtotheappendagegeneratedmostthrustonasmallparticlesubstrate,abundleofartificialsetaeattachedtotheappendagegeneratedmostthrustonalargeparticlesubstrate.Onalltypesofsubstratetested,anappendagewithoutanyattachmentdidsignificantlyworsethanonewith.Thissuggeststhathair-likestructurescanbeadvantageous.