KinematicControlofRedundantRobotArmsUsingNeuralNetworks Kinematic Control of Redundant Robot Arms Using Neural Networks ShuaiLi HongKongPolytechnicUniversity LongJin HongKongPolytechnicUniversity MohammedAquilMirza HongKongPolytechnicUniversity Thiseditionfirstpublished2019 ©2019JohnWiley&SonsLtd Allrightsreserved.Nopartofthispublicationmaybereproduced,storedinaretrievalsystem,ortransmitted,inanyform orbyanymeans,electronic,mechanical,photocopying,recordingorotherwise,exceptaspermittedbylaw.Adviceonhow toobtainpermissiontoreusematerialfromthistitleisavailableathttp://www.wiley.com/go/permissions. TherightofShuaiLi,LongJinandMohammedAquilMirzatobeidentifiedastheauthorsofthisworkhasbeenassertedin accordancewithlaw. RegisteredOffices JohnWiley&Sons,Inc.,111RiverStreet,Hoboken,NJ07030,USA JohnWiley&SonsLtd,TheAtrium,SouthernGate,Chichester,WestSussex,PO198SQ,UK EditorialOffice TheAtrium,SouthernGate,Chichester,WestSussex,PO198SQ,UK Fordetailsofourglobaleditorialoffices,customerservices,andmoreinformationaboutWileyproductsvisitusat www.wiley.com. Wileyalsopublishesitsbooksinavarietyofelectronicformatsandbyprint-on-demand.Somecontentthatappearsin standardprintversionsofthisbookmaynotbeavailableinotherformats. LimitofLiability/DisclaimerofWarranty MATLAB® isatrademarkofTheMathWorks,Inc.andisusedwithpermission.TheMathWorksdoesnotwarrantthe accuracyofthetextorexercisesinthisbook.Thiswork’suseordiscussionofMATLAB® softwareorrelatedproductsdoes notconstituteendorsementorsponsorshipbyTheMathWorksofaparticularpedagogicalapproachorparticularuseofthe MATLAB® software.Whilethepublisherandauthorshaveusedtheirbesteffortsinpreparingthiswork,theymakeno representationsorwarrantieswithrespecttotheaccuracyorcompletenessofthecontentsofthisworkandspecifically disclaimallwarranties,includingwithoutlimitationanyimpliedwarrantiesofmerchantabilityorfitnessforaparticular purpose.Nowarrantymaybecreatedorextendedbysalesrepresentatives,writtensalesmaterialsorpromotional statementsforthiswork.Thefactthatanorganization,website,orproductisreferredtointhisworkasacitationand/or potentialsourceoffurtherinformationdoesnotmeanthatthepublisherandauthorsendorsetheinformationorservices theorganization,website,orproductmayprovideorrecommendationsitmaymake.Thisworkissoldwiththe understandingthatthepublisherisnotengagedinrenderingprofessionalservices.Theadviceandstrategiescontained hereinmaynotbesuitableforyoursituation.Youshouldconsultwithaspecialistwhereappropriate.Further,readers shouldbeawarethatwebsiteslistedinthisworkmayhavechangedordisappearedbetweenwhenthisworkwaswrittenand whenitisread.Neitherthepublishernorauthorsshallbeliableforanylossofprofitoranyothercommercialdamages, includingbutnotlimitedtospecial,incidental,consequential,orotherdamages. LibraryofCongressCataloging-in-PublicationData Names:Li,Shuai,1983–author.|Jin,Long,1988–author.| Mirza,MohammedAquil,1986–author. Title:Kinematiccontrolofredundantrobotarmsusingneuralnetworks/ ShuaiLi,HongKongPolytechnicUniversity,LongJin,Hong KongPolytechnicUniversity,MohammedAquilMirza,HongKongPolytechnic University. Description:Firstedition.|Hoboken,NJ:JohnWiley&Sons,Inc.,2019.| “Mostofthematerialsofthisbookarederivedfromtheauthors’papers publishedinjournalsandproceedingsoftheinternational conferences”–Introduction.|Includesbibliographicalreferencesand index.| Identifiers:LCCN2018048730(print)|LCCN2018050945(ebook)|ISBN 9781119556985(AdobePDF)|ISBN9781119556992(ePub)|ISBN9781119556961 (hardcover) Subjects:LCSH:Robots–Kinematics–Dataprocessing.|Manipulators (Mechanism)–Automaticcontrol.|Redundancy(Engineering)–Data processing.|Neuralnetworks(Computerscience) Classification:LCCTJ211.412(ebook)|LCCTJ211.412.K5632019(print)| DDC629.8/95632–dc23 LCrecordavailableathttps://lccn.loc.gov/2018048730 CoverDesign:Wiley CoverImage:©alashi/iStock.com Setin10/12ptWarnockbySPiGlobal,Pondicherry,India 10 9 8 7 6 5 4 3 2 1 Toourparentsand ancestors,asalways vii Contents ListofFigures xiii ListofTables xix Preface xxi Acknowledgments xxv PartI NeuralNetworksforSerialRobotArmControl 1 1 ZeroingNeuralNetworksforControl 3 1.1 Introduction 3 1.2 SchemeFormulationandZNNSolutions 4 1.2.1 ZNNModel 4 1.2.2 NonconvexFunctionActivatedZNNModel 8 1.3 TheoreticalAnalyses 9 1.4 ComputerSimulationsandVerifications 12 1.4.1 ZNNforSolving(1.13)att =1 12 1.4.2 ZNNforSolving(1.13)withDifferentBounds 15 1.5 Summary 16 2 AdaptiveDynamicProgrammingNeuralNetworksforControl 17 2.1 Introduction 17 2.2 PreliminariesonVariableStructureControloftheSensor–Actuator System 18 2.3 ProblemFormulation 19 2.4 Model-FreeControloftheEuler–LagrangeSystem 20 2.4.1 OptimalityCondition 21 2.4.2 ApproximatingtheActionMappingandtheCriticMapping 21 2.5 SimulationExperiment 23 2.5.1 TheModel 23 2.5.2 ExperimentSetupandResults 24 2.6 Summary 25 viii Contents 3 ProjectionNeuralNetworksforRobotArmControl 27 3.1 Introduction 27 3.2 ProblemFormulation 29 3.3 AModifiedControllerwithoutErrorAccumulation 30 3.3.1 ExistingRNNSolutions 30 3.3.2 LimitationsofExistingRNNSolutions 32 3.3.3 ThePresentedAlgorithm 33 3.3.4 Stability 34 3.4 PerformanceImprovementUsingVelocityCompensation 36 3.4.1 AControlLawwithVelocityCompensation 36 3.4.2 Stability 37 3.5 Simulations 41 3.5.1 RegulationtoaFixedPosition 41 3.5.2 TrackingofTime-VaryingReferences 42 3.5.3 Comparisons 47 3.6 Summary 50 4 NeuralLearningandControlCo-DesignforRobotArmControl 51 4.1 Introduction 51 4.2 ProblemFormulation 52 4.3 NominalNeuralControllerDesign 53 4.4 ANovelDualNeuralNetworkModel 54 4.4.1 NeuralNetworkDesign 54 4.4.2 Stability 56 4.5 Simulations 62 4.5.1 SimulationSetup 62 4.5.2 SimulationResults 63 4.5.2.1 TrackingPerformance 63 4.5.2.2 Withvs.WithoutExcitationNoises 64 4.6 Summary 66 5 RobustNeuralControllerDesignforRobotArmControl 67 5.1 Introduction 67 5.2 ProblemFormulation 68 5.3 DualNeuralNetworksfortheNominalSystem 69 5.3.1 NeuralNetworkDesign 69 5.3.2 ConvergenceAnalysis 71 5.4 NeuralDesigninthePresenceofNoises 72 5.4.1 PolynomialNoises 72 5.4.1.1 NeuralDynamics 73 5.4.1.2 PracticalConsiderations 77 5.4.2 SpecialCases 78 5.4.2.1 ConstantNoises 78 5.4.2.2 LinearNoises 80 5.5 Simulations 81 5.5.1 SimulationSetup 81 5.5.2 NominalSituation 81
Description: