intelligent control

Development

Theideaof​​intelligentcontrolappearedinthe1960s.Atthattime,thestudyoflearningcontrolwasveryactiveandgotbetterapplications.Forexample,self-learningandadaptivemethodshavebeendevelopedtosolvetherandomcharacteristicproblemsofthecontrolsystemandtheunknownproblemsofthemodel;in1965,ProfessorK.S.FuofPurdueUniversityintheUnitedStatesfirstusedAI'sheuristicreasoningrulesLearningcontrolsystem;In1966,AmericanMendel(J.M.Mendel)firstadvocatedtheuseofAIinthedesignofspacecraftcontrolsystems.

In1967,AmericanC.T.Leondesandothersformallyusedtheterm"intelligentcontrol"forthefirsttime.In1971,FuJingsundiscussedthecrossrelationshipbetweenAIandautomaticcontrol.Sincethen,automaticcontrolandAIbegantocollidewithsparks,andanemergingcross-field-intelligentcontrolhasbeenestablishedanddeveloped.Earlyintelligentcontrolsystemsusedrelativelyelementaryintelligentmethods,suchaspatternrecognitionandlearningmethods,andtheirdevelopmentspeedwasveryslow.

Zadepublishedthefamouspaper"FuzzySets"in1965,whichopenedupanewfieldofmathematicsbasedonfuzzylogicthatrepresentsthegeneraluncertaintyofhumanperceptionandlanguageexpression.-Fuzzymathematics.In1975,BritishE.H.Mamdanisuccessfullyappliedfuzzylogicandfuzzyrelationstoindustrialcontrolsystems,andproposedafuzzycontrolmethodthatcandealwithfuzzyuncertaintiesandsimulatehumanoperatingexperiencerules.Sincethen,controlexpertshaveconductedalotofresearchonthetheoryandapplicationoffuzzycontrol,andhaveachievedanumberofinterestingresults,whichareregardedasveryactiveanddeeplydevelopedintelligentcontrolmethodsinintelligentcontrol.

Inthe1980s,AI-basedrulerepresentationandreasoningtechnology(especiallyexpertsystems)rule-basedexpertcontrolsystemsdevelopedrapidly,suchasK.J.Astrom,Sweden.TheexpertcontrolofG.M.Saridis,theexpertcontrolintherobotcontrolofG.M.Saridis,etc.Withthere-emergenceofartificialneuralnetworkresearchinthemid-1980s,researchersinthecontrolfieldproposedandquicklydevelopedneuralnetworkcontrolmethodsthatmakefulluseofthegoodnonlinearapproximationcharacteristics,self-learningcharacteristicsandfaulttolerancecharacteristicsofartificialneuralnetworks.

Withthedevelopmentanddeepeningofresearch,theconditionsforforminganewdisciplineofintelligentcontrolgraduallymature.InAugust1985,IEEEheldthefirstIntelligentControlAcademicSymposiuminNewYork,USA,todiscusstheprinciplesandsystemstructureofintelligentcontrol.Asaresult,intelligentcontrolhasbeenwidelyrecognizedasanemergingdisciplineandhasachievedrapiddevelopment.

Inthepasttenyears.Withthedevelopmentofintelligentcontrolmethodsandtechnologies,intelligentcontrolhasrapidlymovedtovariousprofessionalfieldsandappliedtothecontrolproblemsofvariouscomplexcontrolledobjects,suchasindustrialprocesscontrolsystems,roboticsystems,modernmanufacturingsystems,andtrafficcontrolsystems.

Definition

Thedefinitionofintelligentcontrol:Intelligentcontrolistheprocessbywhichintelligentmachinesachievetheirgoalsautonomously.Anintelligentmachineisdefinedasamachinethatperformstasksprescribedbyhumansautonomouslyorinteractivelyinastructuredorunstructured,familiarorunfamiliarenvironment.

Definition2:KJOstorombelievesthathumans’intuitionandtrial-and-errorandotherintelligencesareformalizedormachine-simulated,andusedintheanalysisanddesignofcontrolsystemstomakethemToacertainextent,realizetheintelligentizationofthecontrolsystem,whichisintelligentcontrol.Healsobelievesthatself-adjustingcontrolandadaptivecontrolarelow-levelmanifestationsofintelligentcontrol.

Definition3:Intelligentcontrolisatypeofautomaticcontrolthatcanautonomouslydriveintelligentmachinestoachieveitsgoalswithouthumanintervention.Itisalsoanimportantfieldinwhichcomputersareusedtosimulatehumanintelligence.

Definition4:Intelligentcontrolisactuallyjustanewbranchofresearchandsimulationofhumanintelligentactivitiesandthelawsofcontrolandinformationtransmissionprocesses,andthedevelopmentofengineeringcontrolandinformationprocessingsystemswithhuman-likeintelligence.

Technicalbasis

Intelligentcontrolisbasedoncontroltheory,computerscience,artificialintelligence,operationsresearchandotherdisciplines,andexpandsrelatedtheoriesandtechnologies.Amongthem,therearemorefuzzyapplications.Logic,neuralnetworks,expertsystems,geneticalgorithmsandothertheories,aswellastechnologiessuchasadaptivecontrol,self-organizingcontrol,andself-learningcontrol.

Expertsystemisacontrolsystemthatusesexpertknowledgetodescribespecialordifficultproblems.Althoughexpertsystemshavebeensuccessfullyappliedinsolvingcomplexadvancedreasoning,theactualapplicationsofexpertsystemsarerelativelyfew.

Fuzzylogicusesfuzzylanguagetodescribethesystem,whichcandescribeboththequantitativemodelandthequalitativemodeloftheapplicationsystem.Fuzzylogiccanbeappliedtoarbitrarilycomplexobjectcontrol.

Asanon-deterministicquasi-naturalstochasticoptimizationtool,geneticalgorithmhasthecharacteristicsofparallelcomputingandrapidsearchforglobaloptimalsolutions.Itcanbemixedwithothertechnologiesforintelligentcontrolofparametersandstructures.Oroptimalcontroloftheenvironment.

Neuralnetworkisanadaptivecontrolmethodthatusesalargenumberofneuronstolearnandadjustaccordingtoacertaintopology.Itcanexpressawealthofcharacteristics,includingparallelcomputing,distributedstorage,variablestructure,highfaulttolerance,nonlinearoperation,self-organization,learningorself-learning.Thesecharacteristicsaresystemcharacteristicsthatpeoplehavelongpursuedandexpected.Neuralnetworkhasuniqueabilitiesinself-adaptation,self-organization,self-learningandothercontrolaspectsofintelligentcontrolparameters,structureorenvironment.

Thecombinationofrelatedtechnologiesofintelligentcontrolandcontrolmethods,orcomprehensivecross-combination,constitutesintelligentcontrolsystemsandintelligentcontrollerswithdifferentstylesandfunctions,whichisalsoamainfeatureofintelligentcontroltechnologymethods.

Researchobject

Themaingoalofintelligentcontrolresearchisnolongerthecontrolledobject,butthecontrolleritself.Thecontrollerisnolongerasingleanalyticalmodelofmathematicalmodel,butageneralizedmodelthatcombinesmathematicalanalysisandknowledgesystem,andisacontrolsystemthatcombinesknowledgeofmultipledisciplines.Intelligentcontroltheoryistoestablishthecharacteristicpatternrecognitionofthecontrolleddynamicprocess,andcontrolbasedonknowledge,experience-basedreasoningandintelligentdecision-making.Agoodintelligentcontrolleritselfshouldhavethecharacteristicsofmulti-mode,variablestructure,variableparameters,etc.Itcanidentify,learnandorganizeitsowncontrolmodeaccordingtothecharacteristicsofthecontrolleddynamicprocess,changethestructureofthecontrollerandadjusttheparameters.

Theresearchobjectsofintelligentcontrolhavethefollowingcharacteristics:

1.Uncertaintymodel

Theresearchobjectsofintelligentcontrolusuallyhaveseriousuncertainties.sex.Themodeluncertaintymentionedherehastwomeanings:oneisthatthemodelisunknownorpoorlyunderstood;theotheristhatthestructureandparametersofthemodelmayvaryinawiderange.

2.Highlynon-linearity

Forhighlynon-linearcontrolobjects,theuseofintelligentcontrolmethodscanoftensolvethecontrolproblemsofnon-linearsystems.

3.Complextaskrequirements

Forintelligentcontrolsystems,thetaskrequirementsareoftenmorecomplicated.

Currently,intelligentcontroliswidelyusedinservosystemapplications,mainlyincludingexpertcontrol,fuzzycontrol,learningcontrol,neuralnetworkcontrol,predictivecontrolandothercontrolmethods.

Features

Themaindifferencebetweenintelligentcontrolandtraditionalcontrolisthattraditionalcontrolmethodsmustrelyonthemodelofthecontrolledobject,andintelligentcontrolcansolvethecontrolproblemsofnon-modeledsystems.Comparedwithtraditionalcontrol.

Intelligentcontrolhasthefollowingbasiccharacteristics:

1)Thecoreofintelligentcontrolishigh-levelcontrol.Itcancarryouteffectiveglobalcontrolofcomplexsystems(suchasnonlinear,fasttime-varying,complexmulti-variable,environmentaldisturbance,etc.).Realizegeneralizedproblemsolving.Andhasstrongfaulttolerance.

2)Theintelligentcontrolsystemcanbeanon-mathematicalgeneralizedmodelrepresentedbyknowledgeandahybridcontrolprocessrepresentedbymathematics.Itadoptsamulti-modalcontrolmethodcombiningopen-closedloopcontrolandqualitativedecision-makingandquantitativecontrol.

3)Itsbasicpurposeistoanalyzeandsynthesizethesystemfromtheperspectiveofsystemfunctionandoveralloptimization.Toachievethepredeterminedgoals.Theintelligentcontrolsystemhasthecharacteristicsofvariablestructureandcanfindthebestoverall.Withself-adaptation,self-organization,self-learningandself-coordinationcapabilities.

4)Theintelligentcontrolsystemhassufficientknowledgeabouthumancontrolstrategies,controlledobjectsandtheenvironment,andtheabilitytousetheseknowledge.

5)Theintelligentcontrolsystemhascompensationandself-repaircapabilities,aswellasjudgmentanddecision-makingcapabilities.

Application

Thespecificapplicationofintelligentcontrolismainlymanifestedinthefollowingaspects:

1)Intelligentcontrolintheproductionprocess

Intelligentcontrolintheproductionprocessmainlyincludeslocal-levelintelligentcontrolandglobal-levelintelligentcontrol.

Local-levelintelligentcontrolreferstotheintroductionofintelligenceintoacertainunitintheprocessforcontrollerdesign.TheresearchhotspotistheintelligentPIDcontroller,becauseithasobviousadvantagesinparametertuningandonlineadaptiveadjustment,anditcanbeusedtocontrolsomenonlinearcomplexobjects.

Global-levelintelligentcontrolismainlyaimedattheautomationoftheentireproductionprocess,includingthecontroloftheentireoperationprocess,processfaultdiagnosis,planningprocessoperationandhandlingabnormalities,etc.

2)Intelligentcontrolinadvancedmanufacturingsystems

Intelligentcontroliswidelyusedinthemachinerymanufacturingindustry.Inmodernadvancedmanufacturingsystems,itisnecessarytorelyonincompleteandinaccuratedatatosolvedifficultorunpredictablesituations.Artificialintelligencetechnologyprovidessomeeffectivesolutionstothisproblem.

(1)Usefuzzymathematicsandneuralnetworkmethodstomodelthedynamicenvironmentofthemanufacturingprocess,andusesensorfusiontechnologyforinformationpreprocessingandsynthesis.

(2)Usingexpertsystemasfeedbackmechanism,modifycontrolmechanismorchoosebettercontrolmodeandparameters.

(3)Usefuzzysetdecision-makingselectionmechanismtoselectcontrolactions.

(4)Usethelearningfunctionofneuralnetworksandtheabilitytoprocessinformationinparalleltoperformonlinepatternrecognitionandprocessinformationthatmaybeincomplete.

3)Intelligentcontrolinpowersystem

Design,production,operationandcontrolofelectricalequipmentsuchasgenerators,transformersandmotorsinpowersystemsItisacomplicatedprocess.Electricalworkersathomeandabroadhaveintroducedartificialintelligencetechnologyintotheoptimizationdesign,faultdiagnosisandcontrolofelectricalequipment,andachievedgoodcontrolresults.

(1)Usinggeneticalgorithmtooptimizethedesignofelectricalequipmentcanreducecosts,shortencalculationtime,andimprovetheefficiencyandqualityofproductdesign.

(2)Theintelligentcontroltechnologyappliedtothefaultdiagnosisofelectricalequipmentincludesfuzzylogic,expertsystemandneuralnetwork.

(3)TheapplicationofintelligentcontrolincurrentcontrolPWMtechnologyisoneoftherepresentativetechnicalapplicationdirections,anditisalsooneofthenewhotspotsofresearch.

Inrecentyears,theintelligentcontroltechnologyhasbeengreatlydevelopedathomeandabroad,andhasenteredthestageofengineeringandpracticalapplication.Asanemergingtheoreticaltechnology,itisstillinaperiodofdevelopment.Withtherapiddevelopmentofartificialintelligencetechnologyandcomputertechnology,intelligentcontrolisboundtousherinaneweraofitsdevelopment.