2024 全球6G技術(shù)大會(huì) -10.0P Network Energy Saving Technology White Paper-final-20240409

Contents

1.Preface 2

2.Significance 2

3.PerformanceIndicators 3

4.StandardProgress 4

5.KeyTechnologiesforNetworkEnergySaving 5

5.1NetworkArchitecture 5

5.1.1SAGINArchitecture 5

5.1.2NewDistributedRANArchitecture 6

5.1.3WirelessIntelligentCloudNetwork 9

5.2AirInterfaceEnergySavingTechnologies 10

5.2.1EnergySavingTechnologiesinSpatialDomain 10

5.2.2EnergySavingTechnologiesinTimeDomain 12

5.2.3EnergySavingTechnologiesinFrequencyDomain 13

5.2.4EnergySavingTechnologiesinPowerDomain 14

5.2.5NewAirInterfaceHardware 14

5.3IntegrationwithNewTechnologies 16

5.3.1IntegrationwithAITechnologies 16

5.3.2Integrationwith6GAirInterfaceTechnologies 19

5.4OtherTechnologies 20

6.SummaryandOutlook 22

7.References 23

8.MainContributors 24

9.Abbreviations 24

1/25

2/25

1.Preface

ThisWhitePaperdescribesthesignificanceandseriouschallengesof6Gnetworkenergysaving,brieflyintroducestheenergyconsumptionperformanceindicatorsandthestandardprogressof3GPPonnetworkenergysaving,andhighlightsthekeytechnicalsolutionsfornetworkenergysavinginaspects,suchasthenetworkarchitecture,energysavingtechnologiesforairinterfaces,integrationwithnewtechnologies,andothertechnologies.Finally,thisWhitePapersummarizesthemaincontentandrelatedconclusions,andlooksforwardtothefuturedevelopmenttrends.

2.Significance

Withtherapiddevelopmentofglobaleconomyandsciencetechnology,energyissuesarebecomingmoreandmoreprominent.Globalcarbondioxideemissionhasincreasedsignificantlysince2000.Withtherapidincreaseofcarbondioxideintheair,globaltemperaturerisesrapidly,andextremeweathersuchasstormsandheatwavescausedbyglobalwarmingseriouslyendangershumanlifeandproperty.

InChina,thedualcarbongoals,thatis,peakcarbondioxideemissionby2030andcarbonneutralizationby2060areincludedinthe14thFive-YearPlan.Thedualcarbongoalsaremajorresponsibilitiesoftheglobaltocopewithclimatechangesandimportantcornerstonesforsustainabledevelopmentofindustriesandenterprises.Inthetelecommunicationsindustry,carbondioxideemissionsaremainlyfromconsumedelectricity.Theenergyconsumptionofbasestations,communicationequipmentrooms,anddatacentersaccountsforthemajorproportionofthetotalenergyconsumption.Therefore,itiscriticaltosaveenergyfortheseitems.Theenergyconsumptionofa5Gbasestationatfullloadisabout3to4timesthatofa4Gbasestation.Especiallywiththeformalcommercialuseof5Gnetworks,energyconsumptionincreasessignificantly.

Thesixtypicalscenariosandfifteencapabilityindicators[1]proposedfor6Gposehigherrequirementsonspeed,capacity,latency,positioning,anduserexperiencefrommultipledimensionssuchasintelligence,sensing,andubiquity.Thisdrives6Gtohigherfrequency,largerbandwidth,andmorecomputingpower,whichbringsseverechallengesto6Gnetworkenergysaving.

I.Higherfrequency:Thecoverageradiusof6Gmillimeterwavebasestationsisonly30%thatof5G3.5GHzbasestations,andthepoweramplificationefficiencyof6Gmillimeterwavebasestationsisabout7%to15%.Thespecificvaluevariesdependingontheprocess.Forexample,itis7%+forthesilicon

3/25

germanium(SiGe)processand15%+forthegalliumnitride(GaN)process,whichisonly1/7to1/3thatoftraditional5Gbasestations.Therefore,moreenergyisrequiredtosupportnormaloperationofthepoweramplifier(PA)of6Gmillimeterwavebasestations.

II.Largerbandwidth:Largebandwidthandmultipleantennasarethemainfactorsforincreasedpowerconsumptionof5Gbasestations.Thepowerconsumptionofa5Gbasestationis3to4timesthatofa4Gbasestation.Accordingtotheintergenerationalgrowthpatternofbandwidth,itisexpectedthatthe6Gbandwidthcanreach500MHzto1GHz.Ifthetransmitpowerperunitofbandwidthremainsunchanged,itisestimatedthatthetransmitpowerof6Gbasestationsismorethanfivetimesthatof5Gbasestations,andtheoverallpowerconsumptionofa6Gbasestationismorethanfourtimesthatofa5Gbasestation.

III.Morecomputingpower:Endogenousintelligenceisanimportantfeatureof6G.Commonlyusedartificialintelligence(AI)modelsincludeadozenofmegabytestohundredsofgigabytesofmodelparameters.Forexample,ChatGPTcontains175billionmodelparameters,anduses10,000V100GPUsformodeltraining.AccordingtoaroughcalculationbytheGlobalZeroEmissionResearchCenter(GZR),itspowerconsumptionexceeds1.68millionkilowatt-hours.Ifthenumberofvisitorsperdayis1million,about12,000kilowatt-hoursofelectricityareconsumedeveryday.

Greenandenergysavingshouldbethebasicprinciplesfordevelopingnewinnovative6Gtechnologies,soastoimprovesystemenergyefficiencyandimplementagreenandecologicaloperatingmodel.Inaddition,6Gtechnologiesshouldempowerthousandsofindustriestohelpallwalksoflifeperformdigitaltransformationthoroughly,implementgreendevelopmentstrategies,andjointlywriteanewchapterofasharedfutureforthemankind.

3.PerformanceIndicators

Energyefficiencyisanimportantperformanceindicatorforevaluatingnetworkenergyconsumption.Wecanfindeffectivemethodsfornetworkenergysavingfromthedefinitionofenergyefficiency.

Energyefficiencyisdefinedinacademicresearchastheamountofdatathatcanbetransmittedunderunitofenergyconsumedandismeasuredinbit/joule(bit/J).Toimproveenergyefficiency,wecanconsidertheamountofdatatransmittedandenergyconsumption.First,wecaneffectivelyimprovethetransmissionrate.Second,wecanreducetheenergyconsumedfortransmittingtheunitamountofdata.

4/25

Inaddition,theinternationaltelecommunicationunion(ITU)definestheenergyefficiencyoftraditional5Gnetworksastheairinterfacecapabilities[2]thatarerelatedtoprovidedservicesandusedtominimizeenergyconsumptionoftheradioaccessnetwork(RAN).Theenergyefficiencyincludestwofactors:(1)Networkside:thenumberofinformationbitstransmittedorreceivedbytheuserequipment(UE)ontheRANperunitofenergyconsumed;(2)UEside:unitofenergyconsumedbycommunicationmodules,measuredinbit/J[3].Therefore,toimprovenetworkenergyefficiencyandachievenetworkenergysaving,wecantakemeasuresfromthenetworkandUE.

As6Gappliestodiversifiedapplicationscenarioswithdifferenttransmissionperformancerequirementsonairinterfaces,thenetworkenergyefficiencycanbedefinedastheratiooftheperformanceindicatortothepowerconsumptioninaspecificscenario.Thishelpsreflecttheactualperformancerequirementsandenergyconsumptioninthescenariomorecomprehensivelyandobjectively.

Specifically,energyefficiencycanbedefinedasthedatarateperunitofenergyconsumedinhigh-speedscenarios,measuredinbps/J,thetransmissionlatencyperunitofenergyconsumedinlow-latencyscenarios,measuredins/J,andthecoveragedistanceperunitofenergyconsumedinwidecoveragescenarios,measuredinm/J.Similarly,inintegrated6Gscenarios,theenergyefficiencycanbeconsideredinmultipledimensions.Theenergyefficiencycorrespondstoperformanceindicatorsin6Gscenarios,whichhelpsevaluatetheairinterfaceperformanceperunitofenergyconsumedmorecomprehensivelyandmeetrequirementsindiversified6Gapplicationscenarios.

4.StandardProgress

3GPP,aninternationalmobilecommunicationstandardorganization,carriedoutdiscussionsonnetworkenergysavingtechnologiesinRelease18,whichincludesthestudyitem(SI)andworkitem(WI).

(1)SIStage

AttheRANmeeting#94eheldinDecember2021,theresearchcontentofnetworkenergysavingwasformallydetermined,whichmainlyincludesthefollowingthreeaspects:

.Establishanenergyconsumptionsimulationmodelforbasestationstoevaluatetheperformanceofnetworkenergysavingsolutions.

.ProvideevaluationmethodsandKPIsfornetworkenergysaving.

5/25

.ResearchandidentifyenergysavingtechnologiesonthegNBandUEsides.

TheSIstagelastsfornearlyoneyear,focusesonthesimulationmodelsandevaluationmethodsfornetworkenergysaving,andthenetworkenergysavingtechnologiesintime,frequency,spatial,andpowerdomains,UEauxiliaryinformation,andotheraspects,andprovidesthesimulationresults.RelatedresearchcontentformsTR38.864[4].

(2)WIStage

BasedontheresearchprogressintheSIstage,theWIstagefocusesonsometechnologieswithhighenergysavinggains.AttheRANmeeting#98heldinDecember2022,theworkcontentofnetworkenergysavingwasdeterminedasfollows:

.Networkenergysavingtechnologiesinspatialandpowerdomainsbasedonchannelstateinformation(CSI)enhancement

.Celldiscontinuoustransmission/reception(DTX/DRX)

.SSB-lessSCellininter-bandmulti-carrier(CA)scenarios(applicableonlytoFR1andco-locatedcells)

.SolutionforpreventinglegacyUEsfromcampingincellsimplementingRel-18networkenergysavingtechnologies

.Inter-nodebeamactivationandenhancedpagingwithinlimitedareas

.Enhancedcellhandoverprocess

SomenetworkenergysavingtechnologiesstudiedintheSIstagearenotstandardized.Tofurtherreducenetworkenergyconsumptionandlaythefoundationfor6G,3GPPRelease19furtherstandardizesnetworkenergysaving.

5.KeyTechnologiesforNetworkEnergySaving

5.1NetworkArchitecture

5.1.1SAGINArchitecture

FuturenetworkswillrealizetheInternetofEverything(IoE),andmorenaturalspaces,suchasspace,air,ground,andoceanwillbecovered,ensuringubiquitousconnectivityinalldomains.Asdigitalizationisacceleratedinallwalksoflife,digitalinfrastructureinalldomainswillbeexpandedrapidly,andthecontradictionbetweenthedevelopmentofdigitaleconomyandtheincreaseofenergyconsumptionandcarbondioxideemissionwillbecomeincreasinglyprominent.

6/25

Greenandenergysavingwillbecomeendogenousdemandsforfuture6Gnetworkarchitecture.ToachievethedevelopmentgoalsofIoE,green,andlowcarbon,the6Gnetworkarchitecturewillundergoasubversivereconstruction.Thelegacyborderedandchimney-likeRANarchitecturewillbeconvertedtothespace-air-groundintegratednetwork(SAGIN)architecturefeaturingubiquitous,green,andenergysaving.

TheSAGINarchitectureconsistsofthreelayers:satellitenetwork,airbornenetwork,andlow-altitudeandgroundnetwork,formingathree-dimensionalall-domaincoveragenetworkbasedonterrestrialnetworksandexpandedbynon-terrestrialnetworks.Theterrestrialandnon-terrestrialnetworksare

interconnectedanddeeplyintegrated,useaunifiedprotocolstack,andsupportperception-free,simplified,andubiquitousaccessofmassivevolumeofusers.Theterrestrialandnon-terrestrialnetworkscanusenewnetworkingmodessuchassupercellularandnon-cellularthatareinlinewiththedevelopmenttrendofgreencommunication.Inthesupercellulararchitecture,thecontrolplaneanduserplaneofabasestationaredecoupled,andcontrolandservicebasestationscanbedeployedindependentlyasrequired.ThecontrolbasestationconnectstoUEs,transmitscontrolsignals,andcanadoptthelargeareacoveragemode.Theservicebasestationprovidesuserswithhigh-speeddatatransmissionandcanbeflexiblydeployedasrequired.Multipleservicebasestationscanbedeployedwithinthecoverageofacontrolbasestation,andtheservicebasestationscandynamicallysleepbasedonchangesinserviceload.Inthisarchitecture,networkcoveragecanbedynamicallyadjustedbasedonservicerequirements.Whenthecoverageperformanceisnotaffected,servicebasestationscansleepatappropriatetime,ensuringmoreflexiblesleepandimprovingnetworkenergysaving.Thenon-cellulararchitectureisUE-centered,withmultipledistributedaccesspoints(APs)andacentralizedunit(CU)connectedtoallAPsdeployed.ThroughcentralizedsignalprocessingoftheCU,widelydistributedAPscanachievehigh-levelcollaborationandformasuperbasestationthatcoverstheentirearea.EachUEaccessesaspecificgroupofAPs.Spatialmacrodiversityandlowpathlosscanbeusedtoimprovethespectralefficiencyandenergyefficiencyofthenetwork.Whentherearefewusersinanarea,someAPscanbeshutdowntofurtherreducesystemenergyconsumption[5][6].

5.1.2NewDistributedRANArchitecture

Tobettersupportverticalindustriessuchasautonomousdriving,intelligentmanufacturing,andtelemedicine,therearehigherlatencyandreliabilityrequirements.Especiallyforubiquitousconnectionsoffuture6Gnetworks,the

7/25

traditionalcentralizedintelligentnetworkarchitecturecannotmeettherequirements.Therefore,theindustryhasproposedaseriesofnewdistributedRANarchitectures,whichintroduceadistributedintelligentcomputingframeworktofullyutilizethemulti-dimensionaldataandcomputingresourcesheldbyUEsandnodes.However,thenewdistributedRANarchitecturefacesmanychallenges,suchascontinuousexpansionofdistributednodes,transmissionofmassivevolumeofhigh-dimensionalmodelparameters,andsupercomputingpower.Asaresult,6Gnetworkenergyconsumptionhasbecomeoneofthemainbottlenecksforitslarge-scaledeploymentandwidespreadapplication.Thedistributed,hierarchical,andintelligentRANarchitecturedesignisadoptedtoeffectivelyreducetheenergyconsumptionofthe6GdistributedRAN[7].

AstheRANadoptsamulti-layernetworktopology,intelligentfunctionalcomponentscanbedeployedatdifferentlayers,suchasmacro/microbasestations,CU/DU,cloud/edgetocarryoutwirelessdistributedlearning.Frombottomtotop,thearchitectureconsistsoftheintelligentUElayer,thefirstintelligencelayerdeployedontheDU,thesecondintelligencelayerdeployedontheCU,thethirdintelligencelayerdeployedontheedgenode,andthefourthintelligencelayerdeployedonthecloud,asshowninthefollowingfigure.Differentintelligentlayersgeneratedifferentfunctionalconfigurationsfordifferentgoals,buildingadistributed,hierarchical,andintelligentRANarchitecturefor6Gnetworks.Onthenetwork,intelligentfunctionalcomponentscanbeflexiblyandquicklyorchestratedandused,multi-leveldataanalysisnetworkelementscanbedeployed,andadistributedcollaborativecontrolsystemcanbeformedatallnetworklayerstoachievedistributedintelligentinteractionandcollaborationbetweenwirelessnodeshorizontally[8].

8/25

Figure1SchematicDiagramofHierarchicalDeploymentofIntelligentFunctionalComponentsofa6GNetwork

InthenewdistributedRANarchitecture,federatedlearning(FL),asthemostpromisingdistributedintelligentcomputingframeworkfor6Ginfrastructure,canconductbroadermachinelearning(ML)whileprotectinguserdataprivacy.Itisexpectedtoplayanimportantrolein6Gintelligentservicesandapplications.ThroughintegrationbetweenFLandmulti-layernetworktopology,multi-levelfederatedaggregationcanbeperformed.Asshowninthefollowingfigure,athree-layernetwork(macrobasestation-microbasestation-UE)formsanFL-based,distributed,hierarchical,andintelligentRANarchitectureintheverticaldirection.Federatedaggregationcanbesplitintolow-levelfederatedaggregationatmicrobasestationsandhigh-levelfederatedaggregationatmacrobasestations.Thisensureslowcommunicationcostsandwiderdatasharing.

Figure2SchematicDiagramofFLNodeDeploymentatMultipleLayersofa6GNetwork

Specifically,ontheedgenetwork,earlymodelaggregationhaslowcommunicationcostsandcaneffectivelyalleviateuncertainmodelupdatesduetorandomlocaldata.Latermodelaggregationperformsmodelaggregationandupdatethroughhigh-levelFLservers,achievingmoreandwiderdatasharingandacceleratingconvergenceofglobalaggregation.Therefore,federatedaggregationcanbeflexiblydeployedatdifferentnetworklayersbasedontheactualneedsofthenetworkforlearningperformance,latency,capacity,energyconsumption,andotherindicatorstoachievedynamicmulti-levelFL.

Tobettersavenetworkenergy,thecommunicationfrequencyofhigh-levelglobalaggregationwithhighcommunicationcostscanbereduced.Thisreducesthecommunicationoverhead,soastoreducethetotalenergyconsumption.ComparedwiththetraditionalFLsolution,thedistributed,hierarchical,and

9/25

intelligentRANarchitecturethatintroducesmulti-levelfederatedaggregationcaneffectivelyreducenetworkenergyconsumptionunderthesamelearningaccuracy.

5.1.3WirelessIntelligentCloudNetwork

TheIMT-2030Framework[9]highlightstheimportanceofenvironmentaladaptabilityandenergysavingandemissionreductionofnetworksandUEs.Networkenergyefficiencyisaquantitativeindicatorthatattractsthemostattention,whichismeasuredinbit/Joule.

Around2010,majoroperatorsinChinasharedtheirobservationsonenergyconsumptionofmobilenetworks.Itwasfoundthathalfoftheenergyconsumptionisfromairconditionersandotherfacilities,andcentralizedRANdeviceswereproposedtoreduceenergyconsumption,whichiscalledcloudRAN(C-RAN).Thefocuscanbeontheradioorbasebandprocessing,asshowninthefollowingfigure.

Figure3C-RANSchematicDiagram

Accordingtoanalysis[10][11],C-RANiseco-friendlyinfrastructure.First,centralizedprocessingoftheC-RANarchitecturecanexponentiallyreducethenumberofbasestationsandsignificantlyreducepowerconsumptionofon-sitesupportdevices,suchasairconditioners.Second,thecooperativeradiotechnologycanreduceinterferencebetweenremoteradioheads(RRHs)andallowdenserRRHs.Therefore,thedistancefromRRHstoUEscanbeshortened,andsmallercellswithlowertransmitpowercanbedeployedwithoutaffectingnetworkcoveragequality.Theenergyrequiredforsignaltransmissionwillbereduced.ThishelpsreducepowerconsumptionintheRANandextendthebatterystandbytimeofUEs.Finally,thebasebandunit(BBU)poolisasharedresourceamongalargenumberofvirtualbasestations.Therefore,higherresourceutilizationandlowerpowerconsumptioncanbeachieved.Whenavirtualbasestationisidleatnightanddoesnotrequiremostofitsprocessingcapabilities,itcanbeshutdownorenterthelower-powerstate,whichdoesnotaffectthe24/7servicecommitment.

OpenRAN(O-RAN)inheritstheadvantagesofC-RANanddefinesopen

10/25

interfacesofO-CU/O-DU/O-RU.Thisensuresflexibleanalysisofthepowerconsumptionofdifferentnetworkentities.Thereport[12]pointsoutthatformostmobilenetworks,morethan80%ofenergyconsumptionisfromtheRAN,withtherestfromthecorenetwork,supportsystems,andrelatedcloudinfrastructure.Itisestimatedthatofthe80%energyconsumedontheRAN,approximately80%ofitisusedtopowerradios,andtheremaining20%ofitisusedfordistributedunits(DUs).Radioenergyconsumptioncanbegreatlyreducedbyusingnewtechnologies,suchasMicroSleepTxandmulti-bandradiodevicedesignandintegration.Inaddition,withahigh-performancegeneral-purposeprocessor,thecloud-basedDUconsumeslessenergyandiscompatiblewiththesoftwaredevelopmentecosystem.Torealizeitsfullpotential,theSMO/RICsoftwareiscarefullydesigned,andtherAPPmeetstheautomatedandnon-real-timenetworkmanagementrequirements.Thishelpsreduceoperatingcosts,improvenetworkperformance,andreduceenergyconsumption.

3GPPcarriedoutdiscussionsandstandardworkonnetworkenergysavingtechnologiesinRelease18andRelease19.Fordetails,seeChapter4.3GPPpointsoutthepotentialdirectionsforreducingRANenergyconsumption.TheO-RANarchitectureseparatesnetworkentitiesandprogrammablerAPPs,maximizingflexibility.Integratingnewenergysavingfeatures,theO-RANarchitectureisexpectedtobringbrightprospectsinthe6Gera.

5.2AirInterfaceEnergySavingTechnologies

5.2.1EnergySavingTechnologiesinSpatialDomain

ThepowerconsumptionoftheActiveAntennaUnit(AAU)accountsforabout80%thatofanNRbasestationandisthemaincomponentofnetworkenergyconsumption.Withtolerablesystemperformanceloss,spatialdomainenergysavingtechnologiescanadapttospatialelementsonthenetworktosignificantlyreducenetworkenergyconsumption.Dependingonthegranularity,spatialelementsmayincludeantennaelements,TxRUs,antennaports,antennapanels,andtransmissionandreceptionpoints(TRPs).Comparedwithsemi-staticspatialelementadaptation,dynamicspatialelementadaptationensuresfineradaptationgranularity,bettermatchestheserviceloadandactualtransmissionenvironment,andbetterservesUEs,thereforereducingenergyconsumptionofnetworks.

Massivemultiple-inputmultiple-output(MIMO)iswidelyusedonexisting5Gnetworksbecauseitsupportsspatialdomainmultiplexingormultipathdiversity.EventhoughmassiveMIMObringslargecapacity,italsoincreasespowerconsumptionofbasestationsduetoitslargenumberofTxRUsandrelatedhardwareprocessingunits(includingPAs).Aneffectivenetworkenergysaving

11/25

solutionturnson/offTxRUsofbasestationsbasedonthetrafficloadorthenumberofUEsserved.Asshowninthefollowingfigure,whenthenumberofUEsinacellbecomessmall,someTxRUsofbasestationscanbeturnedoffsothatthepowerconsumptionofbasestationscanbereducedwithoutaffectingthecapacity.LegacyspatialdomainshutdownsolutionscannotshutdownTxRUsdynamicallyandquickly.Forexample,thereisobviouscapacitylosswhenpowersavinggainsareobtained,orthesolutionscannotbeusedaroundtheclockbecausetheswitchingtimeforstaticshutdownistoolong.Asaresult,theantennastatusofabasestationcannotbequicklyadjustedbasedonthechannelstatus,thatis,theantennastatusdoesnotmatchthechannelstatus,resultinginagreatperformanceloss.

Figure4AdaptiveandDynamicTxRUShutdown

BasedonthedynamicloadandmultiplesetsofCSIcorrespondingtodifferentTxRUshutdownmodes,theentiretimeisdividedintomanymillisecond-levelschedulingwindows.Advanceddynamicshutdownusesadynamicschedulertomaximizeenergysavinggainswithlimitedcapacitylossbasedonenergyefficiency.Ineachschedulingwindow,allshutdownmodes(includingnoshutdown)aresimulatedandtraversed,soastoquicklyselecttheoptimalshutdownmode.Inthedynamicshutdownsolution,thehardwareresponsetimeinenergy-savingstateisakeyfactoraffectingnetworkindicatorsanduserexperience.Thehardwareresponsetimeneedstobeshortenedfromminute-leveltomillisecond-level,sothatenergyissavedallthetimeinsteadofidletimeonly.Inaddition,thenumberofhigh-levelconfigurationparametersforCSIreportsislimited.IftoomanyCSIreportsareconfiguredforadaptivechannelshutdown,theconfigurationofCSIreportsforotherusesmaybeaffected.Therefore,animportantresearchdirectioninthedynamicTxRUshutdownsolutionisCSIreportenhancementthatis,reportingmultiplesetsofCSIinoneCSIreportandreducingCSIoverhead.

ChannelshutdowncanreducethepowerconsumptionofthePA,andstaticpowerconsumptionoftheradiofrequency(RF)channel.Spatialdomainenergysavingtechnologiesbasedonchannelshutdownhaveobviousadvantagesinensuringservicecontinuity,arenotlimitedtobasestationswithlowserviceload,andareregardedasprevailingsolutionsforspatialdomainenergysaving.

12/25

Large-scaledistributedantennassupportmulti-TRP.Thatis,multipleantennapanelscanberegardedasapartoftheAAUinspatialdomain.Semi-static/dynamicadaptivemulti-TRPadjustmentisaspecialcaseofadaptivetransmissionantennaadjustment.Inactualtransmission,abettertransmissionlink(forexample,closerphysicaldistance)mayexistbetweenaUEandaTRP.AsusingmultipleTRPstotransmitdatafortheUEisnotalwaysnecessary,dynamicallyshuttingdownsomeTRPscansignificantlyreducethenetworkenergyconsumption.EspeciallyinfuturetransmissionsystemsthatdeploydistributedmassiveMIMOarraystosupportshort-rangetransmission,moreTRPsareusedfortransmission.Dynamicmulti-TRPshutdowncanbetterbalanceUEperformanceandnetworkenergyconsumption.

5.2.2EnergySavingTechnologiesinTimeDomain

Inmediumandlowsystemloadscenarios,methodssuchasdynamicandintelligentcellshutdown,intelligentDTX/DRXcoordination,andadaptivereductionofbroadcastsignaltransmissiontimeareusedtoensure