面向人工智能的低轨卫星接人网无线资源管控架构研究

打开文本图片集

中图分类号：TN911.22 文献标志码：A文章编号：1003-3114（2025）04-0647-09

Abstract：WiththerapidriseofbusinessessuchasbroadbandInternetandtheInternetofThings，applicationfieldsofsatelite communicationhavebeencontinuously expanding.However，theexponential growthofsateliteusers’servicedemands hasputforwardmorestringentrequirementsfosateliecommunicationsystemsintermsofcapacity，coverage，qualityofservice，anddyamic resource scheduling.Thestaticresourceallocationstrategyadoptedbytraditionalsatelitecommunicationsystemshasexposedmany problemssuchaslowresourceutlization，limitedsystemcapacity，andlargefluctuationsinservicequality.Althoughmulti-beam technologyhasmadesomeprogress，existing technicalsystemstilfcestheoreticalbotlenecksthataredificulttobreakthrough whenfacingcomplexandchangeablecommunicationenvironments.Inviewofthesignificantchallengesofhigh dynamics，strong heterogeneityandservicediversityinteLowEarthOrbit（LEO）sateliteInterescenariothispape，referrigtothedision resultsof the3GPPServiceandSystemArchitecture WorkGroupontheAIservicesystemarchitecture，innovativelyproposesasatellitenetworkresourcemanagementandcontrolarchitecturethatintegratescentralizedanddistributedcontrolideas.Thisacitecture adoptsahierarchicalmodulardesign.Bybuildingamulti-levelcolaborativeworking mechanism，itaimstoachieveefectiveallcationofspaceandgroundresources.Inadditiontispaperdeeplyexploreskeycapabilitisofthisachitecture，coveringfour-dimensionalresourcemanagementoftiefreqencypaceandcodeupportforthe5thenerationon-TestrialNetwork（5GNTN） system，satelltegroundcoordinaonmehanisms，asellsteeplmntadinteractionsigideasofIagorisIise pectedthat through thesestudies，strong technical supportcanbeprovided forempowering LEOsateliteRadioAcessNetwork （RAN）withAI，helpingLEOsatelitecommunicationsystems toachieveacomprehensiveperformanceimprovementandleap-forward development.

Keywords： LEO satellite; AI; resource management; RAN

0引言

过去几十年，卫星通信技术取得了飞速发展，并在全球范围内得到了广泛应用。（剩余13145字）