In-Network Channel Decoding Using%0D%0AConsensus on Log-Likelihood Ratio Averages英文电子资料

《In-Network Channel Decoding Using%0D%0AConsensus on Log-Likelihood Ratio Averages英文电子资料》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、1In-NetworkChannelDecodingUsingConsensusonLog-LikelihoodRatioAveragesyHaoZhu,AlfonsoCano,andGeorgiosB.GiannakisDepartmentofElectricalandComputerEngineering,UniversityofMinnesota,200UnionStreet,Minneapolis,MN55455,USA.Abstract—ThispaperconsiderstheproblemofdistributedThedi

2、stributedconsensusproblemhasattractedgreatinter-decodingofacodedmessagefromamobileaccesspoint(AP)estsintherecentyears.Thedistributedconsensusaveragingsenttoasensornetwork.Thedistributeddecodingproblemisproblemwasinvestigatedin[1],[2],fromtheperspectivesolvedusingaconsensu

3、salgorithmonthelog-likelihoodratioofacceleratingconvergenceandminimizingtheeffectsofaverages,whichensuresthatallsensorsattainthesamedecisionasifallreceivedsignalswereavailableatacentralizedlocation.inter-sensornoise.TheconsensusestimationproblemhasbeenUnlikeexistingdistri

4、butedhypothesistestingproblems,whoseconsideredby[3]thedeterministicsignals.Theintroductionofinter-sensorcommunicationcostgrowsexponentiallywiththebridgesensorsin[3]exploitedthedecomposablestructureofcodewordsize,thenovelconsensus-baseddistributeddecoderthecentralizedoptim

5、izationproblem’saugmentedLagrangianexploitsknowledgeovertheproblemtoreducethecosttobewiththemethodofmultipliers[4].Moreover,forthelattercaselinear.Linkfailuresandinter-sensorcommunicationnoisearealsoaccountedfor.Analysisandsimulationsconfirmthatthenoiserobustnessisprovedbo

6、ththeoreticallyandnumerically.resultantdistributeddecoderattainsbiterrorrateperformanceTheideaofcooperationamongnodesfordistributedde-approachingitscentralizedcounterpartwithinafewiterations.tectionproblemshasalsoattractedmuchattentionrecently.IndexTerms—WirelessSensorNet

7、works(WSNs),DistributedConsensus-baseddistributedhypothesistesting(DHT)basedDecoding,ConsensusAverageonthebeliefconsensusalgorithmisproposedin[5],anditisshownthatthisalgorithmworksforanyconnectednetworkswithloops.ThisbeliefconsensusalgorithmbenefitsfromI.INTRODUCTIONscalab

8、ility,robustnesstolinkfailures,andasynchronismintheWeconsidertheproblemofcooperativelydecodingam