资源描述:
《Analysis on crossover probability estimation using LDPC syndrome .pdf》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、SCIENCECHINAInformationSciences.RESEARCHPAPERS.September2011Vol.54No.9:1895–1904doi:10.1007/s11432-011-4311-yAnalysisoncrossoverprobabilityestimationusingLDPCsyndromeFANGYongCollegeofInformationEngineering,NorthwestA&FUniversity,Yangling712100,ChinaReceivedMarch19,2010;accep
2、tedApril12,2010;publishedonlineJune24,2011AbstractCorrelationestimationisacriticalissuethatimpactstheapplicationofSlepian-Wolfcoding(SWC)inpractice.Dynamiconlinecorrelationestimationisatypeofnewly-appearingapproaches,inwhichthedecoderestimatesthevirtualcorrelationchannelbetw
3、eentwocorrelatedsourcesusingbothsideinformationandthecompressedSWCbitstreamofthesource.SincethecompressedSWCbitstreamusuallycontainspartialinformationofthesource,theemergenceofdynamiconlinecorrelationestimationishelpfultosolvingtheproblemofcorrelationestimationintheSWCandfur
4、thermakestheSWCrealisable.Currently,theSWCisusuallyimplementedbyLDPCcodes.Inthiscase,theSWCbitstreamisjusttheLDPCsyndromeofthesource.IthasbeenrevealedthatthereareresidualredundanciesinLDPCsyndromes,whichcanbeusedtoestimatethecrossoverprobabilitybetweentwocorrelatedbinarysequ
5、ences.However,thisalgorithmhasnotbeenwelljustifiedyet.Thispapermakesuseofthecentrallimittheorem(CLT)toestablishamathematicmodelforanalyzingtheperformanceofthisalgorithm.Especially,forirregularLDPCcodes,theoptimizationofweightvectorsisdiscussedindetail.Representativeexperiment
6、alresultsareprovidedtovalidatetheanalysis.Keywordsdistributedsourcecoding,Slepian-Wolfcoding,LDPCsyndrome,correlationestimationCitationFangY.AnalysisoncrossoverprobabilityestimationusingLDPCsyndrome.SciChinaInfSci,2011,54:1895–1904,doi:10.1007/s11432-011-4311-y1IntroductionN
7、owadays,distributedsourcecoding(DSC)isgettingmoreandmoreapplicationsinpractice.In[1],SlepianandWolfprovedthatcorrelateddiscretesourcescanbecompressedwithoutlossefficiently,whetherornottheencodersofseparatesourcescommunicatewitheachother.WynerandZivsoonthereafterconsideredthepr
8、oblemoflossyDSCwithdecodersideinformation(SI).BothSlepian-WolftheoremandWyn