信道建模与仿真

ModelingandSimulatingtheWirelessChannelsProf.LiPinganWuhanUniversityofTech.Page:2FlatfadingchannelmodelTwoclassicalsimulationmodelearlyR.H.Clarke,BellSyst.Tech.J.,1968W.C.Jakes,MicrowaveMobileComm.,1974ConventionalflatfadingmodelJakesmodelClarkeGaussiannoisefiltermodelM.PatzoldfadingmodelincellmobilecommunicationsystemPage:3JakesmodelThesumofasetofsinusoidswithuniformlydistributiononphase)2cos(1tf)2cos(tfM)2cos(21tfmXXX+1cos2Mcos21cos2MXN2)2cos(1tf)2cos(tfM)2cos(21tfmXXX+1sin2Msin21sin2MXN2)(tTc)()()(tjTtTtTsc(N=4M+2)Page:4JakesmodelMnMnMnn,,2,114MnNnfMnffmmn,,2,12cos1mfwheredenotesthemaximalDopplerfrequencyPage:5ImprovedJakesmodel)2cos(11tf)2cos(211MmtfXXX+1cos2Mcos21cos2MXN2)2cos(11tf)2cos(MMtf)2cos(211MmtfXXX+1sin2Msin21sin2MXN2)(tTc)()()(tjTtTtTsc(N=4M+2))2cos(MMtfPage:6ImprovedJakesmodelMnMnMnn,,2,114MnNnfMnffmmn,,2,12cos1mfwheredenotesthemaximalDopplerfrequency,areindependentrandomvariablesuniformlydistributedoverforalln.)11(Mnn),[Page:7GaussiannoisefilterThesumoftwoquadratureindependentGaussianrandomvariablesIndependentcomplexGaussianwhitenoisesource12Ng*12Ng2Ng*2Ngmfmf12Ng*12Ng2Ng*2Ngmfmf)(fSDmfmf)(fSDmfmfIFFTIFFT22o90Page:8SuzukiprocessSuzukiprocessisaproductprocessofaRayleighprocessandlog-normalprocess)()()()(2221tt、tt)()()(21tjtt)2cos()(,,1,niniNnniitfcti)(3)(tsmetComplexGaussianGaussianRayleighLog-normal)(t)(t)(t)()()(ttt??Page:9PDFandauto-correlationforJakespowerspectraldensityfunctionTheautocorrelationfunctionThemeanpowerofmaxmax2maxmax20)/(1)(0ffffffffSi2,1),2()(max020itfJtriidffSi)(202,1iiPage:10Methodofequaldistances:Rayleigh-distribution212,21,,,02/)0(,,2,,,1),12(2)12(2iiiiiNiniininifiimfnicrNnffNnnNfnf24,25,2,1)]}arcsin()[arcsin({212/114,20NNiciiNnNnniPage:11PDFandauto-correlationfor3)2exp(21)(223ccffS2ln2ccfPowerspectraldensityfunctionTheautocorrelationfunction3dBcutoffF23)(2)(tcetrPage:12Methodofequaldistances:Log-normaldistribution4..,cccgef25;,,1)]}2)1(()2([2{332/133,3NNnNnerfNnerfcccn)12(23,3nNfccnTolimittheF-regionPage:13SuzukiProcessSimulator1,1c)2cos(1,11,1tf)2cos(2,12,1tf)2cos(11,1,1NNtfXXX+2,1c1,1Nc)2cos(1,21,2tf)2cos(2,22,2tf)2cos(11,2,2NNtfXXX+1,2c2,2c1,2Nc)(~t)(~1t)(~2t)(~t)2cos(1,31,3tf)2cos(2,32,3tf)2cos(11,3,3NNtfXXX+1,3c2,3c1,3NcX+e)(~3tsmX)(~t)(~)(~)(~tttRayleighProcessLog-normalProcessSuzukiProcessPage:14SimulationApproachesforSuzukiProcess等距离法MethodofEqualDistances(MED)等面积法MethodofEqualArea(MEA)蒙特卡罗方法MonteCarloMethod(MCM)均方误差法MeanSquareErrorMethod(MSEM)Page:15SimulationResultsbyusingMED瑞利过程仿真对数正态过程仿真器Page:16理论值和仿真值比较Itcanbeseenthatthesimulationmodelcoincidesverycloselytothetheoreticalmodel0,2exp)(222xxxxpii理论模型:Page:17理论值和仿真值比较Itcanbeseenthatthesimulationmodelcoincidesverycloselytothetheoreticalmodel0,2lnexp21)(2xsmxsxpxx理论模型:Page:18理论值和仿真值比较（相关函数）Page:19Multi-pathsChannelModel0TDL(Tap-Delay-Line)Modeltimetap1tap2tap3tap4XXXXRayleighFadingSimulatorRayleighFadingSimulatorRayleighFadingSimulatorRayleighFadingSimulator123TxRxPage:20Questions!