altera原版文章数字预矫正（一）.pptVIP

DigitalPredistortion

AgendaIntroductionAlgorithmStandardlookuptablemethodPhaserelatederrorsMemoryeffectImplementationMultipliersMemoryCordicProcessors

Introduction:Purpose

Introduction:PredistortionVRFVinVRFVinOverallLinear

ResponseVrf=kVinVRFVinPAVrf=kVinIdealPAVdVPAPAVrf=fnlkVdRealPADPDVd=1/fnlVinVdVinPredistorter

Algorithms:OverviewAdaptiveLookupTable(LUT)LookuptableforphaseandmagnitudecorrectionvaluesDealswithmagnitudedependenterrorsVolterramodellingofPADirectimplementationIndirectLUTimplementation

Algorithm:DistanceGradientMethodAssumption:Erroronlydependsonmagnitudeφ

arctan(I/Q)LUT

(?r?φ)IQ

Demodaddressr

(I2+Q2)1/2φ

arctan(I/Q)r

(I2+Q2)1/2delaydelayQ

r*cos(φ)I

r*sin(φ)IQ

modPAI,QinI,Qout·(-1)·(-1)

Algorithm:DistanceGradientMethodMATLABsimulationresultsusingSALEHPAmodelEVMimprovedinregionof90%

Dist–GradientLUTFreqPlotsPredistortionimprovesACLRby70dB(considering700-900and1100-1300asside-bandregionsformeasurements)Simplifiedsimulationenvironment(usingSALEHPAmodel)

Algorithm:Phaserelatederror

Algorithm:MemoryeffectModelseffectofshorttermtemperatureincreaseonsiliconThreepossibleapproachesAdd“deltalook-uptable”(Intersilsolution)AddnewdimensiontoLUT(fullyadaptive)UseFIRinmagnitudeaddresscalculation

Algorithm:MemoryeffectPAmodels:Saleh[1]Memoryless,ErrorsonlydependentoninputmagnitudeVolterrabased,withorwithoutmemory:[2,3]Errordependoninputmagnitudeandphase[1]A.Salehand,AdaptiveLinearizationofPowerAmplifiersinDigitalRadioSystems,BellSyst.Tech.J.,Vol.62,No.4,pp1019-1033,Apr.1983[2],,,,,and,Memorypolynomialpredistorterbasedontheindirectlearningarchitecture,[3].H.Qian,andG.T.Zhou,Aneuralnetworkpredistorterfornonlinearpoweramplifierswithmemory,Proc.10thIEEEDSPWorkshop(DSP’2002),PineMountain,GA,October2002

Algorithm:MemoryeffectErrormechanism: