GNURAdioDoc-9

来源：年旅网

Tutorial9:ADictionaryoftheGNURadioblocks

DaweiShen∗

August21,2005

Abstract

AdictionaryoftheGNURadioblocks.Thisarticletakesatouraroundthemostfrequentlyusedblocks,explainingthesyntaxandhowtousethem.

1Introduction

WhenweuseMatlabtodosimulation,itisbelievedthatinordertowritethecodecleanlyandeﬃciently,weneedtomemorizeanumberofMatlabbuilt-infunctionsandtoolboxeswellandusethemskillfully.ThesameappliestoGNURadio.About100frequentlyusedblockscomewithGNURadio.Foracertainnumberofapplications,wecancompletethedesigningusingtheseexistingblocks,programmingonlyonthePythonlevel,withouttheneedtowriteourownblocks.Sointhisarticle,wewilltakeatouraroundtheGNURadioblocks.

AgoodwaytogothroughtheblocksistostudythetwoGNURadiodocumentationsgeneratedbyDoxygen.Theywillbegeneratedwhenyouinstallgnuradio-coreandusrppackages,assumingyouhaveDoxygeninstalled.Theyarelocatedat:

/usr/local/share/doc/gnuradio-core-x.xcvs/html/index.html/usr/local/share/doc/usrp-x.xcvs/html/index.html

Youmayliketobookmarktheminyourwebbrowser.Theﬁrstoneisalsoavailableon-line.AblockisactuallyaclassimplementedinC++.Forexample,intheFMreceiverexample,weusetheblockgr.quadrature_demod_cf.Thisblockcorrespondstotheclassgr_quadrature_demod_cfimplementedinC++.TheSWIGcreatesitsinterfacetoPython.TheSWIGdoessomemagicworksothatfromPython’spointofview,theblockbecomesaclasscalledquadrature_demod_cfdeﬁnedinthemodulegr,sothatwecanaccesstotheblockusinggr.quadrature_demod_cfinPython.WhenyoulookattheDoxygendocumentations,thepreﬁxsuchasgr,qa,usrpcorrespondstothemodulenameinPythonandthepartafterthepreﬁxistherealnameoftheblockinthatmodule,suchasquadrature_demod_cf,fir_filter_fff.Soifwetalkaboutablocknamed‘A_B_C_...’,wewilluseitas‘A.B_C_...’inPython.

authorisaﬃliatedwithNetworkingCommunicationsandInformationProcessing(NCIP)Laboratory,

DepartmentofElectricalEngineering,UniversityofNotreDame.Welcometosendmeyourcommentsorinquiries.Email:dshen@nd.edu

∗The

2Tutorial9-DictionaryofGNURadioblocks

2.1

SignalSources

Sinusoidalandconstantsources

Block:gr.sig_source_X.Usage:

gr.sig_source_c[f,i,s]

(doublesampling_freq,gr_waveform_twaveform,doublefrequency,doubleamplitude,

gr_complex[float,integer,short]offset)

Notes:ThesuﬃxXindicatesthedatatypeofthesignalsource.Itcanbec(complex),f(ﬂoat),i(4byteinteger)ors(2byteshortinteger).Theoffsetargumentisthesametypeasthesignal.Thewaveformargumentindicatesthetypeofthewaveform.gr_waveform_tisanenumerationtypedeﬁnedingr_sig_source_waveform.h.Itsvaluecanbe:

gr.GR_CONST_WAVEgr.GR_COS_WAVEgr.GR_SIN_WAVE

Whenyouusegr.GR_CONST_WAVE,theoutputwillbeaconstantvoltage,whichistheamplitudeplustheoﬀset.

2.2Noisesources

Block:gr.noise_source_X.Usage:

gr.noise_source_c[f,i,s](gr_noise_type_ttype,

floatamplitude,longseed)

Notes:ThesuﬃxXindicatesthedatatypeofthesignalsource.Itcanbec(complex),f(ﬂoat),i(4byteinteger)ors(2byteshortinteger).Thetypeargumentindicatesthetypeofthenoise.gr_noise_type_tisanenumerationtypedeﬁnedingr_noise_type.h.Itsvaluecanbe:

GR_UNIFORMGR_GAUSSIANGR_LAPLACIANGR_IMPULSE

ChoosingGR_UNIFORMgeneratesuniformlydistributedsignalbetween[-amplitude,amplitude].GR_GAUSSIANgivesusnormallydistributeddeviatewithzeromeanandvarianceamplitude2.Sim-ilarly,GR_LAPLACIANandGR_IMPLUSErepresentaLaplaciandistributionandaimpulsedistributionrespectively.Allthesenoisesourceblocksarebasedonthepseudorandomnumbergeneratorblockgr.random.Youmaytakealookat/gnuradio-core/src/lib/general/gr_random.h(cc)toseehowtogeneratearandomnumberfollowingvariousdistributions.

D.Shen3

2.3Nullsources

Block:gr.null_source.Usage:

gr.null_source(size_t

sizeof_stream_item)

Notes:gr.null_sourceproducesconstantzeros.Theargumentsizeof_stream_itemspeciﬁesthedatatypeofthezerostream,suchasgr_complex,float,integerandsoon.

2.4Vectorsources

Block:gr.vector_source_X.Usage:

gr.vector_source_c[f,i,s,b](conststd::vector&data,

boolrepeat=false)(gr_complexcanbereplacedbyfloat,integer,short,unsignedchar)

Notes:ThesuﬃxXindicatesthedatatypeofthesignalsource.Itcanbec(complex),f(ﬂoat),

i(4byteinteger),s(2byteshortinteger),orb(1byteunsignedchar).Thesesourcesgettheirdatafromavector.Theargumentrepeatdecideswhetherthedatainthevectorissentrepeatedly.Asanexample,wecanusetheblockinthiswayinPython:

src_data=(-3,4,-5.5,2,3)

src=gr.vector_source_f(src_data)

2.5Filesources

Block:gr.file_sourceUsage:

gr.file_source(size_t

constchar*bool

itemsize,filename,repeat)

Notes:gr.file_sourcereadsthedatastreamfromaﬁle.Thenameoftheﬁleisspeciﬁedbyfilename.Theﬁrstargumentitemsizedeterminesthedatatypeofthestream,suchasgr_complex,float,unsignedchar.Theargumentrepeatdecideswhetherthedataintheﬁleissentrepeatedly.Asanexample,wecanusetheblockinthiswayinPython:

src=gr.file_source(gr.sizeof_char,\"/home/dshen/payload.dat\TRUE)

2.6Audiosource

Block:gr.audio_sourceUsage:

gr.audio_source(intsampling_rate)

Tutorial9-DictionaryofGNURadioblocks

Notes:audio_sourcereadsdatafromtheaudioline-in.Theargumentsampling_ratespeciﬁesthedatarateofthesource,insamplespersecond.

2.7USRPsource

Block:usrp.source_c[s]Usage:

usrp.source_c(s)(int

unsignedintintintint

which_board,decim_rate,nchan=1,mux=-1,mode=0)

Notes:whenyoudesignareceiverusingGNURadio,i.e.workingontheRXpath,probablyyouneedtheUSRPasthedatasource.Thesuﬃxc(complex),ors(short)speciﬁesthedatatypeofthestreamfromUSRP.Mostlikelythecomplexsource(I/Qquadraturefromthedigitaldownconverter(DDC))wouldbemorefrequentlyused.Someoftheargumentshavebeenintroducedintutorial4.which_boardspeciﬁeswhichUSRPtoopen,whichisprobably0ifthereisonlyoneUSRPboard.decim_ratetellsthedigitaldownconverter(DDC)thedecimationfactorD.nchanspeciﬁesthenumberofchannels,1,2or4.muxsetsinputMUXconﬁguration,whichdetermineswhichADC(orconstantzero)isconnectedtoeachDDCinput(seetutorial4fordetails).‘-1’meanswepreservethedefaultsettings.modesetstheFPGAmode,whichweseldomtouch.Thedefaultvalueis0,representingthenormalmode.Quiteoftenweonlyspecifytheﬁrsttwoarguments,usingthedefaultvaluesfortheothers.Forexample:usrp_decim=250

src=usrp.source_c(0,usrp_decim)

3.1

SignalSinks

Nullsinks

Block:gr.null_sink.Usage:

gr.null_sink(size_t

sizeof_stream_item)

Notes:gr.null_sinkdoesnothingbuteatupyourstream.Theargumentsizeof_stream_itemspeciﬁesthedatatypeofthezerostream,suchasgr_complex,float,integerandsoon.

3.2Vectorsinks

Block:gr.vector_sink_X.Usage:

D.Shen5

gr.vector_sink_c[f,i,s,b]()

Notes:ThesuﬃxXindicatesthedatatypeofthesignalsink.Itcanbec(complex),f(ﬂoat),i(4byteinteger),s(2byteshortinteger),orb(1byteunsignedchar).Thesesinkswritethedataintoavector.Asanexample,wecanusetheblockinthiswayinPython:dst=gr.vector_sink_f()

3.3Filesinks

Block:gr.file_sinkUsage:

gr.file_sink(size_t

constchar*

itemsize,filename)

Notes:gr.file_sourcewritesthedatastreamtoaﬁle.Thenameoftheﬁleisspeciﬁedbyfilename.Theﬁrstargumentitemsizedeterminesthedatatypeoftheinputstream,suchasgr_complex,float,unsignedchar.Asanexample,wecanusetheblockinthiswayinPython:src=gr.file_source(gr.sizeof_char,\"/home/dshen/rx.dat\")

3.4Audiosink

Block:gr.audio_sinkUsage:

gr.audio_source(intsampling_rate)

Notes:Whenyouﬁnishthesignalprocessingandwishtoplaythesignalthroughthespeaker,youshouldusetheaudiosink.audio_sinkwilloutputthedatatothesoundcardatthesamplingrateofsampling_rate.

3.5USRPsink

Block:usrp.sink_c[s]Usage:

usrp.sink_c(s)(int

unsignedintintint

which_board,interp_rate,nchan=1,mux=-1)

Notes:whenyoudesignatransmitterusingGNURadio,i.e.workingontheTXpath,probablyyouneedtheUSRPasthedatasink.Thesuﬃxc(complex),ors(short)speciﬁesthedatatypeofthestreamgoingintoUSRP.Mostlikelythecomplexsinkwouldbemorefrequentlyused.Someof

6Tutorial9-DictionaryofGNURadioblocks

theargumentshavebeenintroducedintutorial4.which_boardspeciﬁeswhichUSRPtoopen,whichisprobably0ifthereisonlyoneUSRPboard.interp_ratetellstheinterpolatorontheFPGAtheinterpolationfactorI.NotethattheinterpolationrateImustbeintherange[4,512],andmustbeamultipleof4.nchanspeciﬁesthenumberofchannels,1or2.muxsetsoutputMUXconﬁguration,whichdetermineswhichDACisconnectedtoeachinterpolatoroutput(ordisabled).‘-1’meanswepreservethedefaultsettings.Quiteoftenweonlyspecifytheﬁrsttwoarguments,usingthedefaultvaluesfortheothers.Forexample:usrp_interp=256

src=usrp.sink_c(0,usrp_interp)

4.1

SimpleOperators

Addingaconstant

Block:gr.add_const_XXUsage:

gr.add_const_cc[ff,ii,ss,sf](gr_complex[float,integer,short]k)

Notes:ThesuﬃxXXcontainstwocharacters.Theﬁrstoneindicatesthedatatypeoftheinputstreamwhilethesecondonetellsthetypeoftheoutputstream.Itcanbecc(complextocomplex),ff(ﬂoattoﬂoat),ii(4byteintegertointeger),ss(2byteshortintegertoshortinteger)orsf(shortintegertoﬂoat).Thegr.add_const_XXblockaddsaconstanttotheinputstreamsothatthesignalhasadiﬀerentoﬀset.Notethatforgr.add_const_sf,theargumentkisﬂoat.Weaddaﬂoatconstanttoashortintegerstream,andtheoutputstreambecomesﬂoat.

4.2Adder

Block:gr.add_XXUsage:

gr.add_cc[ff,ii,ss]()

Notes:ThesuﬃxXXindicatesthedatatypeoftheinputandoutputstreams.gr.add_XXaddsallinputstreamstogether.Thetypeofeachincomingstreamandmustbethesame.WhenweuseitinPython,multipleupstreamblockcouldbeconnectedtoit.Forexample:adder=gr.add_cc()

fg.connect(stream1,(adder,0))fg.connect(stream2,(adder,1))fg.connect(stream3,(adder,2))fg.connect(adder,outputstream)

Thentheoutputoftheadderisstream1+stream2+stream3.

4.3Subtractor

Block:gr.sub_XX

D.Shen7

Usage:

gr.sub_cc[ff,ii,ss]()

Notes:ThesuﬃxXXindicatesthedatatypeoftheinputandoutputstreams.gr.sub_XXsubtractsacrossallinputstreams.output=input_0-input_1-input_2....Thetypeofeachincomingstreamandmustbethesame.WhenweuseitinPython,multipleupstreamblockcouldbeconnectedtoit.Forexample:subtractorfg.connectfg.connectfg.connectfg.connect

=gr.sub_cc()

(stream1,(subtractor,0))(stream2,(subtractor,1))(stream3,(subtractor,2))(subtractor,outputstream)

Thentheoutputofthesubtractorisstream1-stream2-stream3.

4.4Multiplyingaconstant

Block:gr.multiply_const_XXUsage:

gr.add_const_cc[ff,ii,ss](gr_complex[float,integer,short]k)

Notes:ThesuﬃxXXindicatesthedatatypeoftheinputandoutputstreams.Thegr.multiply_const_XXblockmultipliestheinputstreambyaconstantk.

4.5Multiplier

Block:gr.multiply_XXUsage:

gr.multiply_cc[ff,ii,ss]()

Notes:ThesuﬃxXXindicatesthedatatypeoftheinputandoutputstreams.gr.multiply_XXcomputestheproductofallinputstreams.Thetypeofeachincomingstreamandmustbethesame.WhenweuseitinPython,multipleupstreamblockcouldbeconnectedtoit.Forexample:multiplierfg.connectfg.connectfg.connectfg.connect

=gr.multiply_cc()

(stream1,(multiplier,0))(stream2,(multiplier,1))(stream3,(multiplier,2))(multiplier,outputstream)

Thentheoutputofthemultiplierisstream1×stream2×stream3.

4.6Divider

Block:gr.divide_XXUsage:

gr.divide_cc[ff,ii,ss]()

Tutorial9-DictionaryofGNURadioblocks

Notes:ThesuﬃxXXindicatesthedatatypeoftheinputandoutputstreams.gr.divide_XXdividesacrossallinputstreams.output=input_0/input_1/input_2....Thetypeofeachincomingstreamandmustbethesame.WhenweuseitinPython,multipleupstreamblockcouldbeconnectedtoit.Forexample:divider=gr.divide_cc()

fg.connect(stream1,(divider,0))fg.connect(stream2,(divider,1))fg.connect(stream3,(divider,2))fg.connect(divider,outputstream)

Thentheoutputofthesubtractorisstream1÷stream2÷stream3.

4.7Logfunction

Block:gr.nlog10_ffUsage:

gr.nlog10_ff(floatn,

unsignedvlen)

Notes:gr.nlog10_ffcomputesn×log10(input).Inputandoutputarebothﬂoatnumbers.Ignoretheargumentvlen,thevectorlength,usingitsdefaultvalue1.

5.1

TypeConversions

ComplexConversions

Block:

gr.complex_to_floatgr.complex_to_realgr.complex_to_imaggr.complex_to_maggr.complex_to_argUsage:

gr.complex_to_float(unsignedintvlen)gr.complex_to_real(unsignedintvlen)gr.complex_to_imag(unsignedintvlen)gr.complex_to_mag(unsignedintvlen)gr.complex_to_arg(unsignedintvlen)

Notes:Theargumentvlenisthevectorlength,wealmostalwaysusethedefaultvalue1.Sojustignoretheargument.Theseblocksconvertacomplexsignaltoseparatereal&imaginaryﬂoatstreams,justtherealpart,justtheimaginarypart,themagnitudeofthecomplexsignalandthephaseangleofthecomplexsignal.Notethattheblockgr.complex_to_floatcouldhave1or2outputs.Ifitisconnectedtoonlyoneoutput,thentheoutputistherealpartofthecomplexsignal.Itseﬀectisequivalenttogr.complex_to_real.

D.Shen9

5.2

Block:

FloatConversions

gr.float_to_complexgr.float_to_shortgr.short_to_floatUsage:

gr.float_to_complex()gr.float_to_short()gr.short_to_float()

Notes:Theseblocksarelike‘adapters’,usedtoprovidetheinterfacebetweentwoblockswithdiﬀerenttypes.Notethattheblockgr.float_to_complexmayhave1or2inputs.Ifthereisonlyone,thentheinputsignalistherealpartoftheoutputsignal,whiletheimaginarypartisconstant0.Iftherearetwoinputs,theyserveastherealandimaginarypartsoftheoutputsignalrespectively.

6.1

Filters

FIRDesigner

Block:gr.firdes

Notes:gr.firdeshasseveralstaticpublicmemberfunctionsusedtodesigndiﬀerenttypesofFIRﬁlters.ThesefunctionsreturnavectorcontainingtheFIRcoeﬃcients.ThereturnedvectorisoftenusedasanargumentofotherFIRﬁlterblocks.

6.1.1Usage:

LowPassFilter

vectorgr.firdes::low_pass(double

doubledoubledoublewin_typedoublegain,

sampling_freq,cutoff_freq,

transition_width,

window=WIN_HAMMING,beta=6.76)[static]

Notes:low_passisastaticpublicmemberfunctionintheclassgr.firdes.ItdesignstheFIR

ﬁltercoeﬃcients(taps)giventheﬁlterspeciﬁcations.HeretheargumentwindowisthetypeofthewindowusedintheFIRﬁlterdesign.Validvaluesinclude

WIN_HAMMINGWIN_HANN

WIN_BLACKMANWIN_RECTANGULAR

106.1.2Usage:

vectorgr.firdes::high_pass(double

doubledoubledoublewin_typedouble

HighPassFilter

Tutorial9-DictionaryofGNURadioblocks

gain,

sampling_freq,cutoff_freq,

transition_width,

window=WIN_HAMMING,beta=6.76)[static]

6.1.3Usage:

BandPassFilter

vectorgr.firdes::band_pass(double

doubledoubledoubledoublewin_typedoublegain,

sampling_freq,low_cutoff_freq,high_cutoff_freq,transition_width,

window=WIN_HAMMING,beta=6.76)[static]

6.1.4Usage:

BandRejectFilter

vectorgr.firdes::band_reject(double

doubledoubledoubledoublewin_typedoublegain,

sampling_freq,low_cutoff_freq,high_cutoff_freq,transition_width,

window=WIN_HAMMING,beta=6.76)[static]

6.1.5Usage:

HilbertFilter

vectorgr.firdes::hilbert(unsignedintntaps,

win_typewindowtype=WIN_RECTANGULAR,doublebeta=6.76)[static]Notes:gr.firdes::hilbertdesignsaHilberttransformﬁlter.ntapsisthenumberoftaps,which

mustbeodd.6.1.6Usage:

RaisedCosineFilter

D.Shen11

gain,

sampling_freq,symbol_rate,alpha,ntaps)[static]

vectorgr.firdes::root_raised_cosine(double

doubledoubledoubleint

Notes:gr.firdes::root_raised_cosinedesignsarootcosineFIRﬁlter.Theargumentalphaistheexcessbandwidthfactor.ntapsisthenumberoftaps.

6.1.7Usage:

GaussianFilter

vectorgr.firdes::gaussian(double

doubledoubledoubleintgain,

sampling_freq,symbol_rate,bt,

ntaps)[static]

Notes:gr.firdes::gaussiandesignsaGaussianﬁlter.Theargumentntapsisthenumberof

taps.

6.2

Block:

FIRDecimationFilters

gr.fir_filter_cccgr.fir_filter_ccfgr.fir_filter_fccgr.fir_filter_fffgr.fir_filter_fsfgr.fir_filter_sccUsage:

gr.fir_filter_ccc(intdecimation,

conststd::vector<

gr.fir_filter_ccf(intdecimation,

conststd::vector<

gr.fir_filter_fcc(intdecimation,

conststd::vector<

gr.fir_filter_fff(intdecimation,

conststd::vector<

gr.fir_filter_fsf(intdecimation,

conststd::vector<

gr.fir_filter_scc(intdecimation,

conststd::vector<

gr_complex>&taps)float>&taps)gr_complex>&taps)float>&taps)float>&taps)gr_complex>&taps)

Notes:Theseblocksallhavea3-charactersuﬃx.Theﬁrstoneindicatesthedatatypeoftheinputstream,thesecondonetellsthetypeoftheoutputstream,andthelastonerepresentsforthedatatypeoftheFIRﬁltertaps.

12Tutorial9-DictionaryofGNURadioblocks

Eachblockhastwoarguments.TheﬁrstoneisthedecimationrateoftheFIRﬁlter.Ifitis1,thenit’sjustanormal1:1FIRﬁlter.ThesecondargumenttapsisthevectoroftheFIRcoeﬃcients,whichwegetfromtheFIRﬁlterdesignerblockgr.firdes.

6.3

Block:

FIRInterpolationFilters

gr.interp_fir_filter_cccgr.interp_fir_filter_ccfgr.interp_fir_filter_fccgr.interp_fir_filter_fffgr.interp_fir_filter_fsfgr.interp_fir_filter_sccUsage:

gr.interp_fir_filter_ccc(unsignedinterpolation,

conststd::vector&

gr.interp_fir_filter_ccf(unsignedinterpolation,

conststd::vector&taps

gr.interp_fir_filter_fcc(unsignedinterpolation,

conststd::vector&

gr.interp_fir_filter_fff(unsignedinterpolation,

conststd::vector&taps

gr.interp_fir_filter_fsf(unsignedinterpolation,

conststd::vector&taps

gr.interp_fir_filter_scc(unsignedinterpolation,

conststd::vector&

taps))taps)))taps)

Notes:Theseblocksallhavea3-charactersuﬃx.Theﬁrstoneindicatesthedatatypeoftheinputstream,thesecondonetellsthetypeoftheoutputstream,andthelastonerepresentsforthedatatypeoftheFIRﬁltertaps.

Eachblockhastwoarguments.TheﬁrstoneistheinterpolationrateoftheFIRﬁlter.ThesecondargumenttapsisthevectoroftheFIRcoeﬃcients,whichwegetfromtheFIRﬁlterdesignerblockgr.firdes.

6.4

Block:

DigitalDownConverterwithFIRDecimationFilters

gr.freq_xlating_fir_filter_cccgr.freq_xlating_fir_filter_ccfgr.freq_xlating_fir_filter_fccgr.freq_xlating_fir_filter_fcfgr.freq_xlating_fir_filter_sccgr.freq_xlating_fir_filter_scfUsage:

gr.freq_xlating_fir_filter_ccc[ccf,fcc,fcf,scc,scf]

D.Shen13

(intdecimation,

conststd::vector&doublecenter_freq,doublesampling_freq)

taps,

Notes:Theseblocksallhavea3-charactersuﬃx.Theﬁrstoneindicatesthedatatypeoftheinput

stream,thesecondonetellsthetypeoftheoutputstream,andthelastonerepresentsforthedatatypeoftheFIRﬁltertaps.

TheseblocksareusedasFIRﬁlterscombinedwithfrequencytranslation.Recallthatintutorial4,intheFPGAtherearedigitaldownconverters(DDC)translatingthesignalfromIFtobaseband.Thenthefollowingdecimatordownsamplesthesignalandselectsanarrowerbandofthesignalbylow-passﬁlteringit.Theseblocksdothesamethingsexceptinsoftware.Theseclasseseﬃcientlycombineafrequencytranslation(typicallydownconversion)withanFIRﬁlter(typicallylow-pass)anddecimation.Itisideallysuitedfora‘channelselectionﬁlter’andcanbeeﬃcientlyusedtoselectanddecimateanarrowbandsignaloutofwidebandwidthinput.

6.5HilbertTransformFilter

Block:gr.hilbert_fcUsage:

gr.hilbert_fc(unsignedintntaps)

Notes:gr.hilbert_fcisaHilberttransformer.Therealoutputisinputappropriatelydelayed.ImaginaryoutputisHilbertﬁltered(90degreephaseshift)versionofinput.Weonlyneedtospecifythenumberoftapsntapswhenusingtheblock.TheHilbertﬁlterdesignergr.firdes::hilbertisusedimplicitlyintheimplementationoftheblockgr.hilbert_fc.

6.6Filter-DelayCombinationFilter

Block:gr.filter_delay_fcUsage:

gr.filter_delay_fc(conststd::vector&

taps)

Notes:Thisisaﬁlter-delaycombinationblock.Theblocktakesoneortwoﬂoatstreamandoutputsacomplexstream.Ifonlyoneﬂoatstreamisinput,therealoutputisadelayedversionofthisinputandtheimaginaryoutputistheﬁlteredoutput.Iftwoﬂoatsareconnectedtotheinput,thentherealoutputisthedelayedversionoftheﬁrstinput,andtheimaginaryoutputistheﬁlteredoutput.Thedelayintherealpathaccountsforthegroupdelayintroducedbytheﬁlterintheimaginarypath.Theﬁltertapsneedstobecalculatedusinggr.firdesbeforeinitializingthisblock.

6.7IIRFilter

Block:gr.iir_filter_ffdUsage:

gr.iir_filter_ffd(conststd::vector&

conststd::vector&

fftaps,fbtaps)

14Tutorial9-DictionaryofGNURadioblocks

Notes:Thesuﬃxffdmeansﬂoatinput,ﬂoatoutputanddoubletaps.ThisIIRﬁlterusestheDirectFormIimplementation,wherefftapscontainsthefeed-forwardtaps,andfbtapsthefeedbackones.fftapsandfbtapsmusthaveequalnumbersoftaps.Theinputandoutputsatisfyadiﬀerenceequationoftheform

MN󰀁󰀁

bkx[n−k]y[n]−aky[n−k]=

k=1

k=0

withthecorrespondingrationalsystemfunction

󰀂M

H(z)=

1−

−k

k=0bkz󰀂N−k

k=1akz

Notethatsometextsdeﬁnethesystemfunctionwitha‘+’inthedenominator.Ifyou’reusingthatconvention,you’llneedtonegatethefeedbacktaps.

6.8SinglePoleIIRFilter

Block:gr.single_pole_iir_filter_ffUsage:

gr.single_pole_iir_filter_ff(doublealpha,

unsignedint

vlen)

Notes:ThisisasinglepoleIIRﬁlterwithﬂoatinput,ﬂoatoutput.Theinputandoutputsatisfyadiﬀerenceequationoftheform:

y[n]−(1−α)y[n−1]=αx[n]

withthecorrespondingrationalsystemfunction

H(z)=

1−(1−α)z−1

Notethatsometextsdeﬁnethesystemfunctionwitha+inthedenominator.Ifyou’reusingthatconvention,you’llneedtonegatethefeedbacktap.Theargumentvlenisthevectorlength.Weusuallyuseitsdefaultvalue1,sojustignoreit.

7FFT

Block:

gr.fft_vccgr.fft_vfcUsage:

gr.fft_vcc(int

boolbool

gr.fft_vfc(int

boolbool

fft_size,forward,window)fft_size,forward,window)

Notes:TheseblocksareusedtocomputethefastFouriertransformsoftheinputsequence.Forgr.fft_vcc,itcomputesforwardorreverseFFT,withcomplexvectorinandcomplexvectorout.Forgr.fft_vfcitcomputesforwardFFTwithﬂoatvectorinandcomplexvectorout.

D.Shen15

8.1

Otherusefulblocks

FMModulationandDemodulation

Block:

gr.frequency_modulator_fcgr.quadrature_demod_cfUsage:

gr.frequency_modulator_fc(doublesensitivity)gr.quadrature_demod_cf(floatgain)

Notes:Thegr.frequency_modulator_fcblockistheFMmodulator.Theinstantaneousfre-quencyoftheoutputcomplexsignalisproportionaltotheinputﬂoatsignal.Wehaveseengr.quadrature_demod_cfintheFMreceiverexample.Itcalculatestheinstantaneousfrequencyoftheinputsignal.

8.2NumericallyControlledOscillator

Block:gr.fxpt_nco

Usage:Thisblockisusedtogeneratesinusoidalwaves.Wecansetoradjustthephaseandfrequencyoftheoscillatorbyseveralpublicmemberfunctionsoftheclass.Thefunctionsinclude:

voidvoidvoidvoidvoidvoidfloatfloatvoidfloatfloat

set_phase(floatangle)

adjust_phase(floatdelta_phase)set_freq(floatangle_rate)

adjust_freq(floatdelta_angle_rate)step()

step(intn)

get_phase()constget_freq()const

sincos(float*sinx,float*cosx)constcos()constsin()const

Weusecos()orsin()functiontogetthesinusoidalsamples.Theycalculatesinandcosvaluesaccordingtothecurrentphase.Thefreqisactuallythephasediﬀerencebetweenconsecutivesamples.Whenwecallstep()method,thecurrentphasewillincreasebythefreq.

8.3

Block:

Blocksfordigitaltransmission

gr.bytes_to_symsgr.simple_framer

gr.simple_correlatorUsage:

gr.bytes_to_syms()

gr.simple_framer(intpayload_bytesize)

gr.simple_correlator(intpayload_bytesize)

Tutorial9-DictionaryofGNURadioblocks

Notes:gr.bytes_to_syms()convertsabytesequence(forexample,aunsignedcharstream)toa±sequence,i.e.thedigitalbinarysymbols.gr.simple_framerpacksabytestreamintopackets,withthelengthofpayload_bytesize.Thennecessarysynchronizationandcommandbytesareaddedatthehead.gr.simple_correlatorisadigitaldetector,whichsynchronizesthesymbolandframe,ﬁnallydetectsthedigitalinformationcorrectly.Therealimplementationoftheseblocksisalittlebitcomplicated.PleasestudytheFSKexamplefsk_r[t]x.pyandthesourcecodeoftheseblocks.Theseblocksareveryhelpfulwhenwewanttodesigndigitaltransmissionschemes.

9Wrapup

ThistutorialreviewssomeofthemostfrequentlyusedblocksinGNURadio.Usingtheseblocksskillfullywouldbeveryimportantanduseful.Certainlyweonlydiscussafractionoftheavailableblocks.Therearesomeothermoreadvanced,lessusedblocksthatwedidn’ttalkabout.AlsomoreandmoreblocksarecomingasGNURadiogetsmorepopular(maybeincludingyourblockoneday).Ourintroductionisverysimple.Ifyouwishtoknowallthedetailsaboutablock,pleasegotoitsdocumentationpageandthenreaditssourcecodedirectly.Thesourcecodeisthebestplacetounderstandwhat’sgoingoninablockthoroughly.Studyingsomeoftheexamplestoseehowablockausedisalsoaverygoodway.

References

[1]GNURadio2.xDocumentationhttp://www.gnu.org/software/gnuradio/doc/index.html

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文