Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
1 1
6/24/09
ElectronicsandSignalProcessingLaboratoryEPFL‐STI‐IMT‐ESPLABRueA.‐L.Breguet2CH‐2000Neuchâtel
Head:Prof.P.‐A.FarineTeamLeader:Dr.C.BoHeronAuthors:AleksandarJovanovic,YoussefTawk
1
2
Outline CBOC/TMBOCconcept(A.Jovanovic)
Trackingmetrics&TrackingalgorithmsforMBOC ComparisonofMBOCalgorithms ResultsandrecommendationsforMBOCtracking
AltBOCconcept(Y.Tawk) Trackingmetrics&TrackingalgorithmsforAltBOC ComparisonofAltBOCalgorithms ResultsandrecommendationsforAltBOCtracking
6/24/09 2
3
CorrelaJonandTrackingofMBOC OverviewofMBOCreceivertrackingschemes Performcomparisonw.r.t.trackingmetrics
ACFpeakamplitude Multipatherrorenvelope Codetrackingerror Provideadvantages/disadvantagesofeachalgorithm
Recommendationstoreducereceivercomplexity
6/24/09 3
4
CBOC/TMBOCdefiniJon CBOCvs.TMBOC Samepowerspectraldensity(PSD)
Pilotanddatachannelshavedifferentstructures
DifferenttrackingperformanceforCBOC/TMBOC
CBOC structure TMBOC structure
6/24/09
BOC(1,1)+BOC(6,1)
BOC(6,1) 4chips/33 BOC(1,1)
Pilot TMBOC
Data chanel BOC(1,1)
Data + Pilot channels
Phase Antiphase
4
5
CBOC/TMBOCsignals
6/24/09
• GPSL1COSTMBOC(6,1,4/33) BOC(1,1)ondatachannel(25%),TMBOConpilot(75%)
Multiplexwide‐bandBOC(6,1)withnarrow‐bandBOC(1,1)• GalileoE1OSCBOC(6,1,1/11,’+/‐’)
Equallysplitpower(50%),signalsinantiphase WeightedsumofBOC(1,1)andBOC(6,1)ondata+pilot
5
6
Trackingingeneral Firststeptoconsider:Autocorrelationfunction(ACF) Maintrackingparameters:
Trackingsensitivity MinimumSNRtomaintaintracking
Trackingrobustness Stabilitytrackingregion
Trackingaccuracy Errors:multipath,noise,interference
• Multipathmainsourceoferror
Phasetracking(PLL) Codedelaytracking(DLL)
6/24/09
E P L
E P L
E P L
Search
Track
Lock
6
7
TradiJonalCBOCreceiverarchitecture
SamereplicaforE,P,Lcorrelator
I&D
NCO
cos()
sin()
I&D
Gene L1B
BOC(6,1)
BOC(1,1)
-Q NCO P
Discri Filter
I&D
I&D
E,P,L
E,P,L
E,P,L P
E,
L
6/24/09 7
8
TradiJonaltrackingperformance TraditionalCBOCtrackingrequiresreplicationoffour‐levelreplica Complicatesthereceiverarchitecture
ImprovementintrackingcomparingtoBOC(1,1)isdifferentfordataanddata+pilotchannels=>C/Nocriterionforcomparison
CBOC(6,1,1/11,’‐’)betterthanTMBOCmodulation
6/24/09
TrackingErrorImprovementvsBOC(1,1)inTermsofEquivalentC/N0(dB)withBW2=24MHzandE‐L=1/12Chips
CBOC(6,1,1/11,’+’) TMBOC(6,1,4/33) CBOC(6,1,1/11,’‐’)
1.9 2.4 3.1
8
9
TM61trackingarchitecture
Useone‐bitreplica,usingBOC(1,1)andBOC(6,1)only
I&D
NCO
cos()
sin()
I&D
Gene L1B
BOC(6,1)
BOC(1,1)
NCO
Discri Filter
I&D
I&D
E,L
P
E,P,L
P
E,
L
6/24/09 9
10
TM61forTMBOCtracking TMBOCtrackingusingTM61ismorecomplicated
CorrelationlosseshighwhenonlyBOC(6,1)isused
Thiscanbeavoidedusingtime‐multiplexsubcarriers Usingzero‐padding WhenBOC(6,1)istracked,BOC(1,1)replicacontainszeros
andvice‐versa UsingthismethodTMBOCtrackingissimilartoCBOCtracking
Still0.2dBworstintrackingthanCBOC
6/24/09
Signalvs.Receiver BOC(1,1) BOC(6,1)
TMBOC(6,1,4/33) 0.85 0.15
CBOC(6,1,1/11,’‐’) 0.95 0.3
10
11
DualCorrelatortracking DualCorrelatortechniquecomparestwoparallelcorrelations:
IncomingMBOCwithBOC(1,1)replica IncomingMBOCwithBOC(6,1)replica
Correlationsareweighted,withparameters: βforBOC(1,1)andρforBOC(6,1)
Optimalβ/ρisintherange[1.6,3.2]w.r.t.MEE(Multipatherror) Thismethodusestwiceasmanycorrelators
Andbinaryreplicas Goodmultipathperformance Easilyconfigurableinthesoftware
6/24/09
NCO
Gene L1B
BOC(6,1) E,L
P
E,P,L
BOC(1,1)
11
12
S‐Curveshapingmethod Methodproposedistofindtheidealdiscriminator
UsingidealS‐curve Fittingthecurve,parametersofthecorrelatorarefound
Correlatorspacing Weightofthecorrelator
Greatnumberofcorrelatorsneeded Multipathreducedstrongly
Usesmulticorrelatorstructure Narrowcorrelator Doubledelatcorrelator
6/24/09 12
13
ASPeCTBOC(1,1)algorithm ASPeCThasbeenproposedforBOC(n,n)signals
Unambiguoustechnique‐eliminationofthesidepeaksofACF
Limiteddegradation[0.6‐1dB]comparingtotraditional ASPeCTcanbeappliedtoCBOC(6,1,1/11,’‐’) Pricetopay:
ReducingtheBOC(6,1)inthelocalCBOCreplicabyhalf
Thisimplies: anegligiblereductionofresistancetonoise aslightdegradationofthemultipatherrorenvelope
6/24/09 13
14
Compositetrackingalgorithm Compositetracking
Dataandpilottogether Convexcombinationofdiscriminatoroutputs
Non‐coherent+coherentchannelcombining
Trackingjitterreduced SER(SignErrorRate)
Newtrackingmetrics
6/24/09 14
15
TrackingCBOC/TMBOCwithBOC(1,1) Narrowbandreceivers(BW<<12MHz)donothaveaccesstofullrangeofMBOCcharacteristics
Optimaltrackingperformanceisreachedwiththematchingchannel Reference:[Dr.MassimoCrisci,ESTEC,GIOVEworkshop,Oct2008]6/24/09
Transmission Receiver reference
Delta Losses [dB]
Relative Noise Jitter
RAME error
Baseline BOC(1,1) 0.0 1.00 1.00
CBOC BOC(1,1) +0.3 0.94 0.97
TMBOC BOC(1,1) +1.7 1.31 1.03
CBOC CBOC 0.0 0.76 0.84
TMBOC CBOC +0.3 0.71 0.84
15
16
Trackingalgorithmscomparison
6/24/09
Tracking scheme
Tracking complexity
Tracking robustness
Tracking accuracy
Applicability to mass-market
receiver TM61 + + - +
Dual Correlator +/- - +/- +/-
S-Curve shaping
+ - + +
ASPeCT + +/- - +/-
Composite tracking
- +/- + +/-
16
17
MulJpathperformance Multipatherrorenvelopes: TM61providesbetterresultsthanpureCBOCtracking
IfDPdiscriminatorused
HRCnotgoodsolution forCBOCtracking
S‐curveshaping Providesthebestresults
Useofadvancedcorrelators: Strobe Narrow,DoubleDelta [M.Fantino,L.Presti]
6/24/09 17
18
ConclusionsandoutlookMBOC BOC(1,1)receiversreceivingCBOC/TMBOC
Lossestoohigh Newspecifictrackingalgorithmsproposed:
TM61andDualCorrelatoroutperformBOC(1,1) Futurework:
Compositetracking Tobefurtherinvestigated
CommontrackingCBOC/TMBOCarchitecture Realchallenge
SeparatealgorithmforCBOCandTMBOC? Needforadvancedmultipathmitigationalgorithms Useofnewcorrelatortypes
6/24/09 18
19
Outline CBOC/TMBOCconcept(A.Jovanovic)
Trackingmetrics&TrackingalgorithmsforMBOC ComparisonofMBOCalgorithms ResultsandrecommendationsforMBOCtracking
AltBOCconcept(Y.Tawk) Trackingmetrics&TrackingalgorithmsforAltBOC ComparisonofAltBOCalgorithms ResultsandrecommendationsforAltBOCtracking
6/24/09 19
20
GalileoE5Signal AltBOC(15,10)isthenewmodulationontheGalileoE5Band.
6/24/09 20
21
ConstantEnvelope
6/24/09 21
22
ConstantEnvelope(cont’)
6/24/09 22
23
ConstantEnvelope(cont’)
6/24/09 23
24
AltBOCCharacterisJcs
6/24/09
E5 Signal Spectrum
E5 Spreading Code Lengths
E5 Signal Parameter
24
25
PowerDistribuJonPower 1st Harmonic (Power Distribution) 2nd Harmonic (Power Distribution)
SSC 85.36% +fs/-fs (94.36%) +7fs/-7fs (1.36%)
PSC 14.64% +3fs/-3fs (61.5%) +5fs/-5fs (22.2%)
90 MHz Bandwidth Theoretical
Power Power
Transmitted
90 MHz 85.34% 100%
75 MHz 80.54% 94.37%
51 MHz 77.88% 91.25%
75 MHz 51 MHz
N.B: Power is shared equally between the four components of the E5 signal
6/24/09 25
26
AutocorrelaJonFuncJon
Autocorrelation vs. Bandwidth
6/24/09 26
27
TrackingPossibiliJes Thepresenceof4componentsontheE5signalallowthereceiverdesignerstohavedifferentpossibilitiesintrackingthissignal:
FreesubcarriersLocalgeneratedsignal FullBandCorrelation 8‐PSKlikeprocessing
Severalcombinationofsignalcomponentsarepossible:
E5a‐I,E5a‐Q,E5b‐IorE5a‐Q Dataorpilot E5
6/24/09 27
28
FreeSubcarriers Thelocalsignalisgeneratedfreeofsubcarriers TheE5signalcomponentsaretranslatedtobasebandfromtheircenterfrequencies
E5a‐I,E5a‐Q,E5b‐I,E5b‐Q,E5aandE5bcanbetrackedindividuallyresultinginaBPSK(10)correlation
6/24/09 28
29
FreeSubcarriers(cont’)
Autocorrelation Function for a single E5 component with a 20 MHz Bandwidth
Comparison between an infinite bandwidth and a 20 MHz AltBOC
receiver
6/24/09 29
30
FullbandCorrelaJon Thelocalsignalisgeneratedwithsubcarrierfortherequiredsignalcomponent
Additionalpowerisgainedcomparingtothefreesubcarriermethod
E5ab,E5‐pilotandE5‐datacanbetrackedinadditiontotheothersignalcomponents
6/24/09 30
31
FullBandCorrelaJon
Comparison between autocorrelation function for a single E5 component with a 75 MHz Full Band correlation
and a 20 MHz Free Subcarrier correlation
Full Band Correlation for different signal components with
a 75 MHz bandwidth receiver
6/24/09 31
32
8‐PSKLikeProcessing OnlythecompleteE5signalistracked Correlationisdoneusingalookuptable(LUT)of8values
Look up table for AltBOC phase states
6/24/09 32
33
AltBOCSummary
PowerSharing
CodePhaseJitter
ReceiverComplexity
MultipathMitigation
FreeSubcarrier ‐ ‐ + ‐
FullBand ‐/+ ‐/+ ‐/+ +
8‐PSK + + ‐ +
Performance of the different tracking architectures
6/24/09 33
34
Conclusions GoodtrackingperformanceofnewBOCsignals
Bettermultipathmitigation Improvedcodetrackingperformance Newtrackingstrategiesrequirements
Morecomplexreceiverarchitecture Moreprocessingpower,higherfs,morecorrelators Highcost–needsimplifications
6/24/09 34
35
ThankYoufortheattention Q&A?
Contacts:
6/24/09
Aleksandar Jovanovic PhD student
EPFL-IMT-ESPLAB Rue A.-L. Breguet 2 CH-2000 Neuchâtel
Phone: +41 32 718 34 39 Fax. +41 32 718 34 02
E-mail : [email protected]
Youssef Tawk PhD student
EPFL-IMT-ESPLAB Rue A.-L. Breguet 2 CH-2000 Neuchâtel
Phone: +41 32 718 34 42 Fax: +41 32 718 34 02
Email: [email protected]
35
36
Back–upslides
6/24/09 36
37
Algorithmcomparison• Falsecorrelationpeaksat0.6chips
CorrelationlossesTM61 ACFpeakvariationα=[0.1,1]
6/24/09 37