comparison of gps/glonass clock solutions · comparison of gps/glonass clock solutions 1. ... tps...

Comparison of GPS/GLONASS Clock Solutions

R. Dach1, S. Schaer2, and M. Meindl1

1Astronomical Institute, University of BernSidlerstrasse 5, CH-3012 Bern

2Federal Office of Topography swisstopo

IGS–Workshop on GNSS–Biases

Bern, Switzerland, 18.-19. January 2012

AIUB

Outline

Dach, Schaer and Meindl: Comparison of GPS/GLONASS Clock Solution – 2 / 34

Comparison of GPS/GLONASS Clock Solutions

1. Direct comparison of the satellite clocks

2. Comparison of the submitted ISB/IFB

3. Precise Point Positioning solutions

AIUB

Overview on the Contributions


Clock SolutionAnalysis Center Systems satell. stations

COD Center for OrbitDetermination in Europe,AIUB, Switzerland

GPS+GLONASS 30 sec 300 sec

EMR Natural Resources Canada,Canada

GPS+GLONASS 30 sec —

ESA European Space OperationsCenter, ESA, Germany


GFZ GeoForschungsZentrum,Germany


GRG GRGS-CNES/CLS,Toulouse, France


IAC Information-AnalyticalCentre, Russia


AIUB

Satellite Clock Series (Reference clock: G15)


Deviation from the median

−1

0

1

G02

in n

s

COD EMR ESA GFZ GRG IAC

−1

0

1

G12

in n

s


−1

0

1

G22

in n

s

0 4 8 12 16 20 24

DOY 2011−348 (hours)


AIUB




−300

0

300

R02

in n

s


−300

0

300

R12

in n

s


−300

0

300

R22

in n

s

0 4 8 12 16 20 24



AIUB

Biases in GPS/GLONASS (Clock) Processing


� DCB: differential code biasdifferent hardware delays forP– and C–Code

� ISB: inter–system biasdifferent hardware delays formeasurements of differentGNSS

� IFB: inter–frequency biasfrequency–dependent hard-ware delays for the differentGLONASS–signals

We can only extract the sumof delays from a GPS/GLONASSdata processing.

AIUB

Biases in GPS/GLONASS (Clock) Processing


The different biases are realized in different ways in the six contributingsolutions. We can extract the following biases from the (satellite) clock timeseries:

� ISB: inter–system biasone common offset between all GPS and GLONASS clocks from each ofthe solutions

� IFB: inter–frequency biasone offset for each GLONASS clock with respect to all GPS clocks of thesolution

Note, all offsets are realized as median to be robust against outliers.

AIUB



Deviation from the median, ISB subtracted

−30

0

30

R02

in n

s


−30

0

30

R12

in n

s


−30

0

30

R22

in n

s

0 4 8 12 16 20 24



AIUB



Deviation from the median, ISB+IFB subtracted

−1

0

1

R02

in n

s


−1

0

1

R12

in n

s


−1

0

1

R22

in n

s

0 4 8 12 16 20 24



AIUB

ISB/IFB as Extracted from Satellite Clocks


−100

−50

0

50

IFB

in n

s

Analysis center

362.6 ns −17.7 ns −12.8 ns 149.3 ns 11.2 ns−491171344.7 ns


Conclusion

� The ISB/IFB can be estimated from the satellite clock solutions ofdifferent ACs.

� Appling these corrections allows for a GLONASS satellite clockcombination following the same technology as for GPS.

� The estimated ISB/IFB can be rather inhomogeneous if a satellite is notincluded in all solutions (which has no influence on a PPP–solution).

AIUB




−100−50

050

100

R02

in n

s

COD ESA GFZ GRG

−100−50

050

100

R12

in n

s

COD ESA GFZ GRG

−100−50

050

100

R22

in n

s

0 4 8 12 16 20 24


COD ESA GFZ GRG

AIUB



Deviation from the median, ISB subtracted

−30

0

30

R02

in n

s

COD ESA GFZ GRG

−30

0

30

R12

in n

s

COD ESA GFZ GRG

−30

0

30

R22

in n

s

0 4 8 12 16 20 24


COD ESA GFZ GRG

AIUB



Deviation from the median, ISB+IFB subtracted

−1

0

1

R02

in n

s

COD ESA GFZ GRG

−1

0

1

R12

in n

s

COD ESA GFZ GRG

−1

0

1

R22

in n

s

0 4 8 12 16 20 24


COD ESA GFZ GRG

AIUB



−100

−50

0

50

IFB

in n

s

Analysis center

362.6 ns −17.7 ns −12.8 ns 149.3 ns 11.2 ns−491171344.7 ns


−100

−50

0

50

IFB

in n

s

Analysis center

−13.5 ns 2.9 ns −2.8 ns 29.4 ns

COD ESA GFZ GRG

AIUB



−100

−50

0

50

IFB

in n

s

Analysis center

362.6 ns −17.7 ns −12.8 ns 149.3 ns 11.2 ns−491171344.7 ns


Conclusion – 2� To apply the mean ISB/IFB per satellite from the satellite clock solutions helps to

reduce the differences between the solutions.

� Nevertheless, additional ISB/IFB need to be estimated from the satellite clocksolutions of different ACs to achieve the consistency level necessary for thecombination.

� The problem of ISB/IFB estimation from satellites not included in all solutions isreduced but not solved.

AIUB

IFB/ISB Computed by the Analysis Centers


Repeatability of the IFB/ISB (as they are)

−10

0

10

R02

in n

s 49.2 ns −321.1 ns −158.4 ns −327.2 ns

COD ESA GFZ GRG

−10

0

10

R12

in n

s 49.2 ns −331.0 ns −159.6 ns −321.3 ns

COD ESA GFZ GRG

−10

0

10

R22

in n

s

Analysis center

48.9 ns −323.6 ns −158.3 ns −343.2 ns

COD ESA GFZ GRG

Station: MATE – Matera, IT

AIUB



Repeatability of the IFB/ISB (as they are)

−10

0

10

R02

in n

s 35.7 ns −335.6 ns −173.1 ns −331.9 ns

COD ESA GFZ GRG

−10

0

10

R12

in n

s 41.6 ns −340.1 ns −168.0 ns −329.3 ns

COD ESA GFZ GRG

−10

0

10

R22

in n

s

Analysis center

36.8 ns −335.1 ns −171.0 ns −353.4 ns

COD ESA GFZ GRG

Station: CONZ – Concepcion, CL

AIUB



Repeatability of the IFB/ISB (unified reference)

−10

0

10

R02

in n

s 49.6 ns −320.9 ns −158.8 ns −320.1 ns

COD ESA GFZ GRG

−10

0

10

R12

in n

s 49.4 ns −334.5 ns −159.7 ns −321.7 ns

COD ESA GFZ GRG

−10

0

10

R22

in n

s

Analysis center

48.7 ns −321.6 ns −158.4 ns −341.3 ns

COD ESA GFZ GRG


AIUB



Repeatability of the IFB/ISB (unified reference)

−10

0

10

R02

in n

s 35.8 ns −335.1 ns −173.0 ns −332.6 ns

COD ESA GFZ GRG

−10

0

10

R12

in n

s 41.6 ns −342.7 ns −168.1 ns −327.9 ns

COD ESA GFZ GRG

−10

0

10

R22

in n

s

Analysis center

36.8 ns −333.5 ns −171.0 ns −352.0 ns

COD ESA GFZ GRG

Station: CONZ – Concepcion, CL

AIUB



Conclusion� The IFB/ISB series from daily independent solutions need to be unitfied regarding

their reference.

� The stability of the IFB with unified reference depend on the stability of thedifferences between the IFB between two satellites from the day to day in theoriginal series.

� In the best cases we achieve a peak-to-peak stability of 2 to 3 ns.

AIUB



ISB characteristic of the receivers

−100

0

100

200

300

400

500

ISB

in n

s

Stations

ASHTECH Z18 JAVAD TRE_G3T DELTA JAVAD TRE_G3TH DELTA JPS EGGDTJPS E_GGD JPS LEGACY LEICA GRX1200+GNSS LEICA GRX1200GGPRONOV OEMV3 TPS E_GGD TPS GB−1000 TPS LEGACYTPS NET−G3A TPS NETG3 TPS ODYSSEY_E TRIMBLE NETR5TRIMBLE NETR8 TRIMBLE NETR9

Analyse centrum: COD

AIUB




−400

−300

−200

−100

0

100

200

300

ISB

in n

s

Stations


Analyse centrum: GFZ

AIUB




−600

−500

−400

−300

−200

−100

0

100

ISB

in n

s

Stations


Analyse centrum: ESA

AIUB




−400

−300

−200

−100

0

100

200

300

ISB

in n

s

Stations


Analyse centrum: GRG

AIUB



Differences between ISB characteristic of the receivers

−230

−220

−210

−200

−190

−180

ISB

in n

s

Stations


Analyse centrum: COD−GFZ

AIUB




−390

−380

−370

−360

−350

ISB

in n

s

Stations


Analyse centrum: COD−ESA

AIUB




−180

−170

−160

−150

−140

ISB

in n

s

Stations


Analyse centrum: GFZ−ESA

AIUB




−440

−420

−400

−380

−360

−340

−320

−300

−280

−260

−240

ISB

in n

s

Stations


Analyse centrum: COD−GRG

AIUB




−220

−200

−180

−160

−140

−120

−100

−80

−60

−40

−20

ISB

in n

s

Stations


Analyse centrum: GFZ−GRG

AIUB




−60

−40

−20

0

20

40

60

80

100

120

140

ISB

in n

s

Stations


Analyse centrum: ESA−GRG

AIUB




Num. of Mean Median RMSDifference Stations in ns in ns in ns

COD − GFZ 52 −210.6 −209.4 4.9COD − ESA 39 −377.5 −377.6 5.1GFZ − ESA 36 −167.7 −168.2 6.1COD − GRG 50 −371.9 −372.2 18.7GFZ − GRG 46 −162.1 −163.0 19.2ESA − GRG 34 6.1 5.8 20.6

� High consistency (low RMS) with a proper IFB–handling(enough weight for the code measurements?)

� Test whether the ACs select the same type of code observations(CODE differs from ESA and GFZ)

AIUB

Number of Satellites in the Solutions


Number of satellites per system

0

8

16

24

32

Num

ber

of s

at. GPS


0

8

16

24

32

Num

ber

of s

at.

Analysis center

GLONASS


AIUB

RMS of Orbit–fit using Bernese Software


Mean over all satellites of each system

0

5

10

Mea

n in

mm GPS


0

5

10

Mea

n in

mm

Analysis center

GLONASS


AIUB

RMS of Orbit–fit using Bernese Software


Median from all satellites of each system

0

5

10

Med

ian

in m

m GPS


0

5

10

Med

ian

in m

m

Analysis center

GLONASS


AIUB

Comparison in the PPP–Performance


How the following Solutions Have Been Generated?1. Data import and screening

(a) if high–rate satellite clock corrections are available:Phase–data are screened based on PPP with a sophisticated algorithmbefore the first PPP–solution using post–fit residual screening.

(b) without high–rate satellite clock corrections:Only the consistency between code– and phase–data is checked before thefirst PPP–solution using post–fit residual screening.

2. static PPP using GPS/GLONASS dataassuming satellite–specific inter–system/inter–frequency code biases without

ambiguity resolution for the phase data

3. pseudo–kinematic PPP using GPS/GLONASS data(same as above)

AIUB

Statistics of the PPP Solution (after screening)


Number of code measurements (300 sec. sampling)

0

1000

2000

3000

Cod

e ob

serv

. GPS


0

1000

2000

3000

Cod

e ob

serv

.

Analysis center

GLONASS


Station: ONSA – Onsala, SE

AIUB



Number of phase measurements (300 sec. sampling)

0

1000

2000

3000

Pha

se o

bser

v. GPS


0

1000

2000

3000

Pha

se o

bser

v.

Analysis center

GLONASS



AIUB



Number of ambiguity parameters

0

100

200

Am

bigu

ities

GPS


0

100

200

Am

bigu

ities

Analysis center

GLONASS



AIUB



A posteriori unit of weight

0

1

2

RM

S in

mm

GPS/GLONASS with common receiver clock


0

1

2

RM

S in

mm

Analysis center

GPS/GLONASS with independent receiver clock



AIUB

Coordinate Comparison from PPP Solutions


Coordinate difference to IGS08.SNX

−15−10

−505

1015

Nor

th in

mm

GPS


−15−10

−505

1015

Nor

th in

mm

Analysis center

GLONASS



IGS08.SNX

AIUB




−15−10

−505

1015

Nor

th in

mm



−15−10

−505

1015

Nor

th in

mm

Analysis center




IGS08.SNX

AIUB




−15−10

−505

1015

Eas

t in

mm

GPS


−15−10

−505

1015

Eas

t in

mm

Analysis center

GLONASS



IGS08.SNX

AIUB




−15−10

−505

1015

Eas

t in

mm



−15−10

−505

1015

Eas

t in

mm

Analysis center




IGS08.SNX

AIUB




−30−20−10

0102030

Up

in m

m

GPS


−30−20−10

0102030

Up

in m

m

Analysis center

GLONASS



IGS08.SNX

AIUB




−30−20−10

0102030

Up

in m

m



−30−20−10

0102030

Up

in m

m

Analysis center




IGS08.SNX

AIUB

Pseudo–Kinematic PPP Solutions


RMS of the differences w.r.t. the static solutions

0

10

20

Nor

th in

mm

GPS


0

10

20

Nor

th in

mm

Analysis center

GLONASS



AIUB




0

10

20

Nor

th in

mm



0

10

20

Nor

th in

mm

Analysis center




AIUB




0

10

20

Eas

t in

mm

GPS


0

10

20

Eas

t in

mm

Analysis center

GLONASS



AIUB




0

10

20

Eas

t in

mm



0

10

20

Eas

t in

mm

Analysis center




AIUB




0

20

40

Up

in m

m

GPS


0

20

40

Up

in m

m

Analysis center

GLONASS



AIUB




0

20

40

Up

in m

m



0

20

40

Up

in m

m

Analysis center




AIUB

GLONASS: Interfrequency–Biases


IFB as they are estimated in the static PPP–solution

−10

−5

0

5

10

15

IFB

in n

s

GLONASS−only


−10

−5

0

5

10

15

IFB

in n

s

Analysis center




comparison of gps/glonass clock solutions · comparison of gps/glonass clock solutions 1. ... tps...

Documents