ınbul 2004
'S for the
olution. 20
Local Time (hh:mm)
02:23
Rx Clock
23:36 00:10 00:43 01:16 01:50
1 - T T i
etween the
iparing the 18 j
| in Figure ; |a /
he height |
educed in
suggesting
to the two
; shown in
r one-third
, Which is
Receiver Clock Offset (m;
>
==
— Forward |
„= Backwerd_ |
id
23000
GPS Time (s)
BALL —— =
13000 15000 17000 19000 21000
Figure 9: Receiver Clock Error
Time (hh:mm)
223
Position Error (Backward Pass) Local Time (hh:mm)
12 23:36 00:10 90:43 01:16 91:50 02:23
[=="Talitude (RMSE = 0.084 m)
| Longitude (RMSE = 0.076 m) |
1, [-— Height (RMSE = 0,531 m) i
08 |
| i
| ;
1 06+ 3 A A T i
€ | ^ M #
5 04 , : f^. ^
i LET f MW Vi
© te PA AVIA
Ww s AV WA ay twee"
ih de ve NY
SPS Time (s)
a 13000 15000 17000 19000 21000 23000
GPS Time (s)
. oye 3
m bem) Figure 10: Position Errors between P" and the Data
ime mm,
223 Supplier’s In-House DGPS Solution
+ Position Error (Backward Pass) Local Time (hh:mm)
| 23:36 00:10 00:43 01:16 0t50 0223
| 0.6 = 5i a :
| 05
|
| =
| |
| o3r a
| 2:02 ly
| eg 02 1 |
| = { ‘ 4
EL ANY
1
A y U^
o -01| Ï l
3PS Time (s)
-02 [— Latitude (RMSE = 0.04 m)
| Longitude (RMSE = 0.056 m)
— Height (RMSE = 0 166 m)
gt 13000 15000 17000 19000 21000 23000
; GPS Time (s)
ime (hh:mm) 3
m Figure 11: Position Errors between P" and JPL?s GIPSY
| OASIS II Solution
|
| oye Aij» 3 1
| Table 2: Position Error Statistics (cm): P^ vs DGPS
| .
| Latitude Longitude Height
| Mean -14 -3.6 50.7
| Std. Dev. 4.0 6.7 15.9
| RMSE 8.4 7.6 53.1
| ... ce 3 6
Table 3: Position Error Statistics (em): P” vs JPL's
= G IS II
i La de H t
000
PS Time (s) Mean "ll, -3. ] ] 37
9.4
16.6
4.6
5.6
Std. Dev.
RMSE
3.8
0
849
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
5. CONCLUSIONS
The purpose of this paper was to present a method that has been
developed at the University of Calgary for high precision
kinematic positioning using a single dual-frequency GPS
receiver. P), the software developed at the University of
Calgary, was also described. Since the PPP approach does not
require the deployment of base stations, errors associated with
reference station coordinates as well as error de-correlation
with increasing rover-reference receiver distance do not apply.
Aside from globally consistent accuracy (rivaling DGPS
accuracy in many instances), the PPP approach offers a
significant cost saving since base stations do not need to be
deployed.
Two airborne kinematic data sets have been analyzed and
compared with other available solutions. It was found that the
P^ solution agrees well with the GIPSY/OASIS II solution
provided by JPL, with differences at the centimetre level in the
horizontal, and up to a couple of decimetres in the vertical.
About half of the height discrepancy between these two
solutions can be attributed to different precise clock product
resolutions — 5 minutes for P?, and 1 second for JPL.
The second data set presented in this paper showed that the PPP
approach would be more accurate than the DGPS approach in
some instances especially over long baselines since it does not
depend on error de-correlation between the rover and reference
receivers or the coordinates of the reference receivers. As such,
it is capable of providing a globally consistent solution with
reduced logistical complexity in the field.
ACKNOWLEDGEMENTS
The research was supported partially by an NSERC grant and
an NCE GEOIDE grant.
REFERENCES
Gao, Y. and X. Shen (2002). *A New Method Of Carrier Phase
Based Precise Point Positioning," Navigation: Journal of the
Institute of Navigation, Vol. 49, No. 2.
Gabor, M.J. (2000) “Characteristics Of Satellite-Satellite
Single Difference Widelane Fractional Carrier Phase
Biases.” Proceedings of ION GPS 2000, Salt Lake
City, UT, 19-22, September.
IGS, (2000). The IGS-2000 Annual Report.
GPS*C Interface Control Document, 2001, Natural Resources
Canada, www.cdgps.com.
IGS Data & Products, November 21,
http://i gscb.jpl.nasa.gov/components/prods.html
2003,
