Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.ipa.nw.ru/conference/wpltn2012/docs/24/1540%20qzs-1%20orbit.pdf Äàòà èçìåíåíèÿ: Mon Oct 1 12:02:24 2012 Äàòà èíäåêñèðîâàíèÿ: Sun Feb 3 20:34:13 2013 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: çàêîí âèíà

Accuracy evaluation of QZS-1 orbit solutions with Satellite Laser Ranging
Kyohei Akiyama (Japan Aerospace Exploration Agency: JAXA) and Flight Dynamics Team And Toshimichi Otsubo (Hitotsubashi University)
Sep. 24, 2012 WPLTN Technical Workshop@St.Petersburg, Russia
1


Introduction
QZSS-1(Quasi-Zenith Satellite-1) launched in Sep.2010
· A Japanese original positioning system using multiple satellites that have the

same orbital period as geostationary satellites with about 45deg inclinations. · Transmit GPS compatible signal and LEX comm. signal based on MultiGNSS scheme. QZS-1 orbit parameters Semi-major Axis 42,164 km (average) Eccentricity 0.075 ± 0.015 Orbital Inclination Argument of Perigee Central Longitude of Ground Track 43° ± 4° 270° ± 2° 135° 5° East ±

Ground trace from QZS-1 orbit

2


QZSS monitoring stations
A QZSS tracking network of 9 monitoring stations are currently operated.

3


Precise Point Positioning using LEX signal channel of QZS and application
LEX: L-band experiment signal

Development Status of MADOCA, T. Takasu, 2012
4


MADOCA
· · ·

Multi-GNSS Advanced Demonstration tool for Orbit and Clock Analysis
Software for Precise Orbit/Clock estimation for Multiple GNSS For JAXA Precise Point Positioning (PPP) experiment via QZSS LEX channel Key-technology for precise positioning with GNSS

· Requirements
· ·

Satellites: GPS, GLONASS, QZSS and Galileo Offline (in this study) and real-time functions

· Goal of Orbit/Clock Accuracy
· ·

Offline : 3 cm/0.1 ns (GPS), 7 cm/0.25 ns (Glonass/QZS) Real-time: 4 cm/0.1 ns (GPS), 9 cm/0.25 ns (Glonass/QZS)
Development Status of MADOCA, T. Takasu, 2012

5


Concept of real time PPP user using MADOCA

Development Status of MADOCA, T. Takasu, 2012
6


Accuracy evaluation of QZS-1 MADOCA solutions with SLR
This Study
· · ·

SLR residuals to the QZS-1 orbits processed with MADOCA. Differences between the MADOCA-orbits and those with SLR tracking data. The orbits with SLR tracking data are provided by JAXA and Hitotsubashi Univ. (HIT).
J AXA orbit determination with SLR MADOCA orbit determination Accuracy Evaluation
Quality Check SLR stations

QZSS Monitoring Stations

HIT. orbit determination with SLR

IGS AC. final orbit
Quality Check
7


Evaluation Procedure
Step.1 Prior evaluation: GPS orbits using SLR bservations
· GPS orbit determination using SLR observations. (JAXA/HIT) · Differences between JAXA/HIT-orbits and IGS final orbits. · Differences between JAXA-orbits and HIT-orbits.

Step.2 Evaluation of the QZS-orbits using SLR observations
· QZS-1 orbit determination using SLR observations. (JAXA/HIT) · Differences between JAXA-orbits and HIT-orbits.

Step.3

Evaluation of the orbits processed with MADOCA

· SLR residuals to the orbits processed with MADOCA. · Differences between the MADOCA-orbits and the JAXA/HIT-orbits using SLR observations.

8


STEP1:

GPS orbit estimation with SLR
Evaluation
· Differences between JAXA/HIT-orbits and IGS final orbits. · Differences between JAXA-orbits and HIT-orbits.

GPS orbit determination using SLR observations
· Estimation periods are selected so that the SLR data at no less than 3 stations

is provided at the same time.
Models/Parameters Site position Satellite mass Difference between CoM and optical center Troposphere delay model SRP model JAXA ilrsb 972.9 kg [862.58, -524.51, 671.7] m (common) Marini-Murray model CO D E m o d e l Mendes & Pavlis model Canon ball HIT ITRF2008 930.0 kg

Estimation parameters

Orbit elements (6)

Orbit elements (6) SRP correction coefficient (1) Constant and 1/rev accelerations in the along-track direction (3)

9


GPS36: JAXA-IGS final orbit
5 Days Arc Solution Example

Short Arc Solution Example

10


Step1: Results

GPS orbit estimation with SLR
Differences between JAXA/HIT-SLR orbits and IGS final orbits · HIT vs. IGS final · JAXA vs. IGS final
· ·

Radial : ~ 5 cm Along track : ~20 cm
248 225

· Radial : ~ 10 cm (exc. arc1)

140 120 100 RMS Error [cm] 80 60 40 20 0

Differences between JAXA/HIT orbit and IGS final orbit
HIT (LONG) JAXA (LONG) JAXA (SHORT)

Arc Length · Long (7~9 days) · Short (4~7 hours)

Radial Cross Along Radial Cross Along Radial Cross Along Radial Cross Along track track ARC1 track track ARC2& track track ARC4 track track ARC5

11


Step2:

QZS-1 orbit estimation with SLR
Evaluation · SLR residuals of the orbits determined by JAXA. · Differences between JAXA-orbits and HIT-orbits.
· Estimation periods are selected so that the SLR data at no less than

3 stations is provided at the same time.
Models / Parameters Site position Satellite mass Center of mass Optical reflection center Troposphere delay model SRP model Cross-section area SPR Coefficient (Cr) JAXA HIT ilrsb ITRF2008 2280.7 [kg] (common) (Xs, Ys, Zs) = (-0.8, 2.9, 1819.3 ) [mm] (common) (Xs, Ys, Zs) = (-1150.0, -550.0, 4517.64 ) [mm] (common) Marini-Murray model Mendes & Pavlis model Canon ball Canon ball 60.0 [m2] 52.0 [m2] 1.2 C a se - 1 Orbit elements (6) SRP correction coefficient (1) Constant accelerations in the alongtrack direction(1) C a se - 2 1/rev accelerations in the along-track are estimated in addition to Case-1.

Estimation parameters

Orbit elements (6) SRP correction coefficient (1) Constant and 1/rev accelerations in the along-track direction (3)

12


Step2 Results:

QZS-1 orbit estimation with SLR
SLR residuals of the orbits determined by JAXA
·

Long arc (5~7 days) · SLR residuals: ~10 cm RMS · Periodic variation assumed to be due to the model error ware detected.

·

Short arc (5~10 hours) · SLR residuals: ~1 cm RMS

13


Step3:

QZS-1 orbit based on MADOCA
SLR residuals to the orbits processed with MADOCA · U-shaped residuals that have a peak-to peak amplitude of 20 cm were detected in each arc. · The average of SLR residuals in Arc1 were opposite in sign to those in Arc2 and 3.
Evaluation period Arc1 : 2011/09/02 00:00 ~ 09/07 00:00 Arc2 : 2011/11/19 00:00 ~ 11/24 00:00 Arc3, 4 : 2011/12/15 00:00 ~ 12/22 00:00 SLR residuals (cm) AVE STD RMS

14.2 -15.3 -24.1

7.5 9.4 11.0

16.0 17.9 26.5

14


Step3 results:QZS-1 orbit based on MADOCA and SLR orbits
Differences between the MADOCA-orbits and JAXA/HIT-orbits · Radial biases that have an average of 20-30 cm from the orbits using SLR data were detected in each arc. · It seems that the orbits processing with MADOCA obtain the accuracy of 20-30 cm in the radial direction.
Differences between JAXA/HIT and MADOCA
5 4 RMS Error [m] 3 2 1 0 Radial Cross track ARC1 Along track Radial Cross track ARC2 Along track Radial Cross track ARC3 Along track 7.0
HIT(LONG) JAXA(LONG) JAXA(LONG)-1PRX JAXA (SHORT)

Arc Length · LONG (5~7 days) · SHORT (5~10 hours) -1PRX · 1/rev accelerations in along-track is additionally estimated.

15


Conclusion
·

JAXA and HIT evaluate QZS-1 MADOCA using SLR data QZS-1 MADOCA periodic systematic error found by Residual Analysis
·

·

A U-curved P-P 20cm and bias mean difference by each arc in residual analysis were found. Periodic residual (comes from MADOCA analysis) should be studied.

QZS-1 SLR only Orbit Determination
JAXA and HIT orbit matches by about 20cm level in the radial direction. In which cross and Along direction orbit determination were not sensitive well. · More SLR data of QZS-1 is needed to evaluate QZS-1 MADOCA orbit. · Even in a campaign basis SLR to get 3 or more stations participate at the same time for short arc solution to be important.
·

16


Thank you for your attention.

17


GPS orbit estimation with SLR
Differences between JAXA-orbits and HIT-orbits
·

JA X A v s. H I T
· ·

Radial track : ~ 5 cm Along track : ~20 cm

Estimation period (Long) Arc1 : 2011/3/19 00:00 ~ 3/26 00:00 Arc2, 3 : 2011/6/20 00:00 ~ 6/29 00:00 Arc4 : 2011/7/08 00:00 ~ 7/15 00:00 Arc5 : 2011/9/26 00:00 ~ 10/03 00:00

Differences (cm) RMS R ad i al Cross Along

27.4 10.8 11.6 4.8

272.4 106.7 76.7 30.9

238.5 64.8 75.1 27.8

Defferences between JAXA and HIT orbits
35 Radial direction Error [cm] 25 15 5 -5 AVE STDRMSMAXAVE STDRMSMAXAVE STDRMSMAXAVE STDRMSMAX ARC1 ARC2& ARC4 ARC5
JAXA (LONG)

Estimation period (Short) Arc1 : 2011/03/23 00:00 ~ 07:00 Arc2 : 2011/06/21 20:00 ~ 24:00 Arc3 : 2011/06/27 20:00 ~ 24:00 Arc4 : 2011/07/11 19:00 ~ 23:00 Arc5 : 2011/09/29 12:00 ~ 18:00

Differences (cm) RMS Radial Cross Along 25.3 13.0 2.4 5.1 3.6 227.5 89.3 90.3 169.6 34.1 237.7 69.5 10.9 10.7 9.0

18


Step2 results:

QZS-1 orbit estimation with SLR
Differences between JAXA-orbits and HIT-orbits
· Differences
· ·

Radial : ~30cm Along/Cross track: several meters
Differences (cm) RMS Radial Cross Along

· Biases that have an average of 10 cm in the radial direction exist in each arc.
50 Radial error [cm] 40 30 20 10 0

Defferences between JAXA and HIT orbits
105

Estimation period (Long) Arc1 : 2011/09/02 00:00 ~ 09/07 00:00 Arc2 : 2011/11/19 00:00 ~11/24 00:00 Arc3, 4 : 2011/12/15 00:00 ~ 12/22 00:00

24.2 26.5 15.1

335.3 198.1 99.9

853.4 126.2 87.5

JAXA (LONG)

Estimation period (Short) Arc1 : 00:0 Arc2 : 20:0 Arc3 : 20:0 Arc4 : 19:0 2011/03/23 0 ~ 07:00 2011/06/21 0 ~ 24:00 2011/06/27 0 ~ 24:00 2011/07/11 0 ~ 23:00

RMS (cm) Radial Cross Along

-10 -20 AVE STD RMS MAX AVE STD RMS MAX AVE STD RMS MAX ARC1 ARC2&3 ARC4

13.8 23.9 4.6 10.2

100.7 200.1 257.5 106.1

156.5 186.9 8.8 136.3

19


MADOCA internal orbit accuracy by 24H-Overlap analysis

Development Status of MADOCA, T. Takasu, 2012
20