Abstract

1

• Abstract
• Since 2006 the Navigation Support Office at ESOC
  has been working on its new analysis software,
  called Napeos, for all its activities. Although
  the main focus for Napeos was GNSS processing the
  developments kept the SLR and Doris processing
  capabilities intact. The development was
  completed by September 2007 at which point the
  software was extensively tested for its IGS
  operational capabilities. For that purpose the
  whole year 2007 was reprocessed generating a GPS
  only and a GPS/GLONASS combined solution. The
  main focus of this reprocessing was robustness of
  the software to ensure that it could process the
  full year without any malfunctioning. After
  successful completion of the tests Napeos was put
  into operations for the ESOC IGS activities. It
  replaced the old software for the ESOC IGS final
  and IGLOS products starting with GPS week 1463
  (20-January-2008). Next followed the ESOC rapid
  products on day 2 of GPS week 1469
  (4-March-2008). The Ultra-rapid products were
  replaced starting with the solution of 18 hours,
  on day 2 of GPS week 1471 (18-March-2008). This
  poster highlights the significant improvements of
  the quality of the ESOC IGS contributions. The
  key features of the Napeos IGS analysis are
• Full IERS 2003 compliance
• Undifferenced GNSS processing. GPS, GLONASS, and
  Galileo
• Non-overlapping processing using 24-hour
  solutions
• No day boundary continuity constraints
• Fast!!! (Final orbit solution takes 30 minutes
  using 100 stations on a Linux PC with Intel
  Fortran compiler)
• Currently for improving our IGS products we are
  focusing on the following themes
• Earth Albedo and infra-red radiation modeling
• Orbit predictions to improve the Ultra-rapid
  orbits
• Normal equation stacking for longer (GLONASS)
  arcs.
• Effects of network geometry and number of
  stations
• ESOC is heavily involved in LEO precise orbit
  determination (POD), especially ENVISAT and
  METOP-A. ENVISAT POD is based on SLR and Doris
  tracking whereas METOP-A is based on GPS
  tracking. With the IGS results and the
  comparisons of the LEO orbits we have
  demonstrated that Napeos is fully state of the
  art in these areas. This implies that we could
  use Napeos also to become an analysis centre for
  the IDS and ILRS. We are currently studying the
  efforts required regarding joining these
  technique services. In particular we are
  interested in contributing a TRF solution from
  each of these three techniques for the upcoming
  ITRF realization.

• ESOC Innovation Corner
• At ESOC we have the ambition to be the best in
  class within the IGS. For this purpose we have
  weekly meetings of our IGS group and within these
  meetings we have what we call an innovation
  corner. In this innovation corner group member
  can propose special study topics. If the topic is
  accepted one of the group members can volunteer
  to do the study. Results from two innovation
  corner topics, which were studied earlier this
  year, are shown below.
• Station Selection
• In this study the effect of the number of
  stations and the station selection feature of
  Napeos were investigated. For this purpose we
  used our normal IGS final routine settings. The
  only change was that we did not select any
  preferred stations. The routine was run starting
  with 30 stations, increasing the number by 10 for
  each next step, and relying on the Napeos station
  selection feature to select a network with a good
  spatial distribution. The quality of the
  solutions was assessed to determine which number
  of stations is optimal. The figure below shows
  the result for the orbits. There are two
  interesting features
• 60 to 70 well distributed stations seem to give
  sufficient quality, additional stations do not
  bring any significant improvements.
• The ambiguity free results do not seem to
  converge to the same level as the ambiguity
  fixed results. This is somewhat surprising!?

Non-GNSS Activities Besides our IGS GNSS
activities our group at ESOC is also heavily
involved in LEO POD using DORIS, SLR, PRARE, and
Altimetry observations, in particular for the ESA
missions ERS-1 and 2 and Envisat. We do, however,
have concrete plans to become a reprocessing
Analysis Center for both the IDS and the ILRS.
For the IDS a first 1-year solution is currently
being generated. Below some interesting results
from our DORIS activities. DORIS Phase Center
Variations In GNSS we have had significant issues
with the location of the receiver and transmitter
phase center position. DORIS, also being a
microwave technique, suffers from the same
problems. The lessons learned in GNSS in this
area are very useful for the DORIS technique. The
top two figures below show the residuals of the
DORIS observations as function of azimuth and
elevation for two DORIS stations. One station,
SCRB, located around the equator, the other one,
FAIB, near the North-Pole. The bottom two figures
show the residuals for the Envisat transmitting
antenna, once in colour and once in black and
white. In all cases some clear signals may be
observed which are most likely caused by the
phase center variations of the DORIS antennas.

• True GNSS Processing
• The ESOC IGS final analysis solution is generated
  in two variants. The first variant is a
  classical GPS only solution. The second variant
  is a true GNSS solution in which the GPS and
  GLONASS observations are processed simultaneously
  and contribute equally to all estimated
  parameters. We have ensured that both solutions
  use the same station selection by sharing the
  same preprocessing step. So the only difference
  between the two solutions is the inclusion of the
  GLONASS data. This allows us to study the impact
  the GLONASS observations have on our solutions.
• For testing our new software, before putting it
  into operations for the IGS, we did a full
  reprocessing of the year 2007 generating both a
  GPS only and a GPS-GLONASS combined solution. The
  figure below shows the RMS and Median of
  comparing the GPS parts of the orbits against the
  IGS final orbits. For comparison also the CODE
  orbits and the original ESOC solutions are
  included in this picture.
• We may make the following observations
• The new ESOC software solutions (labeled GPS and
  GPS/GLO) outperform the old ESOC solutions.
• The GPS and GPS/GLO solutions perform equally
  good.
• No negative influence of GLONASS!
• The inclusion of GLONASS data has a noticeable
  impact on the solutions
• It is our ambition to stop submitting our
  GPS-only solutions and start submitting our true
  GNSS solutions to the IGS in the near future. The
  required software modifications have been made
  and are now in testing. Two further enhancements
  are envisioned
• Improved pre-processing of the GLONASS
  observations
• GLONASS ambiguity fixing

Orbit Prediction Tests Over
Improvement of ESOC IGS Contributions The two
figures below show the improvement of the ESOC
IGS contributions. The top figure shows it for
the ESOC Rapid orbits. The ESOC rapid products
are based on our new software Napeos starting
with the products of day 2 of GPS week 1469. The
bottom figure shows the improvement for our ESOC
Final clocks. The ESOC final products are based
on our new software Napeos starting with GPS week
1463. In both cases a significant improvement of
the results may be observed.

• Orbit Prediction Tests
• The second innovation study is investigating the
  quality of our orbit predictions. Despite the
  fact that the estimated part of our Ultra-rapid
  orbits is amongst the best within the IGS, the
  predicted part does not perform as good. Main
  goal of this investigation is to improve the ESOC
  orbit predictions to become best in class. For
  this study we use
• ESOC final orbits as input, IGS final orbits to
  compare the predictions
• Different combination of all 18 orbit parameters,
  and different arc lengths
• 9 CODE parameters (D0 Y0 B0 Dp Yp Bp)
• 9 CPR parameters (R0 A0 C0 Rp Ap Cp)
• The figure below gives a summary of the obtained
  results from which we may derive
• For 1-day fit a few parameters may be saved but
  the predictions do not get much better
• The 3-day and 4-day predictions can be improved
  using more parameters but still do not reach the
  same level as the 2-day predictions.
• We will continue these orbit predictions studies
• Focusing on the Albedo and IR effects
• Look at ERP issues using the real Ultra Rapid
  results
• Using 1.5 and 2.5-day arc lengths for the fitting
  interval.

• GIOVE-A Processing
• At ESOC we have analyzed the GIOVE-A microwave
  data (150 days in 2006/2007)
• IGS final orbits and clocks fixed.
• Estimated parameters
• GIOVE-A orbit and clock.
• GESS station coordinates and clocks.
• Receiver dependent bias between GPS and GIOVE-A.
• Daily solutions saving the normal equations.
• Generate n-day solutions based on normal equation
  stacking (n 1, 2, 3, 4, and 5).
• The figure below shows the orbit quality based on
  a 2-day fit through the middle day of the
  solutions. For comparison a GPS satellite
  (SVN-35) was treated in the same way as the
  GIOVE-A satellite. The performance of the GIOVE-A
  orbit estimates is not as good as that of the GPS
  satellite.

• Conclusions
• Improvement of ESOC IGS Contributions
• ESOC has joined the best of class for the IGS
  rapid, final, and reprocessed products!
• True GNSS Processing
• ESOC is running a GPS-only and a true GNSS
  solution. Soon, hopefully before the end of 2008,
  only the true GNSS solution will be submitted for
  the IGS final products.
• GLONASS data can be included in the IGS
  processing without any negative impact on the GPS
  results. Butno improvement observed either.
• ESOC is actively involved in processing the
  GIOVE-A and GIOVE-B data. The new observables
  offered by these satellites are really
  interesting.
• ESOC Innovation Corner
• ESOC has an active innovation corner which
  should ensure that we remain amongst the best of
  class for all IGS products, by pro-actively
  investigating possible improvements and
  enhancements.
• A significant list of ideas exist, several being
  pursued at present, e.g.
• Intersystem bias stacking, improved data cleaning
  and cycle slip detection, N-day solutions (e.g.
  weekly orbits to have orbits and SINEX fully
  consistent), M-hourly solutions (for real time
  applications), Ambiguity fixing improvements, and
  advanced clock modelling.
• Non-GNSS Activities
• ESOC is active in routine processing of DORIS and
  SLR observations from ESA satellite missions.
  Also involved in reprocessing of all historic
  observations of these missions.
• ESOC is planning to become an reprocessing
  Analysis Centre for all space geodetic
  techniques. Starting with the IDS and ILRS in
  2008. IVS possibly in 2009.

ESOC has the ambition to remain amongst the best
in class in the IGS and in addition contribute
significantly to the IDS, ILRS, and IVS!