Short-session Static and Kinematic Processing

1
Short-session Static and Kinematic Processing

• Short-session static GAMIT processing,
  sessions 1-3 hours long
• Kinematic TRACK processing, coordinates
  estimated at each epoch one or more sites may
  be moving
• High frequency multipath and sometimes
  atmospheric errors dominate and fail to average
• Shorter satellite tracks mean less time to
  separate the signatures in the data, leading to
  higher correlations and thus higher uncertainties
  in coordinates and ambiguity parameters
• Less averaging of noise and higher geometric
  uncertainties make ambiguity resolution more
  difficult when it fails with short tracks, the
  uncertainties become much larger

2
Results may vary significantly with time of day
A 2-hr session between 1200 and 1600 (left
plot) would be much more likely to resolve
ambiguities than a 2-hr session between 2000 an
2400 (right plot)
3
Continuous network in Italy used to test the
effect of session length and network
configurationon coordinate repeatabilities ?
Test site is PRATFiruzabadi and King (2011)
4
Time series for 2-hr sessions with 4 reference
sites.Note large sigmas on day 61 and outliers
on days 62 and 63 with these removed the rms is
2 mm horizontal and 7 mm vertical
5
Precision vs Session Length for Network Processing
Bars show repeatability in position over 31 days
of test site with respect to networks of 2 to 16
sites spanning 180-600 km. With at least 4
reference stations, outliers were less than 5
for sessions of 2 hrs or more.
6
Precision vs Session Length for Single Baselines
Bars show repeatability in position over 31 days
of a test site with respect to each of 7 sites
26-585 km away in single-baseline
processing. 10 of the 1- and 2-hr sessions had
large uncertainties and were omitted. 1-hr
results degrade significantly for baselines
longer than 100 km

7
Cerca del Cielo earthflow, Ponce, Puerto Rico
10 GPS monuments (including one continuous) on
the landslide, and 2 reference monuments
outside Steady-state flow 0.5 - 2.0 mm/d
Maximum 100 mm/d
From G. Wang, 2010
8
Time series of bi-weekly GPS surveys Mar-Dec 2008
20-minute occupations GPS 07 and 13 near the
head scarp GK03 and GP18 mid-slope
From G. Wang, 2010
9
Time series of hourly and daily positions over 10
days
10
GAMIT Settings for Sessions lt 3 Hours

• Consider using 15s sampling
• Run sh_gamit with the sesinfo option specifying
  the start time, sampling interval, and
  number of epochs
• If more than one session per day, run sh_gamit
  with the netext option, using a different
  letter for each session
• Dont decimate the preliminary or final solutions
  
• Decimation factor 1
• Quick-pre decimation factor 1

11
Kinematic GPS

• The style of GPS data collection and processing
  suggests that one or more GPS stations is moving
  (e.g., car, aircraft).
• The moving stations are kept stationary at the
  beginning and/or end of the track to resolve
  ambiguities then phase lock is maintained (as
  best as possible) through the track
• To obtain good results for positioning as a
  function of time it helps if the ambiguities can
  be fixed to integer values. Although with the
  back smooth option in track this is not so
  critical.
• Program track is the MIT implementation of this
  type of processing
• Unlike many kinematic processors,track pre-reads
  all data before processing. (But there is a
  real-time version, trackRT.)

12
General aspects

• The success of kinematic processing depends on
  separation of sites
• There are one or more static base stations and
  the moving receivers are positioned relative to
  these.
• For separations lt 10 km, usually easy
• 10gt100 km more difficult but often successful
• gt100 km very mixed results depending on quality
  of data collected. (Seismological example results
  are from 400km baselines.)

13
Track features

• Track uses the Melbourne-Wubbena Wide Lane to
  resolve L1-L2 and then a combination of
  techniques to determine L1 and L2 cycles
  separately.
• Bias flags are added at times of cycle slips
  and the ambiguity resolution tries to resolve
  these to integer values.
• Track uses floating point estimate with LC, MW-WL
  and ionospheric delay constraints to determine
  the integer biases and the reliability with which
  they are determined.
• Kalman filter smoothing can be used.
  (Non-resolved ambiguity parameters are constant,
  and atmospheric delays are consistent with
  process noise). When atmospheric delays are
  estimated, the smoothing option should always be
  used.

14
Basic input

• Track runs using a command file
• The base inputs needed are
• Obs_file specifies names of rinex data files.
  Sites can be K kinematic or F fixed
• Nav_file orbit file either broadcast ephemeris
  file or sp3 file
• Mode air/short/long -- Mode command is not
  strictly needed but it sets defaults for variety
  of situations

15
Some results

• Moving vehicle used for gravity measurments
  5-second sampling with stop and go
• GPS seismology 1 HZ tracking of earthquake
  surface wave arrivals

16
Track of Map-view track of vehicle motion over 8
km
17
Vehicle height vs time
18
Zoom of height just before power failure
19
Surface waves from the December, 2000, M 6 San
Simeon, Calliforna earthquake1 Hz sampling
20
Detail around arrival time.
Descriptiion and data on web site.
21
Summary

• Under favorable conditions and especially for
  short inter-site distances, both short-session
  static and kinematic processing can produce
  excellent results
• Use of more than a single reference site improves
  reliability
• TRACKs forward-backward filtering improves
  reliablity of non-real-time kinematic tracking
• -- BUT there is now a real-time
  version (trackRT) available