ash_z12.dft REPORT ON TEST AND DEMONSTRATION of the
ASHTECH Z-12 and
ASSOCIATED PROCESSING SOFTWARE
Federal Geodetic Control Subcommittee
Instrument Working Group
ABSTRACT
The Federal Geodetic Control Subcommittee (FGCS) tested and evaluated the Ashtech Z-12 and associated receiver processing software in March, 1994. The dual frequency receivers were tested in several independent operational tests: static, rapid static, pseudo-kinematic, and stop-and-go kinematic. The post-processing software, PRISM II, was employed to process data in the field using the predicted broadcast ephemerides, and to convert all data to Receiver Independent Exchange (RINEX) format. All kinematic and rapid static baselines were processed using the Ashtech PNAV software. Kinematic lines included baselines as long as 42 km with occupation times as short as 5 minutes. Least squares adjustment and analysis were performed using the Ashtech software, SNAP, which was also used for the creation of the G-file for National Geodetic Survey (NGS) bluebooking of data, analysis of repeat baselines, and performance of loop closure analysis. All baselines were used in the final adjustment and no scaling of a priori errors was required. All loop closures were better than 1 part per million (ppm). Horizontal positional error for the adjusted positions were all estimated to be 0.005 meter (m), or better. Vertical error ellipses were 0.012 m, or better.
All repeat baselines repeated at better than 1 ppm, except for two, which repeated at 4 ppm. Results, presented in the form of repeat vectors and adjustment errors, are compared with FGCS geometric accuracy standards and specifications for Global Positioining System (GPS) relative positioning. The analysis of results is based upon historical vector data gathered from testing that has been conducted since 19XX. The final adjustment indicated that FGCS Order A specification were met (1:10,000,000). Baselines varied in length from 43 meters to 108 kilometers. Overall, the results from the use of the predicted ephemeris in the solutions indicate that the Ashtech Z-12 will yield accuracies that meet or exceed the vendor's specifications.
INTRODUCTION
In March, 1994, the Federal Geodetic Control Subcommittee (FGCS) Instrument Working Group (IWG) conducted a test and demonstration of the Ashtech Z-12, a dualband (L1 and L2) GPS surveying system developed by Ashtech, Incorporated (Ashtech, Inc.) based in Sunnyvale, California. This was the nineteenth in a series of tests by FGCS to evaluate the performance of a surveying system based upon the GPS.
Field data was collected over 4 days, beginning on March 27, 1994 (DAY086), and ending on March 31, 1994 (DAY090). The first 3 days were reserved for static observations on the FGCS test networl consisting of 19 stations. DAY089 was devoted to data processing. The fourth day of observations on DAY090 was the manufacturer's choice to observe rapid static and kinematic observations, tied to 3 static base stations. Preliminary results were presented at a public meeting held at the National Institute for Standards and Technology (NIST), on Friday, April 1, 1994.
TEST DESCRIPTION
The test and demonstration were conducted over a 6-day period beginning Sunday, March 27, 1994, and ending Friday, April 1, 1994, on stations of the FGCS test network located in the vicinity of Washington, D.C. (Hothem and Fronczek 1983). Figure 1 is a sketch of the general layout of the FGCS test network. Many closely spaced stations are located within the grounds of the NIST in Gaithersburg, Maryland (Figure 2). Tests are conducted by the FGCS on various geopositioning systems, including those utilizing GPS equipment. The tests are not a certification process and are not conducted routinely on all equipment. The equipment that is tested is deemed to have a unique enough feature such that it warrants FGCS testing.
The FGCS testing network has been surveyed in to the highest accuracy standards and is tied to the Maryland State High Accuracy Reference Network (HARN), the Very Long Baseline Interferometry (VLBI) system, and oncludes stations that are part of the Continuously Operating Reference System (CORS) network. used to determine vectors between stations, holding only one position fixed. The analysis of the
For GPS testing, GPS observational data is obtained between stations of the FGCS testing network. Baseline vectors are obtained from the observational data, holding an FGCS specified point fixed. The coordinates for that specified station are supplied by the FGCS at the time of testing and change slightly from test to test, since they are based upon historical vector data that has been obtained from testing conducted since 19XX. The vector data is compared against historical vector data, and analysis is conducted based upon the findings of this type of comparison, inspection of loop closures and the like. Prior to the Ashtech Z-12 test, no adjustment was conducted on the data as part of the testing. With this test, a constrained adjustment was conducted holding three FGCS stations fixed to compare with the test vector data. Other sub-tests were conducted to assist with analysis of the total capability of the surveying system. This included a zero baseline test, a RINEX test and comparison with other receivers set up to gather data at the same time.
INSERT FIGURE
Figure 1. FGCS test network located in the vicinity of Washington, D.C.
INSERT FIGURE
Figure 2. Portion of the FGCS test network near Gaithersburg, Maryland.
The Ashtech GPS Z-12, as configured for the test, is an instrument capable of operation under any conditions, including when Anti-Spoofing (A/S) is activated. When A-S is activated, the Z-12 will track using an Ashtech patented Z-trackingTM mode that mitigates the effects of A/S. When A/S is not present, the Z-12 automatically reverts to P-code mode tracking. The receiver can measure in any mode: C/A pseudorange, L1 carrier, P1 code and P2 code, L2 carrier, all with full wavelength. Ashtech claims a decibel (dB) improvement in the signal-to-noise ratio (SNR) over the cross-correlating type receivers. It functions with full wavelength carrier phases on both P-code bands when A/S is enabled.
The second significant feature is dual-bit analog-to-digital (A/D) conversion that converts the received GPS signals into four levels of digital information. This provides substantial improvement in the performance of GPS receivers in the presence of interfering radio signals such as civilian communications, aircraft avionics, etc. Until now, with the exception of few military receivers, all commercial GPS receivers sued "one-bit conversion".
The Z-12 is a 12-channel all-in-view receiver, capable of static, rapid static, kinematic and pseudo-kinematic survey modes of operation. It has an approximately 2 minute cold start period, and less than 30 seconds for a warm start. It is capable of real-time differential GPS, and utilization of the RTCM format. The PRISM II and PNAV software packages are optional.
The higher SNR allows maintenance of lock on lower satellites, as well as maintaining lock during high ionospheric conditions. When in "Z-tracking" mode, the receiver maintains the same observables as when in P-mode, with independent P1 and P2, as well as full wavelength L2. There are two great advantages to having shorter correlation times for the same SNR:
1] The ability to track rapidly varying ionosphere with full observable accuracy. This cannot be accomplished with cross correlating receivers.
2] Acquisition transients settle in seconds while the competition has to wait minutes before their A/S observables reach equivalent accuracy.
The ability to derive useful information at low elevation is critically tied to SNR. When faced with low SNR, the user must either integrate for such a long time that there is essentially no data at low elevations, or accept huge errors. For all non-classified A/S on solution, the SNR falls off with elevation angle as the square of normal code SNR. That is, if the P mode SNR drops (with elevation angle) by a factor of 4, all civilian A/S techniques yield a drop in the SNR factor of 16.
There is also the ZY-XII option for organizations with appropriate clearance, which has an internal decryption board for full SA/AS Y-code capability. Currently, DMA, USMC and USAF are using this capability.
The receiver has an improved method of conversion of analog signal to digital over previous models, and allows better representation of the analog signal. This results in better jam immunity.
Ashtech employs a dual-line digital processing capability, which improves the jam immunity over other single bit receivers. The receiver does not lose lock near transmitters or high voltage power lines.
All data collected with the Ashtech Z-12 receiver were processed with the predicted broadcast satellite orbital coordinates. Data collected simultaneously with one or more other receivers are processed to determine vectors ( X, Y, Z) between occupied station points. The results of this post-processing and analysis were presented n the public meeting held on April 1, 1994, at NIST. The data was also post-processed using the precise ephemerides provided by the National Geodetic Survey (NGS). Those results are presented in this report.
Various observational modes were tested that employ one or more stationary receivers, and one or more mobile or roving receivers. These modes included conventional static, rapid static, stop-and-go, and kinematic.
Table 1. Observations attempted and achieved
Station Sunday Monday Tuesday Thursday 4-char. 27 March 1994 28 March 1994 29 March 1994 31 March 1994 ID DAY086 DAY087 DAY088 DAY090 ASTW Static / M/L ATHY Static / S/M GORF Static / S/M Static / M/L KINA Kinematic1 KINB Kinematic1 KINC Kinematic1 KIND Kinematic1 KINE Kinematic1 KINF Kinematic1 MDPT Static / M/L N102 NBS0 Static / S NBS1 Static / S NBS2 NBS3 Static / S NBS5 Static / S Static / S/M Static / M/L Kin. / RS Base OPTK Static / S/M Static / M/L Kin. / RS Base ORM1 Static / S Static / S/M SCOL Static / S/M Static / M/L Kin. / RS Base
Static / SB : Static Survey, Short Baselines
Static / S/M B : Static Survey, Short and Medium Baselines
Static / LB : Static Survey, Long Baseline
Kinematic1 : Two runs. First run occupation time, 1 minute. Second run, 0.1 min.
Kin. : Kinematic
RS : Rapid Static
Table 3. Status of GPS satellite constellation.
BLOCK I SATELLITES BLOCK II SATELLITES PRN CODE 3, 12,13,1,2,4,5,6,7,9,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,31 PLANE SLOT C4,A1,C1, F1,B3,D4,B4,C1,C4,A1,E1,D2,E3,D3,F3,A4,B2,E2,B1,E4,D1,A2,F2,A?,C2,F4,C3 CLOCK RB,RB,RB,CS,CS,CS,CS,RB,CS,CS,CS,CS,CS,CS,CS,CS,CS,CS,CS,CS,RB,CS,CS,CS,CS,CS,CS
Plane: six planes, A through F; Slot: four slots in each plane;
CS = cesium; RB = rubidium.
Table X. Ashtech Z-12 FGCS Test Chronology and Summary
DAY SURVEY MODE: STATIONS USED
DATE VECTORS COMMENTS
SURVEY TYPE
086 STATIC: NBS0, NBS1, Two 2 hour sessions:
27 March 1994 Short NBS3, NBS5, Session A: 1630-1830 UTC
baselines ORM1 Session B: 1835-2035 UTC
5 second recording interval; 5o elevation mask
Fixed height tripolds on every station
Power failure
on NBS0 and
NBS5, with
loss of
approx. 10
minutes data
each
087 STATIC: NBS5, SCOL, One 3 hour session, 1630-1930 UTC
28 March 1994 Medium and ATHY, GORF, 5 second recording interval; 5o elevation mask
long OPTK, ORM1 Fixed height tripolds on every station
baselines
Power failure
on SCOL, with
loss of
approx. 20
minutes of
data
088 STATIC: NBS5, SCOL, One 3 hour session, 1630-1930 UTC
29 March 1994 Medium and ASTW, GORF, 5 second recording interval; 5o elevation mask
long OPTK, MDPT Fixed height tripolds on every station except
baselines GORF
No problems
089 KINEMATIC and BASE RAPID STATIC RUN ON NIST Test course
30 March 1994 RAPID STATIC: STATIONS: No stationat less than 500 m away from NBS5
Day set aside NBS5, SCOL, base station;
for data OPTK First run rapid static occupation for 1 minute
processing; No ROVERS: Second run rapid static occupation for 0.1
surveying minute
conducted
090 KINEMATIC TRACK RUN
1 April 1994 5 stations spaced 5 km apart, placing the
furthest station25 km from NIST test site
Outbound test: Receiver left on between site
occupations of 5 minutes per occupation
Returning
Inbound test:
Rapid static
on same five
stations with
receiver
turned off
inbetween
site
occupations
of varying
duration
(4-13
minutes)
FIELD OBSERVATIONS. Field data collection by FGCS observers and Ashtech field personnel took place between March 27, 1994 (DAY086) and March 31, 1994 (DAY090). Observations were done on DAY086, DAY087, DAY088, and DAY090. DAY089 was used to concentrate on data processing. Data collected during the first three days were for static processing required by the FGCS. The fourth day was the manufacturer's choice of rapid static and kinematic observations tied to three static base stations.
STATIC MODE OBSERVATIONS. For all static baseline sessions, the receivers were set to a 5 second sampling rate, with a 5 degree masking angle. Survey data with accompanying meteorological data ??? were collected during the XX independent station occupations. There were 3 days of static baseline observations, over baselines ranging from 43 m to 108 km.
Short Baselines. The short baselines ranged from 43 m to 1300 m (1.3 km). Data for the short baselines was collected using 5 receivers on Sunday, Day086. Two sessions of 2 hours (Session A: 1630 UTC until 1830 UTC; Session B 1835 UTC until 2035 UTC) each were collected during a midday window with an average of 6-7 available satellites. For both sessions, data was recorded at a 5 second sampling rate with a 5 degree elevation mask. Battery problems occurred at two sites, NBS5 and NBS0, causing a 10 minute loss of data at each site.
Short to Medium Baselines. The medium baselines ranged from 6 km to 42 km. A single 3 hour session was observed in the field, during the same window (1630 UTC until 1930 UTC) used on Sunday. Six receivers were used for the Monday, Day087 observations. A single session of 3 hours was recorded, using a 5 second sampling rate with a 5 degree elevation mask, just as the day before. On Monday, a bad battery created a 20 minute loss of data at one station, SCHOOL (SCOL).
Medium to Long Baselines. On Tuesday, Day088, baselines of 22 km to 108 km were observed using 6 receivers. Again, just as the day before, a single 3 hour session of data was collected using the same midday window (1630 UTC until 1930 UTC), 5 second data sampling rate and 5 degree elevation mask angle. All collection went well and no problems occurred in the field.
DAY 086
The five stations observed were NBS0, NBS1, NBS3, NBS5, and ORM1. These stations were occupied for two 2-hour sessions at a 5 second recording interval and a 5 degree elevation mask. All stations had fixed height tripods provided by FGCS. Session A started at 1630 UTC and ended at 1830 UTC. Session B was scheduled to obtain redundant measurements for the short baselines measured in Session A. Since fixed height tripods were used, no tripod resetting was deemed necessary. Session B started at 1835 UTC and ended at 2035 UTC. Momentary loss of battery power at stations NBS5 and NBS0 caused an approximate 10 minutes loss of data at each station, otherwise no significant problems were encountered.
DAY 087
The six stations observed were NBS5, SCOL, ATHY, GORF, OPTK, and ORM1.
All stations were occupied for one 3-hour session. Data was collected every 5 seconds with a 5 degree elevation mask. The session began at 1630 UTC and ended at 1930 UTC. Fixed height tripods were used at all sites except GORF. All stations report periods of significant rainfall.
The only problem was a power problem at SCOL. The batteries died at approximately 1730 UTC. Another battery was connected to the receiver, and the session continued. The data files were later joined together. A total of about 20 minutes of data was lost.
DAY 088
The six stations observed were NBS5, SCOL, ASTW, GORF, OPTK, and MDPT.
All stations were occupied for one 3-hour session. Data was collected every 5 seconds with a 5 degree elevation mask. The session began at 1630 UTC and ended at 1930 UTC. Fixed height tripods were used at all sites except GORF.
No problems were reported.
Day 089
No observation. Set aside for data processing.
Day 090
Several kinematic and rapid static tests were performed during an early morning window consisting of an average of 7 to 8 satellites. The utilized three base stations and two rover receivers. The three base stations were NBS5, SCOL, and OPTK. All receivers record data every 2 seconds with a 5 degree elevation mask. The base sites collected data from 0900 UTC to 1300 UTC.
The first rover receiver performed two runs on the kinematic track at NIST. The track consists of a cluster of points no more than 500 meters from the base station NBS5. The first run occupied each point for 1 minutes. The second run utilized an occupation time of 0.1 minutes.
The second rover performed a kinematic run on the kinematic track located outside of the NIST grounds. The track consists of 5 station spaced 5 sm. apart, placing the last station 25 km from the NIST grounds. In one direction, the receiver was left on in between site occupations. Site occupation times on this run were 5 minutes. On the return, a rapid static test was performed on the same five stations. The receiver was turned off between site occupations. The occupation time was determined by the formula: ((baseline length (km) + recording interval (sec)) / 2). Based on this formula, the occupation times for the Rapid Static run were 13 minutes, 11 minutes, 8 minutes, 6 minutes, and 4 minutes respectively.
No problems were reported for either the base sites or the rover receivers.
Table 1. Summary of static survey observations.
Station Sunday Monday Tuesday 4-char. 27 March 1994 28 March 1994 29 March 1994 identification DAY086 DAY087 DAY088 ASTW Static / M/L GAIT GORF Static / S/M Static / M/L KINA KINB KINC KIND KINE KINF MDPT Static / M/L N102 NBS0 Static / S NBS1 Static / S NBS2 NBS3 Static / S NBS5 Static / S Static / S/M Static / M/L OPTK Static / S/M Static / M/L ORM1 Static / S Static / S/M SCOL Static / S/M Static / M/L
Static / SB : Static Survey, Short Baselines Kinematic1 : Two runs. First run occupation
Static / S/M B : Static Survey, Short and Medium Baselines time, 1 minute. Second run, 0.1 min.
Static / LB : Static Survey, Long Baseline
Kin. : Kinematic
RS : Rapid Static
KINEMATIC SURVEY FIELD OBSERVATIONS. On Thursday, several Kinematic tests and a Rapid Static test were performed during an early morning window consisting of an average of 7-8 satellites. All tests utilized three base stations, on located on the NIST grounds, one located approximately 7 km out and one located approximately 16 km out. The data sampling rate was set to a 2 second interval and a 5 degree elevation mask was employed. Two Kinematic runs were performed on NIST grounds on a cluster of points no more than 500 meters from the base station on NIST. The two runs were identical with the exception of occupation time. The first run utilized a point occupation time of 1 minute. The second run utilized an occupation time of 0.1 minute.
Kinematic Observations. On Thursday, several kinematic tests and a rapid static test were completed in an early morning window that had 7-8 satellites usually available. All the kinematic tests used three base stations, all located on the NIST ground. The base stations were located approximately 7 km out and one was located approximately 16 km out.
Two kinematic runs were performed on the NIST grounds on the kinematic test course on a cluster of points no more than 500 meters from the base stations on NIST. The two runs were identical with the exception of occupation time. The first run utilized a point occupation time of 1 minute. The second run utilized a point occupation time of 0.1 minute.
A third kinematic run was performed on a line of 5 stations spaced 5 km apart placing the last station 25 km from the NIST grounds. Occupation times for each station on this run were set to 5 minutes. Satellite lock was lost several times, due to obstructions from overpasses and trees.
Rapid Static Observations. After the third kinematic run was completed, the team turned around and repeated the test on the same points using the rapid static technique. return trip, a Rapid Static test was performed on the same 5 stations. Data was collected on each station after which the receiver was turned off until reaching the next station. Occupation times on the stations were based on the equation 'occup = (length of baseline (km) + recording interval)/2'. Based on this equation , the occupation times for the Rapid Static run on the 5 stations were 13 minutes, 11 minutes, 8 minutes, 6 minutes, and 4 minutes respectively.
After the data collection on Thursday, the Kinematic and Rapid Static data was analyzed and a report was generated for presentation the next day. All material needed to be supplied to FGCS prior to leaving was gathered.
Table 6. Summary of rapid static observations.
Session Number of Number of Average
Satellites
Stations Reference Duration
Stations (min)
Start/Min/Max
/End
Day 5 3 8
090
Table X. Summary of kinematic and rapid static observations.
Station Thursday, Thursday,
4-char. 31 March 31 March
identificati 1994, 1994,
on DAY090 DAY090
Kinematic
Observations
Rapid
Static
Observations
ASTW
ATHY
CENA
EAST
GAIT
GORF
KINA
Kinematic1
KINB
Kinematic1
KINC
Kinematic1
KIND
Kinematic1
KINE
Kinematic1
KINF
Kinematic1
KINM
MDPT Static /
M/L
N102
N104
NBS0
NBS1
NBS2
NBS3
NB42
NBS5 Static /
M/L
Kin. / RS
Base
NIST
ONSC
OPTK Static /
M/L
Kin. / RS
Base
ORM1
POWL
QUAL
REMO
ROD1
SCOL Static / Kin. / RS
M/L Base
kIN / rs
bASE
ts19
Static / SB : Static Survey, Short Baselines Kinematic1 : Two runs. First run occupation
Static / S/M B : Static Survey, Short and Medium Baselines time, 1 minute. Second run, 0.1 min.
Static / LB : Static Survey, Long Baseline
Kin. : Kinematic
RS : Rapid Static
Data were collected for each of the observing modes and processed
with the software PRISM (static-kinematic). Since it
was impractical to establish a method for evaluating positions
for each individual epoch along a kinematic trajectory, kinematic
trajectories were not evaluated. Only a single receiver
was used as a mobile receiver in the rapid static, reoccupation,
kinematic, and stop-and-go tests. All modes were successfully
tested as planned. Broadcast ephemerides were used to forecast
satellite positions for the scheduling of observing sessions.
The observation windows were modified, however, because
of a malfunctioning satellite and because the receiver had no
option for overriding the transmitted health status of a satellite.
This is from Leica report...
Data were acquired independently for each mode tested. There were two exceptions: (1) the reoccupation mode consisted of combining the data from two rapid static sessions for sessions XXXX and (2) receivers used for conventional static testing were also used as the reference trackers for other tests.
Table 2 summarizes for each session the approximate starting and ending time in UTC (Coordinated Universal Time), number of stations occupied, the PRN (pseudorandom noise) code for the satellites tracked, number of satellites at beginning and end of each session, and maximum and minimum number of satellites observed during each session.
Table 2. Observation summary, Ashtech GPS Z-12 FGCS test survey
Session Starting Number Number Satellites observed
and ending of of rover (PRN code)
time (UTC) static stations Number of Satellites
(1) stations S/Max/Min/E
86A 1630 - 1830 5 -- 3,6,14,18,19,22,25,28,
29,31
86B 1835 - 2035 5 -- 6,15,18,19,22,27,28,29
,31
87 1630 - 1930 6 -- 3,6,14,18,19,22,25,27,
28,29,31
88 1630 - 1930 6 -- 3,6,14,18,19,22,25,27,
28,29,31
90A (1 0900 - 1030 3 5 1,5,9,12,15,17,20,21,2
min) 3,25
90B (0.1 1120-1220 3 5 1,5,9,12,15,20,21,23,2
min) 5
90 C(5 0900-1030 3 5 1,5,6,12,15,17,20,21,2
min) 3,25,26,28
90 D 1030-1230 3 5
(RS)
TOTALS
(1) Subtract 4 hours to convert UTC to local time.
The status of the GPS satellite constellation, based on information available from the United States Coast Guard Navigation Information Service (USCG NIS) Bulletin Board Service (BBS) on XXXXX date is summarized in Table 3. The accuracy for the predicted (broadcast) satellite orbital coordinate data used in the baseline solutions was estimated to be about 1 mm/km (1 ppm) at the 1-Í level.
ADD TABLE 3 HERE
Table 3. USCG Navigation Information Service (NIS) data on the status of the GPS constellation at the time of the test.
DATA PROCESSING
On Wednesday, the Ashtech team examined the processed data and determined the method of processing which yielded the best looking results:
L1 only on baselines of less than 15 km
Wide lane on baselines of greater than 15 km
Performed final adjustment of data
The three days of data were converted to :
RINEX
Bluebooked
and backed up to tape
On Friday, April 1, 1994, a three phase presentation on the results was given to approximately 20 people. Roy Anderson of NGS presented details on the test course, observation window, and the data collected. Larry Hothem presented the results of the 3 days of static data collection on the FGCS test course. Bill Martin presented the results of the Kinematic and Rapid Static data collection, the results of the analysis of the 3 hour observations broken down into short data segments, and a short presentation on the Z-XII itself.
In addition to the standard processing of the long data sets, three baselines, ranging from 1.3 km to 22 km were selected for further evaluation consisting of chopping up the data sets into shorter occupation spans and processing each of theses spans to compare the results with the total data set.
NOTE THIS IN A TABLE..
Partial processing had been completed by Ashtech at noon on Friday, April 1. This processing included the conversion of all data (except kinematic data) to Receiver INdependent EXchange (RINEX) version 2.0 format, and the processing of n-1 independent radial vectors. NOTE: UPDATE AS NECESSARY A full set of n*(n-1)/2 vectors and adjusted results were provided after the test. Preliminary results were provided for analysis and presentation at a public meeting held on Friday afternoon, April 1, 1994.
Ashtech's PRISM software, version 1.04, was used
to generate vector results. Compatible output files were generated
by PRISM for input to SNAP, a 3-dimensional least squares adjustment
program.
PRISM DESCRIPTION
Two commonly available output file formats, compatible with public domain adjustment software developed and used by the NGS for the inclusion of GPS data into the National Geodetic Reference System (NGRS) database, are the GFILE and BFILE (White and Love, 1991; FGCC, 1991). NOTE: UPDATE AS NECESSARY The GFILE, a file containing solved vectors for a GPS project, was not supported by PRISM software. An independent piece of software had been invoked to reformat PRISM output files into the standard GFILE format. The BFILE, a compendium of site-specific data for a GPS project, was constructed manually from observer log sheets. Code support within PRISM for NGS file formats may be available in the future.
Using the RINEX format allows the processing of data collected using different receiver types to be processed together. As the number of receivers on the market increases, the number of people co-observing with different receiver types is also increasing. Data in RINEX version 2.0 format were provided by Ashtech. Data was converted to the RINEX 2.0 format through an option available within the PRISM software. Though an option to PRISM is available to import RINEX version 2 data from mixed receivers, this was not tested.NOTE: UPDATE AS NECESSARY Did we ever get the Antarctica data processed and what were the results?
PRISM processing is automatic after setting some processing parameters, session sites, and reference stations. For the test the processing parameters were set to the following values, unless otherwise indicated:
Cutoff angle (degrees) : 5 Tropospheric model : Hopfield Ionospheric model : Standard Ephemeris : Broadcast Data used : Use Code and Phase Frequency : L1 + L2 Limit to resolve ambiguities (km) : 20 Kinematic chain computation rate (epoch) : 1 a priori rms (mm) : ???
For future tests, the cutoff angle will be set at 5 degrees. Post-processing packages allow the data to be processed at a user specified cutoff angle, for example, 15 degrees. Collecting data at the lower cutoff angle of 5 degrees will provide more complete data sets for post-processing and analysis efforts.
The PRISM software uses standard weather data with either the Hopfield (Modified Hopfield also and any others???) or Saastamoinen model, to compensate for the effect of tropospheric propagation delays due to the troposphere. (Add in appropriate references to cite here and add to the bibliography). Surface weather measurements for temperature, humidity, and atmospheric pressure were not used in the solutions. (These were provided, were they not?)
Carol's note: Carol! What do you mean to say??? L1 processing for short lines L1C long --> float solution Wide lane -- attempt to fix integers ~15 km ==> float can use wide lane to fix ??? longer processing times.
Need explanation of database system/method here and further explanation, perhaps, of the philosophical choice and direction Ashtech has taken on float vs fixing integers, automation etc.
NOTE: UPDATE AS NECESSARY A sample of the changes in propagation delay encountered in the test is shown in Figure 3. This is a double-difference residual plot made with an independent software package, OMNI, for the longest baseline of 108 km between stations ASTW and SCOL on DAYXXX. The magnitude of residuals for double-difference solutions using single frequency data, either L1 or L2, is approximately one cycle peak to peak. A linear-combination solution (Lc) of the L1 and L2 observables provides a first order estimate for the atmospheric refraction correction. In Appendix 3 are tables containing results provided by Ashtech for conventional static, rapid static and stop-and-go solutions.
RESULTS
Adjustments: The 'usual' minimally constrained adjustment that holds GORF fixed and the constrained adjustment that held GORF, ATHY and ASTW fixed to remove the biases. Include any explanation of the net, testing etc.?
NOTE: UPDATE AS NECESSARY The quality of the solutions are evaluated by examining repeat vectors and residuals from a minimally-constrained least-squares adjustment. First, the components of repeat vectors are compared. This is given as the minimum and maximum component difference., Secondly, the residuals of the minimally constrained adjustments are compared with FGCS geometric accuracy standards (FGCS, 1988). These standards are summarized in Table 4. Lastly, the test results are evaluated with respect to the expected baseline precision as specified by the vendor. The baseline precision with PRISM software for static, rapid static, reoccupation, and stop-and-go kinematic, are stated in the Ashtech technical specifications for the Ashtech GPS Z-12 receiver which can be found in Appendix 1.
INSERT FIGURE
Figure 3. Double difference residuals for 108 km baseline using the OMNI software.
NOTE: GET ACCURATE DESCRIPTION BOOK TOGETHER OF THE NGS STATIONS AND HOW TO ACCESS THE DATABASE. PERHAPS AS A SEPARATE PUBLICATION TO BE UPDATED AS NECESSARY AND PUBLISHED BY THE FGCS.
Environmental factors may have influenced the test results. The results may have been affected by trees, buildings, nearby power transmission lines, and terrain at some of the FGCS test station sites. These test conditions are uniform from test to test, however, and frequently characterize GPS survey conditions. Thus, the FGCS stations and ambient conditions are useful in determining the effectiveness and capabilities of the GPS survey instruments and processing software.
Table 4.FGCS geometric accuracy standards
FGCS Accuracy Standards (2 sigma or 95% confidence limit) Order Definition ± [( 3 mm)¨ + (d x 0.01 mm/km)¨]½ B ± [(10 mm)¨ + (d x 10 mm/km)¨]½ A ± [( 8 mm)¨ + (d x 1 : mm/km)¨]½ AA ± [( 5 mm)¨ + (d x 0.1 mm/km)¨]½
From a set of XXXXX baselines, repeated at least twice in XXXX total sessions, the vector components were compared for each set (Table 5). Using all XXXXX observed baselines, a minimally-constrained adjustment yielded mean component ( X, Y, Z) residuals of XXX, XXX, XXX, respectively. Figure 4 graphically presents the component residuals and FGCS geometric accuracy standards.
Table 5A. Variations in components of repeat baselines using 20-25 degree cutoff angles (results provided by Ashtech): conventional static observations
Vector Minimum (m)
Component Maximum (m)
X 0.000
0.008
Y 0.000
0.023
Z 0.000
0.011
Table 5B. Variations in components of repeat baselines using 20 degree cutoff angle for Day086 and 25 degree cutoff angle for Day087 and Day088: conventional static observations
Vector Minimum (m)
Component Maximum (m)
X 0.000
0.005
Y 0.000
0.018
Z 0.000
0.009
Table 5C. Variations in components of repeat baselines using 20 degree cutoff angle: conventional static observations
Vector Minimum (m)
Component Maximum (m)
X 0.000
0.006
Y 0.000
0.018
Z 0.000
0.009
Table 5D. Variations in components of repeat baselines using 15 degree cutoff angle: conventional static observations
Vector Minimum (m)
Component Maximum (m)
X 0.000
Y 0.000
Z 0.000
Table 5E. Variations in components of repeat baselines using 10 degree cutoff angles: conventional static observations
Vector Minimum (m)
Component Maximum (m)
X 0.000
Y 0.000
Z 0.000
Table 5F. Variations in components of repeat baselines using 5 degree cutoff angles: conventional static observations
Vector Minimum (m)
Component Maximum (m)
X 0.000
Y 0.000
Z 0.000
INSERT FIGURE
Figure 4. Residuals from least-squares adjustment: conventional static observations.
Rapid static observations were taken at 14 stations in 4 sessions. A summary is given in Table 6. The analysis of vector components, independent of recording duration, for all 22 repeat rapid static vectors is found in Table 7.
Table 7. Variations in components of repeat baselines: rapid static observations
Vector Minimum (m)
Component Maximum (m)
X 0.001
0.296
Y 0.005
0.575
Z 0.001
0.776
INSERT FIGURE
Figure 5. Residuals from least-squares adjustment: rapid static observations.
Rapid static observations acquired approximately 3 hours apart at the same stations in sessions XXX and XXX were processed in the reoccupation mode. This resulted in 6 common baselines with lengths less than 2.5 km. The components were evaluated by comparing the residuals from the rapid static adjustment against the differences between the adjusted rapid static baseline and the reoccupation baseline solutions. A summary of these comparisons is given in Table 8. The reoccupation processing method produced slightly better results compared with the rapid static solutions.
Table 8. Comparison of baseline solutions: rapid static observations and reoccupation
Minimum (m) Mean (m)
Maximum (m)
Rapid static
Resx 0.004 0.014
Resy 0.005 0.027
Resz 0.002 0.028
Reoccupation
Resx 0.003 0.140
Resy 0.005 0.274
Resz 0.005 0.207
STOP-AND-GO KINEMATIC
Twelve stop-and-go vectors were repeated at least twice, their maximum length being approximately 0.55 km. The technique requires continuous phase lock on four or more satellites. The data acquisition interval is usually short, in this case 30-60 seconds. Table 9 shows that the minimum and maximum differences for the results from the repeat baselines.
Table 9. Variations in components of repeat baselines: stop-and-go observations
Vector Minimum (m)
Component Maximum (m)
X 0.077
0.892
Y 0.102
1.246
Z 0.031
0.944
A least-squares minimally-constrained adjustment of 31 vectors from 10 stations, yielded mean residuals for cartesian vector components ( X, Y, Z) of 0.003m, 0.003m, and 0.003m, respectively. NOTE: UPDATE
CONCLUSIONS
Results Summary
PRISM II was used for all data processing and raw data file conversion to RINEX. All data processing went very smoothly. The data collected by the Z-XII was extremely clean. No data editing was required throughout the entire processing. All processed baselines were used in the final least squares adjustment. This fact stands out in comparison to the FGCS test of the P-XII receiver where much data editing and manual processing was required to result in a clean adjustment.
Static baselines were processed repeatedly using different processing options (L1 only, L1C, Wide lane). Through analysis of the residual plots and adjustment, it was found that the best baselines resulted from L1 only processing on baselines shorter than 15 km and Wide lane processing on baselines longer than 15 km. All baselines, including the 108 km line, fixed biases. An adjustment was performed on these baselines.
PNAV was used to process all Kinematic and Rapid Static baselines. All stations occupied using Kinematic and Rapid Static procedures processed without any failures. These included baselines as long as 42 km with occupation times as short as 5 minutes.
SNAP was used for the least squares adjustment, creation of the G-file for NGS Bluebooking of the data, analysis of repeat baselines, and performance of loop closure analysis. All baselines were used in the final adjustment. No scaling of the baselines a priori errors were required. The final adjustment met FGCC Order A specifications (1:10,000,000). All repeat baselines repeated at better than 1 ppm except for 2 which repeated at 4 ppm. All loop closure results were better than 1 ppm. Positional errors for the adjusted positions were all estimated to be better than 0.005 meters.
Conclusion
The results of this test could not have been better. All data collection and data processing was performed without a failure or problem. The data was extremely clean, requiring no manual editing of data or processing parameters. The Z-XII receiver is truly in a class of its own. PRISM II and SNAP performed very well, facilitating the accuracy's and repeatability that resulted from this test.
Congratulations are in order for the people directly involved with the FGCS test. The team represented Ashtech in a very professional manner.
NOTE: UPDATE Based on the analysis of observations from the FGCS test survey, the FGCS geometric relative positioning accuracy standards are compatible with the performance of the System 300 for the various observation modes is shown in Table 10. The comments qualify the conditions upon which the classifications are based. To achieve orders A and AA, it is assumed that PRISM software is capable of producing results from fixed orbital coordinate data solutions that are limited only by the accuracy of the orbit coordinate data.
Overall, the results from the predicted ephemeris solutions indicate that the Ashtech GPS Z-12 survey system will produce accurate results that meet or exceed the vendor's specifications. (See Appendix 2).
In conclusion, analysis of the results from the FGCS test survey conducted in April 1995 on the Ashtech Z-12, collected with four or more satellites in an acceptable geometric configuration, processed with double-difference software using orbital coordinate data accurate to 2 mm/km (2 ppm) or better, will yield accuracies that should meet requirements for most geodetic surveying needs.
We were hunting mms here
Mention anything about field procedures ie not black box...too much reliance on automation can lead to less rigorous knowledge of survey practices and the danger exists that data analysis will not be performed to the extent that it should be
Table 10. FGCS accuracy standards achieved for tested survey modes
Observation Standard Maximum
Type Baseline
Achieved(1) Length (km)
(2) Observation
Duration
(min)
Conventional A 108
Static 180
Rapid static A
Reoccupation A
Stop-and-Go A
kinematic
(1) Results from data collected when AS is not activated.
(2) Solutions employ L1 and L2 data from ionospherically free fixed integers for baselines less than 20 km and non-fixed integers for baselines greater than 20 km.
REFERENCES
Federal Geodetic Control Committee, 1984: Standards and specifications for geodetic control networks. September 1984, reprinted February 1991, 29 pp. National Geodetic Survey Information Branch, NOAA, Rockville, MD 20852.
Federal Geodetic Control Committee, 1989: Geometric geodetic accuracy standards and specifications for GPS relative positioning surveys. Version 5.0, May 1988, reprinted with corrections August 1, 1989, 40 pp. National Geodetic Survey Information Branch, NOAA, Rockville, MD 20852.
Federal Geodetic Control Committee, 1991: Input formats and specifications of the National Geodetic Survey Data Base, Volume I. Horizontal Control Data (includes GPS). Reprinted June 1991, 250 pp. National Geodetic Survey Information Branch, NOAA, Rockville, MD 20852.
GEOsurv Inc., 1990: GeoLab Manual, The Baxter Centre, Ottawa, Ontario, Canada.
Hothem, L.D., and Fronczek, C.J., 1983: Report on test and demonstration of MACROMETERTM model V-1000 interferometric surveyor. FGCC Report IS-83-1, Federal Geodetic Control Committee. National Geodetic Information Branch, NOAA, Rockville, MD 20852.
Hothem, L.D., 1990: Test and demonstration of GPS Satellite Survey Systems. FGCC Report IS-90-1, Federal Geodetic Control Committee. National Geodetic Information Branch, NOAA, Rockville, MD 20852.
Milbert, Dennis G. and Kass, William, 1985: ADJUST: The horizontal observation adjustment program, NOAA Technical Memorandum NOS/NGS-47.
White, Madeline B. and Love, John D., 1991: Submitting GPS projects to the National Geodetic Survey: How and what?, Technical Papers of the GIS/LIS ACSM/ASPRS 1991 Fall Convention of the ACSM/ASPRS, Atlanta, GA, 75-83.
APPENDIX 1
Ashtech Z-12
APPENDIX 2
FGCS Geometric Accuracy Standards
vs. Ashtech Specifications
APPENDIX 3
Figures and Tables Provided by Ashtech, Inc.