[ANSS-netops] Azimuth Pointing System (APS) Testing at ASL/USGS

[Kyle Persefield]

All,

I want to stress that we tested the APS when it was new and we had 2 
days to play with it in the May 2009 time frame.  At ASL we never got 
a GPS integrity above 80% with the demo unit.  This figure was a bit 
disappointing because we think the unit is capable of 0.1 degree 
accuracy with a GPS integrity >80%.  At the time, we felt with a GPS 
integrity > than 60% that the unit was capable of 0.4 degrees of 
accuracy and that is a good as most methods.  The best methods are 
still the sun shot with a theodolite and the IxSea Octans FOG 
gyrocompass.  I am sure that the APS has matured since we tested.

I had a recommendation to buy these units.  They are easy to use, but 
they are still GPS and must be used with a good sky view, good 
weather, and a good elevation mask.  The user must understand the 
limitations of GPS.  The unit has a fixed pistol scope and you will 
have to purchase a good sturdy camera tripod.  Lines can be 
translated so long as you don't have to make multiple turns into 
vaults or tunnels.

I have attached the APS user manual (1st addition).  The manual may 
have gone through a revision or two since.


Below is some E-mail from the May 2009 time frame:


Bob and Gary,

Could you approve the below draft for release to the manufacturer.  I 
would like to do this today and return their unit to them.

Dear Bruce,

I want to thank you for allowing Steve to give us the opportunity to 
test the APS for a couple of days here at our lab.  What follows is 
the results from my observations.  I then I asked our lab director 
Lind Gee to have her husband Mark Murray weigh in on the 
results.  Mark works at New Mexico Tech and his life is GPS.  Mark 
ran a sky plot of our GPS monument station P034 (attached).  Mark 
reported that our visibility is excellent at the lab, and that he 
thought our azimuth results were reasonable.  But, I think we were 
all a little surprised that we did not obtain a GPS integrity of over 
80% at any time at our lab.  I do not know how the GPS integrity is 
derived at.  Could you explain this a little bit?  It may be that the 
APS we had is less sensitive than it should be.  Overall, we are 
happy with the results and I made a recommendation to buy.

What follows is some E-mail conversations with regard to testing the APS:

Lind,

I recommend that we buy this at least for the ANSS Depot and the ANSS 
Backbone so long as it is clearly understood that there are 
limitations for ideal GPS reception.  Clearly, a written 
procedure  needs to be developed for the field stating conditions for 
use.  Ideally, I think a GPS integrity of > 60% is desirable in the 
field.  This is a WAAS enabled unit and I don't understand if WAAS 
improves the azimuth determination or what.  I will ask the 
manufacturer.  If WAAS improves the azimuth determination, then it's 
utility might be less accurate for the GSN.

I will write the manufacturer and tell him of our results.  I will 
also ask how the BIS made the initial ECCN 7A994 export restriction 
determination.  Right now, I don't think the APS is quick enough to 
make a azimuth determination to be on the list.

GPS reception factors to consider are the terrain elevation mask, 
large buildings, multi-path sources, foliage, and cloud cover.  The 
manufacturers stated accuracy figures for the azimuth derived from 
the GPS integrity percentages are correct,  with a little fudge 
factor favoring the manufacturer.  In other words, it performs a bit 
better than they state.

The highest GPS Integrity obtained at ASL over the period of testing 
never exceeded 77% and on average was only 67%.  I believe the cloud 
cover and our elevation mask to the East and Southeast, due to the 
mountains, possibly contributed to this lower than expected figure.

Weather conditions on the afternoon of 20 May 2009, were partly 
cloudy with thunderstorms building to the West and East.
Weather conditions on the morning of 21 May 2009, were mostly cloudy 
with light, widely scattered rain.

Using the ASL reference north mark the following applied:

60% GPS INT, -0.3 to + 0.4 from 0.0 reference
67% GPS INT, -0.3 to + 0.2 from 0.0 reference
70% GPS INT, -0.2 to + 0.1 from 0.0 reference
77% GPS INT, -0.1 to + 0.1 from 0.0 reference

GPS integrity check under cottonwood foliage canopy in the Rio Grande 
bosque over 22:03 to 22:33 UTC, 20 May 2009:

37%, 43%, 60%, 27%, 33%, 60%, 53%, 57%, 53%
GPS integrity never exceeded 60%.  The foliage does seem to have an affect.

GPS integrity check at home over 23:00 to 23:58 UTC, 20 May 2009 
without mountain elevation mask as at ASL.

43%, 70%, 67%, 70%, 77%, 67%, 80%, 77%, with most of the time period 
spent at 77% or 10% better than ASL.  This was the only time I had 
80% integrity.

This test under foliage, cloud cover and a less than ideal elevation 
mask, represents less than desirable field conditions.  With a worst 
case 60% GPS integrity, a 0.4 degree accuracy can be obtained.  I 
believe that the APS is better than a Brunton compass and magnetic 
declination correction even at this 60% GPS integrity figure.

Under the conditions of this test, a GPS integrity of between 70 and 
77% was obtained in under a 1/2 hour at my home and at ASL.  The 
azimuth accuracy than jumps to 0.2 to 0.1 degrees in this percentage range.

Lind - Does Mark know of any software or any online service that 
provides a chart or table that predicts the Dilution of Precision 
(DOP) for a given location at a future time?  What aids like this are 
out there that could help pick the best time to look for the best GPS 
constellation?

The APS is very simple to use, but because the scope on the unit is 
fixed, we would have to purchase a sturdy camera tripod that allows 
for vertical movement.  Surveying tripods don't allow for this.

For our application, I see no utility in ordering the laser 
rangefinder.  This would bring the cost to around $6500 per unit.

Below is a small sample of text from a downloaded data file when we 
had 70% GPS Integrity:

$GPGGA,181448.02,3456.75707,N,10627.42710,W,2,08,1.2,1839.1,M,-23.1,M,,STN*26
$PLTIT,HV,,,360.0,D,0.1,D,,*63
$GPGGA,181449.01,3456.75707,N,10627.42710,W,2,08,1.2,1839.1,M,-23.1,M,,STN*24
$PLTIT,HV,,,360.0,D,0.1,D,,*63
$GPGGA,181450.00,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*25
$PLTIT,HV,,,360.0,D,0.1,D,,*63
$GPGGA,181451.00,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*24
$PLTIT,HV,,,360.0,D,0.1,D,,*63
$GPGGA,181451.99,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*24
$PLTIT,HV,,,360.0,D,0.1,D,,*63
$GPGGA,181453.00,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*26
$PLTIT,HV,,,359.9,D,0.1,D,,*60
$GPGGA,181454.01,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*20
$PLTIT,HV,,,0.0,D,0.2,D,,*65
$GPGGA,181455.00,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*20
$PLTIT,HV,,,0.0,D,0.1,D,,*66
$GPGGA,181455.99,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*20
$PLTIT,HV,,,359.9,D,0.1,D,,*60
$GPGGA,181457.00,3456.75707,N,10627.42710,W,2,07,1.4,1839.0,M,-23.1,M,,STN*22
$PLTIT,HV,,,0.1,D,0.1,D,,*67
$GPGGA,181458.02,3456.75707,N,10627.42710,W,2,08,1.2,1839.0,M,-23.1,M,,STN*26
$PLTIT,HV,,,0.1,D,0.1,D,,*67

MSG ID, UTC Time, LAT, N/S, LONG, E/W, FIX, SATS in view, HDOP, MSL, 
Meters, Height above WGS84, Meters, Checksum
MSG ID, Horz Vector, Blank, Blank, Azimuth, Degrees, Inclination 
Angle, degrees, checksum

For a sure response send E-Mail to gsnmaint at usgs.gov, or 
gsn-xxxx(your station code)@usgs.gov

Regards,

Kyle E. Persefield
Honeywell Int'l HTSI, Field Engineer
USGS Albuquerque Seismological Laboratory Contractor

Hi all,

Some comments from Mark regarding Kyle's report on the APS.  He did a
skyview assessment of the PBO site and comments that the visibility is
excellent.  So I am still surprised that the data integrity was so
low .....


----- Forwarded by Lind S Gee/GD/USGS/DOI on 05/22/2009 01:03 PM -----

   From:       Mark Murray 
<murray at ees.nmt.edu> 


   To:         Lind S Gee 
<lgee at usgs.gov> 


   Cc:         Mark Murray 
<murray at ees.nmt.edu> 


   Date:       05/22/2009 12:17 
PM 


   Subject:    Re: APS Differential GPS Observations 
(Revised)


Hi Lind,

I've attached a skyplot for a recent day at P034 (aka ANMO). I can't
get the processing and plotting software working together properly to
plot results below 10 degrees, but this should give a pretty good idea
of the usable data coverage.

The red lines show the satellite tracks for a 24 hour period. The sky
visibility is excellent at this site - I think you can make out the
hills popping up above 10 degrees elevation around 130-150 degrees
azimuth. The hole to the north is due to the satellite geometry - the
size of the hole depends on latitude (in Alaska it will go all the way
up to the zenith direction in the middle of the circle). The green and
yellow areas showing phase residuals plotted along the tracks. As one
might expect, the residuals are greater near the horizon where the
multipath and atmospheric effects are more significant. But I don't
see anything too unusual here...

I am not aware of any online services that generate this sort of plot.
I did a quick search for software packages. The wikipedia entry on GPS
dilution of precision (DOP) does have a link to a Trimble planning
software package that runs on PCs. I have not tried it, but it looks
like its free and outputs lots information about the satellite
geometry that would be good for planning purposes. The strength of the
satellite geometry does vary over the 24 hour period, so depending on
local obstructions one could try to find an optimal time to make
observations. But in general, with few obstructions, its not usually
necessary to do this sort of planning.

A few other comments about Kyle's report:

1) Cloud cover and other weather-related phenomena do not
significantly affect GPS observations (except for lightning strikes
and snow covering the antenna!). GPS signals are sensitive to water
vapor in the atmosphere, which causes a slight delay in propagation
time. This would mostly change the estimated height of the instrument,
which we can usually estimate anyways. For this application, the
propagation paths to both antennas are nearly identical and so any
effects on the differential baseline would cancel out.

2) WAAS (Wide Area Augmentation System) is used for real-time
navigation, where position corrections are sent to a satellite and
redistributed over a wide area. This helps improve real-time
positioning to about the 1-m level - good for aircraft navigation.
However, I don't think it would have any effect on an azimuth
determination from this system, just on its estimate of the absolute
position.

3) I did a some back-of-the-envelope calculations on precision. I
recall some results from a short (~50m) baseline at Pinyon Flat that
suggested that GPS had a system noise level of about 0.1 mm in the
horizontal. I would expect this system using an the even shorter
baseline (500 mm?) to be able to achieve close to this level,
depending on the satellite geometry and length of observations and
antenna design. 0.1 in 500 mm translates to about 0.01 degrees. Given
that the limiting factors might bump it up an order of magnitude, the
reported accuracies seem reasonable to me.

Cheers,

Mark

For a sure response send E-Mail to gsnmaint at usgs.gov, or 
gsn-xxxx(your station code)@usgs.gov

Regards,

Kyle E. Persefield
Honeywell Int'l HTSI, Field Engineer
USGS Albuquerque Seismological Laboratory Contractor

Observatory "Trouble" Hotline Telephone: 505-846-5646
Tel:   505-846-7597
Fax:  505-846-6973
Web:  http://aslwww.cr.usgs.gov/

Mailing Address:
U S Geological Survey
Albuquerque Seismological Laboratory
P O Box 82010
Albuquerque, NM 87198-2010
-------------- next part --------------
A non-text attachment was scrubbed...
Name: APS Manual-1st-Edition-v11.pdf
Type: application/pdf
Size: 1298748 bytes
Desc: not available
URL: <http://geohazards.usgs.gov/pipermail/anss-netops/attachments/20120910/476120fb/attachment-0001.pdf>