[Manifold-l] GPS Accurracy

Mon Oct 2 05:59:53 CDT 2006

Colin,

I am no GPS expert either but I have taken a layman's interest in it.  My
information is very out-of-date now.

Unfortunately your experience may not always work - some areas/times in the
world the geometry of the satellites will be such that for a short time the
accuracy may not be as good as usual.

DOP is a theoretical calculation of the Geometrical/Time error, not the
accuracy.  This was probably more interesting during the development stage
of GPS as they moved the satellites around a lot. Now there should be even
coverage it may not be such an issue, but some parts of the world may be
affected for periods of minutes. 

DOP may be interesting if some of the satellites are blocked from view so
your viewable sat constellation is not so good - I noticed a big positional
error when fixing at a window where all the viewable satellites were in the
South West.

To assess the accuracy of the satellite position in space automatically, one
apparently needs the UERE/URA for each satellite.  This is updated once per
pass and broadcast with the transmission.  Combine this with your DOP to
assess maximum accuracy capability.

What a receiver receives and calculates is another thing altogether -
nothing I know of can help here.  

Ranges to satellites can be affected by other varying factors, so a
confidence level is normally be quoted with any real-time position
capability. This is statistical in nature so any single point should be
regarded as meaningless within this parameter (eg 25m 2D RMS means that 50%
of the time you would be within 25m.  Some of these errors drift with time -
I walked around my cricket square outside my house: everything looked good
and straight and nearly square but I was surprised to see the square didn't
close by a long way (before WAAS/EGNOS).  The impression was 1-2m errors
point to point but as a set over time they drifted away. No idea what the
absolute position was like.

Differential positioning corrects these latter effects down to 3m or less.

Finally, there are non-predictable/non-assessable errors. 

The worst I think is multi-path where some object nearby reflects the
transmission introducing phase and distance errors.  Having a metal shovel
in your back pack is, I found, a lousy idea.  Positioning of the antenna is
important on boats for this reason.  The errors as far as I can tell can be
miles!  In town buildings used to be and may still be a problem.

Interference causes problems too - any conductor with a separation equal to
the transmission wave length (Ghz range) will cause a Young's slit problem
and trees, apparently, are the best for this.  Generally you loose fix here.

So why is it so good?  Because 99.9% of the time it is good enough - one
reading out by a couple of 10s of meters is obviously wrong - a whole set
and the antenna was next to your hip flask all day (shame about the track
you wanted to plot) - Just DON'T BET YOUR LIFE ON IT!!!!  Check it with a
compass/map even MANIFOLD!

The fix flag was/is important - some GPS units used to go into
dead-reckoning mode if they lost a fix and continue pumping out GPS
positions that looked really good but were completely wrong - in these cases
they were accompanied by a "D" flag to signify that the transmission is lost
- somebody on deck just kicked the cable out of the antenna!

Francis

-----Original Message-----
From: manifold-l-bounces at lists.directionsmag.com
[mailto:manifold-l-bounces at lists.directionsmag.com] On Behalf Of Colin
Driscoll
Sent: 29 September 2006 11:32
To: 'Martin Roseveare'; 'Dimitri Rotow'; manifold-l at lists.directionsmag.com
Subject: RE: [Manifold-l] GPS Accurracy

My .02 worth. As the saying goes about the proof of the pudding....I spend
2-3 full days a week in the field with my Garmin 60cs connected (USB/NMEA)
to Manifold on my laptop. With a good quality orthorectified aerial image
under the GPS tracks and good satellite reception (high gain external
antenna) the track line is drawn where I can see I am out the vehicle
window. If I am on the left of a bush road that's where the line is. If I
drive back along a narrow road I have already been on the return line is on
top of the first- in fact I put a wish list item in a while ago to have a
blinking point at the current location because it is generally not easy to
know this on a return trip.

This is all +/- about 3meters but is cetainly reproducible here in Australia
at least- maybe it's the clear sky.

-----Original Message-----
From: manifold-l-bounces at lists.directionsmag.com
[mailto:manifold-l-bounces at lists.directionsmag.com] On Behalf Of Martin
Roseveare
Sent: Friday, 29 September 2006 7:31 PM
To: Dimitri Rotow; manifold-l at lists.directionsmag.com
Subject: Re: [Manifold-l] GPS Accurracy

I guess you're right Dimitri!

Martin Roseveare
Senior Geophysicist

Phone: +44 (0) 1989 730 564
Fax: +44 (0) 7050 369 790
Web: www.archaeophysica.co.uk

This e-mail and any attachment may contain proprietary or confidential
material and be subject to legal privilege. If you have received it in error
please promptly erase it from your system and inform the sender. It should
not be copied, disclosed to, retained or used in any way by any other party.

Nothing in this e-mail should be regarded as representing a binding contract
unless explicitly stated otherwise. All opinions are those of the sender and
not necessarily ArchaeoPhysica Ltd.
----- Original Message -----
From: "Dimitri Rotow" <dar at manifold.net>
To: <manifold-l at lists.directionsmag.com>
Sent: Friday, September 29, 2006 1:30 AM
Subject: RE: [Manifold-l] GPS Accurracy

>
>> Looking through the NMEA web page you referenced confirms that they 
>> do differ.
>>
>> For example GSA has '3' for 3D fix, GGA has '3' for PPS fix
>
> In that link the particular value 3 for PPS fix is cited for the GGA 
> sentence is noted as "Fix quality" whilst in the GSA sentence the 
> value 3 for 3D fix is noted as "3D fix".  These are different 
> attributes, so that the same value "3" occurring in those two 
> different sentences is not assigning two different meanings to the same
thing, it is two different
> attributes occuring in two different sentences.   In this particular case,
> both attributes use integer-coded values where the one attribute uses 
> integers in the range 0 to 8 and the other uses integers in the range 1 to
> 3.   It's just happenstance that in the particular example given an 
> integer
> value of 3 is used in both cases, but it means different things for 
> the two different attributes.
>
> As far as I've been able to tell, a "quality" attribute in a sentence 
> is something that occurs in the GGA sentence only and is a value from 
> 0 to 8 inclusive.  Other attributes such as "3D fix," the various 
> "DOP"s and so on
> appear to be attributes that are different from "quality."   In the web
> crawls I've done I have not located a "quality" attribute except 
> within the GGA sentence.
>
>> and both of these are common sentences. Hence my question about which 
>> Manifold uses when in receipt of more than one sentence per message! 
>> I guess it comes down to the purpose of each type of NMEA sentence, 
>> i.e., rather than 'quality' being an absolute thing it is a purpose 
>> specific indicator only.
>>
>
> I believe "fix quality" has a specific meaning in NMEA-speak as being 
> a specific attribute that's one of those 9 values.  "Quality" in this 
> context is a specific technical term setting forth one of those 9 
> coded meanings and is not a generic English word such as a synonym for 
> circular error probability, etc.
>
> But then, I'm not an NMEA or GPS guy so the above is just my inference 
> from reading what I've been able to lay hands on via web searches...  
> perhaps someone with an expert understanding of NMEA could comment on 
> that inference.
>
> Cheers,
>
> Dimitri
>
> _______________________________________________
> Manifold-l mailing list
> Manifold-l at lists.directionsmag.com
> http://www.directionsmag.com/mailman/listinfo/manifold-l

_______________________________________________
Manifold-l mailing list
Manifold-l at lists.directionsmag.com
http://www.directionsmag.com/mailman/listinfo/manifold-l

_______________________________________________
Manifold-l mailing list
Manifold-l at lists.directionsmag.com
http://www.directionsmag.com/mailman/listinfo/manifold-l