Looking for the correct term about accuracy on my Garmin Geko 201

Hello.

I have just purchased a Garmin Geko 201, which when it has a
position-fix shows accuracy in "meters".

Being used to having only accuracy shown in GDOP, i was a bit mystified
as how it shows accuracy in meter presentation.

I have a book about GPS (written on my native language), that has a
sentence which roughly translates into the following:

"With discount GPS-recievers, one is capable of real-time positioning
with a spread (accuacy) of about 6 meters. Accuracy represented by
spreading, means that 68.3% of the readings will be within the shown
spreading, 95.4% of the readings will be within the shown spreading and
99.7% of the readings within three times the spreading."

In other words if my GPS-reciever shows a accuracy of 6 meters, 68.3%
of the readings is in a circle with a radius of 6 meter from the
recievers current locate, 95.4% of the readings in a circle with a 12
meter radius and so forth.

Now - using google groups i have tried to find out the exact way this
Garmin reciever makes its "guestimate" of the accuarcy, and i'm pretty
sure that it has something to do with GDOP (Geometric Dilution of
Position), DRMS ([Error] Distance RMS), UERE (User Equivalent Range
Error ) and/or URA (User Range Accuracy), but i haven't seen the light
as to exactly how the Garmin does it's estimate :(

Furthermore (than just searching through this fine group) i have also
looked at http://www.gpsinformation.org/dale/wgarmin.htm and Sam
Wormly's page http://edu-observatory.org/gps/gps_accuracy.html - and
still haven't seen the bright light.

So here's my questions:
What is the english technical term of "spreading"?
How is this accuracy thing calculated/estimated?

Regards
Soren

smoelfster@gmail.com wrote:


"Standard deviation". If readings are normally distributed (that is, the
randomness follows a Gaussian or normal distribution pattern), 68.3%
will be within one standard deviation (usually represented by the Greek
letter sigma), 95.4 will be within two sigma, and 99.7% within three
sigma of the mean value. The standard deviation is a measure of the
width of the probability distribution curve.


I'll have to leave that answer to others.

Roy Lewallen

Roy Lewallen wrote:

[snip]


Hi Roy.

Thanks for your prompt answer. This surely gave me a lot more to read
now :)

/Soren

smoelfster@gmail.com wrote:


[snip]


[snip]


I have found out, that this representation of accuracy is not only done
by Garmin, but also by other GPS-recievers and some PDA-software...

I still haven't figured out _how_ excatly this is done though :(

/Soren

smoelfster@gmail.com wrote:


You are correct that many GPS receivers do this to diplsay the number
for the user. No one provides information on the algorithim and exactly
how it works. Garmin considers this a secret and will not reveal how
they do it.

That being said, you are correct in your assumptions of DOP and UERE
numbers being the gateing data for the calculation. It is, of course,
only an estimate.

Dale

--
_ _ Dale DePriest
/`) _ // http://users.cwnet.com/dalede
o/_/ (_(_X_(` For GPS and GPS/PDAs