[Qgis-user] wishing for accurate lattitude/longitude from, a cell phone

[Garth Fletcher]

I gathered some field data using my Garmin eTrex 20.

It was placed in a fixed location in a small field and 17 readings were
taken (using Mark Waypoint) at various intervals between 11 AM and 10 PM
(EST).  The Garmin generally estimated the error band at 3 to 3.6 m. The
Garmin was receiving both US GPS and Russian GLONASS satellites, plus
WAIS (as indicated by a "D" in the Garmin's satellite signal bars).

The same location was accurately measured with a dual-freq iGS3 by
recording during the first ~ 6 hours and post-processing via CSRS-PPP
which resolved to a 95% probability error ellipse of 2.4 cm x 1.2 cm.

The CSRS report indicates that on average 10 to 11 satellites were being
tracked (range of 8 to 13).  The Garmin's display indicated a similar
number of satellites being tracked.

The Garmin waypoints were plotted on a 1 meter grid, along with the
precisely determined location. This plot can be downloaded from

<http://www.Mason-NH.org/Specials/GARMIN_TESTS.pdf>  (21 KB).

Worst case Garmin errors are > 6 m, close to twice the Garmin's
estimates - see above linked plot.

I am not convinced that averaging any couple such readings can do much
to increase the confidence level...

My conclusion is that if precision better than ± dozens of feet is
needed for a number of points (such as in an archeological dig), simple
hand-held instruments such as the Garmin (or cell phones) are not
suitable.

Using base/rover pairs, such as discussed earlier, is a much more
effective solution, albeit considerably more costly (~ 1500 USD). They
have the advantage of rapidly providing precise differential
measurements - which are what are all that is needed to map a site.

High position accuracy can be achieved with a single dual-frequency GNSS
receiver, such as the iGS3 I use, but at the cost of long duration
observations at each point.  In my experience an observation time of at
least 30 minutes is needed (in my wooded environment) to achieve ± 1
meter, and of 1 or more hours to achieve sub-meter accuracy in absolute
positional accuracy (as opposed to relative positional accuracy).
-- 
Garth Fletcher