help understanding GPS accuracy specs

The best accuracy spec I have seen published for a consumer GPS receiver
is for a "mouse" type device. The Holux GM-210 using the SiRF Star II or
SiRF Star IIe/LP chipset, claims 2.2 meter accuracy when receiving WAAS
signals: http://www.holux.com/product/gpsreceiver_m...vel=grandsonson

EGNOS/WAAS/Beacon Position
< 2.2m, horizontal 95% of time
< 5 m, vertical 95% of time

Advertised accuracy for the newer models with the SiRF Star III chipset
seems to be lower. They claim 5 m or 10m horizontal position. The older
(pre-SS3) Garmins like the GPS 60 advertise 10 feet w/WAAS which is
about 3 meters. The newer GPSMAP 60Cx advertises 3-5 meters w/WAAS.

Is this a difference in advertising hype or a real difference? Is the
WAAS 2.2m / 95% spec. from Holux reliable? It seems possible, given the
FAA page that claims WAAS can deliver 1-2 m accuracy for ground vehicle
navigation, not just aircraft http://gps.faa.gov/FAQ/faq-waas.htm#34 but
if that's real, why can't the other manufacturers do as well as Holux?

Are there any independent tests of consumer GPS receiver accuracy under
specified conditions?

corrected Holux accuracy link is
http://www.holux.com/product/search.htm?filename=gpsreceiver_mouse_gr213_gg_wzhjzhd.htm&target=gpsreceiver10&level=grandsonson

user@domain.invalid wrote:

user@domain.invalid wrote:


Antenna design, RF section noise figure, correlation processing
algorithms can all be factors in receiver/error accuracy. I have
equipment that can approach an accuracy of < 0.6 m, horizontal 95%
of time, using the same GPS L1 C/A and WAAS signals. But it cost
a lot more than a consumer handheld does.

There are two encouraging trends
o better receiver designs continue to be developed
o the GPS as a whole is becoming more accurate

user@domain.invalid writes

David Wilson has done some excellent work on GPS accuracy:

http://users.erols.com/dlwilson/gps.html

His WAAS page is here but the whole site is well worth studying in
detail.

http://users.erols.com/dlwilson/gpswaas.htm

--

Dominic Sexton

Dominic Sexton wrote:

The biggest remaining error source in a WAAS solution that is
processing both ionosphere and satellite corrrections is multipath.

A narrow correlator receiver can help, longer carrier smoothing times
using codeless L2 can help, or a benign multipath enviorement can yield
good results. You can get a 80 cm to 90 cm 95% horizontal error level
of accuracy with WAAS with a narrow correlator receiver in a benign
enviroment or with a narrow correlator using codeless L2 to have long
smoothing times. Multipath errors are usually very close to zero mean
over time so time averaging can also beat down the error. Its not hard
to see how a manufacture can claim 2 meters of accuracy, but without
knowing how it was measured it don't expect to always get that accuracy
at your location.