Page 1 of 2   1 2 > last >>
Bookmark this page: Add Zumo accuracy question issue to Yahoo MyWeb Add Zumo accuracy question issue to Google Bookmarks Add Zumo accuracy question issue to Windows Live Add Zumo accuracy question issue to Del.icio.us Digg Zumo accuracy question issue! Add Zumo accuracy question issue to Netscape
  •  
  • Subject
  • Author
  • Date
If you were  Registered and logged in, you could reply and use other advanced thread options
Posted by Ray in SF on June 4, 2008, 3:37 am


Hi all. I just bought my first GPS unit - a Zumo 550. Being new to the
use of a GPS, I'm not sure what acceptable accuracy tolerances are.
My unit says it's between 18-20' accurate (on the tools page) however
whenever I come up to an intersection where I have to make a turn, it
consistently says to turn in 40-50' when I'm right at the edge of the
intersection. Taking into account the GPS's self-generated 18-20'
accuracy reading and that the position of the road is measured from
the center of the width of the road, perhaps 12' for a two lane road,
that would explain a discrepancy of up to 32'. Any idea why I would
consistently get a discrepancy of an additional 10-18' over the
theoretical maximum error?

Enabling/disabling WAAS makes a negligible difference. I've installed
the latest firmware and maps (2009 edition). Is my unit faulty?

Thanks,
Ray in San Francisco

Posted by James Robinson on June 4, 2008, 8:16 am



> My unit says it's between 18-20' accurate (on the tools page)

The discussion of GPS accuracy is a complicated one, which can't be
effectively addressed in short Usenet posts.

Without looking at how they calculate it, the 18 to 20 foot accuracy is
likely based on a 50 percentile measurement, meaning that 50 percent of
the readings will be within that distance, or conversely, 50 percent
will fall farther from that distance, in some cases significantly
farther.

The measurements are also typically taken with the unit sitting in one
position, and when you are moving, the accuracy will be less.

> however whenever I come up to an intersection where I have to make a
> turn, it consistently says to turn in 40-50' when I'm right at the
> edge of the intersection. Taking into account the GPS's self-generated
> 18-20' accuracy reading and that the position of the road is measured
> from the center of the width of the road, perhaps 12' for a two lane
> road, that would explain a discrepancy of up to 32'. Any idea why I
> would consistently get a discrepancy of an additional 10-18' over the
> theoretical maximum error?

There is another source of error that you haven't taken into account: the
accuracy of the maps stored in the unit. Many times my unit shows me
driving on the grass beside roads, particularly in curves or on winding
roads. The program logic will often "snap" to the nearest road, so you
won't see the error, even though it is still there, having been corrected
by the software.

> Enabling/disabling WAAS makes a negligible difference. I've installed
> the latest firmware and maps (2009 edition). Is my unit faulty?

Your unit is probably functioning as well as you should expect for a
consumer GPS. You just have to learn to live with it.

Posted by Ray in SF on June 4, 2008, 2:05 pm


Thanks for the info. Just FYI, this accuracy issue is when I'm at a
dead stop at the intersection. I understand there's additional display
latency while I'm moving. Also, if it was problems with the map data,
wouldn't the inaccuracy always be towards a certain direction? It
doesn't matter from what direction I approach a turn and it doesn't
matter which intersection I'm at. It's always off by the same amount
always assuming the turn is about 50' farther in the direction of
movement than it is. I realize that a consumer GPS likely is not meant
to have exceptional accuracy, but it's a bit disappointing when I read
so many other posts of other GPS units being accurate to within 10-15'
feet. If I'm approaching an intersection AND I'm in motion, then it
really sucks to see the unit says to turn sometimes in as much as 100'
when in fact the turn is immediate. Being on a motorcycle and driving
in an area I'm not familiar with, I've missed many a turn-off because
of this issue. Well, I bought it from Amazon and I'm going to exchange
it as defective. I've got nothing to lose. If the next unit is same
then I guess I will just have to live with it.

> > My unit says it's between 18-20' accurate (on the tools page)
> The discussion of GPS accuracy is a complicated one, which can't be
> effectively addressed in short Usenet posts.
> Without looking at how they calculate it, the 18 to 20 foot accuracy is
> likely based on a 50 percentile measurement, meaning that 50 percent of
> the readings will be within that distance, or conversely, 50 percent
> will fall farther from that distance, in some cases significantly
> farther.
> The measurements are also typically taken with the unit sitting in one
> position, and when you are moving, the accuracy will be less.
> > however whenever I come up to an intersection where I have to make a
> > turn, it consistently says to turn in 40-50' when I'm right at the
> > edge of the intersection. Taking into account the GPS's self-generated
> > 18-20' accuracy reading and that the position of the road is measured
> > from the center of the width of the road, perhaps 12' for a two lane
> > road, that would explain a discrepancy of up to 32'. Any idea why I
> > would consistently get a discrepancy of an additional 10-18' over the
> > theoretical maximum error?
> There is another source of error that you haven't taken into account: the
> accuracy of the maps stored in the unit. =A0Many times my unit shows me
> driving on the grass beside roads, particularly in curves or on winding
> roads. =A0The program logic will often "snap" to the nearest road, so you
> won't see the error, even though it is still there, having been corrected
> by the software.
> > Enabling/disabling WAAS makes a negligible difference. I've installed
> > the latest firmware and maps (2009 edition). Is my unit faulty?
> Your unit is probably functioning as well as you should expect for a
> consumer GPS. =A0You just have to learn to live with it.


Posted by James Robinson on June 5, 2008, 9:40 am



> I realize that a consumer GPS likely is not meant
> to have exceptional accuracy, but it's a bit disappointing when I read
> so many other posts of other GPS units being accurate to within 10-15'
> feet. If I'm approaching an intersection AND I'm in motion, then it
> really sucks to see the unit says to turn sometimes in as much as 100'
> when in fact the turn is immediate. Being on a motorcycle and driving
> in an area I'm not familiar with, I've missed many a turn-off because
> of this issue. Well, I bought it from Amazon and I'm going to exchange
> it as defective. I've got nothing to lose. If the next unit is same
> then I guess I will just have to live with it.

No inexpensive consumer GPS is accurate to within 10 or 15 feet. There are
too many errors, from radio propagation to those in the electronics, that
add up to more error than that. A more likely accuracy would be in the
order of 50 feet, and on occasion, maybe 100 feet. You should expect less
accuracy when moving in areas where the radio signal can bounce around, or
where the view of some of the satellites is obscured, like around heavy
forests, tall buildings or close to mountains.

Posted by Simon Slavin on June 6, 2008, 4:06 pm


On 04/06/2008, Ray in SF wrote in message <b9b5b9ac-59d7-45fa-9ae6-
5d0ef3726076@e53g2000hsa.googlegroups.com>:

> Just FYI, this accuracy issue is when I'm at a
> dead stop at the intersection.

Contact Garmin and ask them whether this is expected of the model you
bought. We are not Garmin. We can't talk for them.

Simon.
--
http://www.hearsay.demon.co.uk

Page 1 of 2   1 2 > last >>