GPS Error: Vertical vs. Horizontal

[Marco Leon]

When we look at a GPS error value in a so-equipped GPS, does that value
reflect *both* the vertical (i.e. altitude) and the horizontal
component? Can a separate value for each be determined?

The reason I ask is that in firing up my Garmin 496 (an early Christmas
present, comments to follow in another post :)), I noticed the altitude
initially register at 100 ft below actual then right itself after a few
minutes. All of this happened while the horizontal location did not
noticeably move (had it fully-zoom by the way) which led me to believe
that the error can be different from each other at least initially.

Marco

[Gig 601XL Builder]

"Marco Leon" > wrote in message
ups.com...
> When we look at a GPS error value in a so-equipped GPS, does that value
> reflect *both* the vertical (i.e. altitude) and the horizontal
> component? Can a separate value for each be determined?
>
> The reason I ask is that in firing up my Garmin 496 (an early Christmas
> present, comments to follow in another post :)), I noticed the altitude
> initially register at 100 ft below actual then right itself after a few
> minutes. All of this happened while the horizontal location did not
> noticeably move (had it fully-zoom by the way) which led me to believe
> that the error can be different from each other at least initially.
>
> Marco
>

The two errors are separate. So if they spec 100 ft +/- it could be 100ft in
the horizontal plane and 100ft in the vertical plane.

[Mxsmanic]

Marco Leon writes:

> When we look at a GPS error value in a so-equipped GPS, does that value
> reflect *both* the vertical (i.e. altitude) and the horizontal
> component?

No. The vertical error is usually much greater than the horizontal
error. The displayed error, if there is only one, will generally be
the horizontal error.

> Can a separate value for each be determined?

Yes, but it's only practical for the computer to do it.

> The reason I ask is that in firing up my Garmin 496 (an early Christmas
> present, comments to follow in another post :)), I noticed the altitude
> initially register at 100 ft below actual then right itself after a few
> minutes. All of this happened while the horizontal location did not
> noticeably move (had it fully-zoom by the way) which led me to believe
> that the error can be different from each other at least initially.

Initially and continually. GPS altitudes should be treated as general
indications, not as precise altitudes that can be substituted for
altimeter readings.

--
Transpose mxsmanic and gmail to reach me by e-mail.

[Travis Marlatte]

"Marco Leon" > wrote in message
ups.com...
> When we look at a GPS error value in a so-equipped GPS, does that value
> reflect *both* the vertical (i.e. altitude) and the horizontal
> component? Can a separate value for each be determined?
>
> The reason I ask is that in firing up my Garmin 496 (an early Christmas
> present, comments to follow in another post :)), I noticed the altitude
> initially register at 100 ft below actual then right itself after a few
> minutes. All of this happened while the horizontal location did not
> noticeably move (had it fully-zoom by the way) which led me to believe
> that the error can be different from each other at least initially.
>
> Marco
>

They have separate error characteristics and separate error indicators -
although they are dependent. Typically, if a GPS unit displays any kind of
error value, it is for the horizontal component simply because that is what
most users will want to see. It is most likely an interpretation of the
statistical error in units of meters or feet. The statistical error is
called dilution of position and most GPS receivers determine an HDOP, a
VDOP, and a couple of others.


--
-------------------------------
Travis
Lake N3094P
PWK

[Ron Lee]

"Travis Marlatte" > wrote:

>"Marco Leon" > wrote in message
ups.com...
>> When we look at a GPS error value in a so-equipped GPS, does that value
>> reflect *both* the vertical (i.e. altitude) and the horizontal
>> component? Can a separate value for each be determined?
>>
>> The reason I ask is that in firing up my Garmin 496 (an early Christmas
>> present, comments to follow in another post :)), I noticed the altitude
>> initially register at 100 ft below actual then right itself after a few
>> minutes. All of this happened while the horizontal location did not
>> noticeably move (had it fully-zoom by the way) which led me to believe
>> that the error can be different from each other at least initially.
>>
>> Marco
>>
>
>They have separate error characteristics and separate error indicators -
>although they are dependent. Typically, if a GPS unit displays any kind of
>error value, it is for the horizontal component simply because that is what
>most users will want to see. It is most likely an interpretation of the
>statistical error in units of meters or feet. The statistical error is
>called dilution of position and most GPS receivers determine an HDOP, a
>VDOP, and a couple of others.
---------------------
>Travis
>Lake N3094P
>PWK

Actually the DOPs are related to satellite geometry and are unitless.
It is only when you multiply the DOP by the signal/user equipment
error that you get some sort of error in feet/meters.

Ron Lee

[Travis Marlatte]

"Ron Lee" > wrote in message
...
> "Travis Marlatte" > wrote:
>
>>"Marco Leon" > wrote in message
ups.com...
>>> When we look at a GPS error value in a so-equipped GPS, does that value
>>> reflect *both* the vertical (i.e. altitude) and the horizontal
>>> component? Can a separate value for each be determined?
>>>
>>> The reason I ask is that in firing up my Garmin 496 (an early Christmas
>>> present, comments to follow in another post :)), I noticed the altitude
>>> initially register at 100 ft below actual then right itself after a few
>>> minutes. All of this happened while the horizontal location did not
>>> noticeably move (had it fully-zoom by the way) which led me to believe
>>> that the error can be different from each other at least initially.
>>>
>>> Marco
>>>
>>
>>They have separate error characteristics and separate error indicators -
>>although they are dependent. Typically, if a GPS unit displays any kind of
>>error value, it is for the horizontal component simply because that is
>>what
>>most users will want to see. It is most likely an interpretation of the
>>statistical error in units of meters or feet. The statistical error is
>>called dilution of position and most GPS receivers determine an HDOP, a
>>VDOP, and a couple of others.
> ---------------------
>>Travis
>>Lake N3094P
>>PWK
>
> Actually the DOPs are related to satellite geometry and are unitless.
> It is only when you multiply the DOP by the signal/user equipment
> error that you get some sort of error in feet/meters.
>
> Ron Lee

The DOPs represent a statistical confidence. They do have a unit but you're
right, it's not in units of distance and it is not standardized.

Does the 496 display a value for the GPS error? Or, was it a hypothetical
question. If a GPS receiver does indicate the accuracy with something like
"+/- 5 meters" then it is really expressing a confidence and is more
correctly stated as "+/- 5 meters, 95% of the time" Or whatever confidence
they think looks the best to their consumer.

--
-------------------------------
Travis
Lake N3094P
PWK

[peter]

Travis Marlatte wrote:
> "Ron Lee" > wrote in message
> ...

> > Actually the DOPs are related to satellite geometry and are unitless.
> > It is only when you multiply the DOP by the signal/user equipment
> > error that you get some sort of error in feet/meters.

> The DOPs represent a statistical confidence. They do have a unit but you're
> right, it's not in units of distance and it is not standardized.

No, Ron is correct that the various DOPs (PDOP, VDOP, HDOP, TDOP, etc.)
are unit-less measures based on the satellite geometry. They are
well-defined and standardized as specified in Section 3.1.3 of the
Navstar User Equipment guide:
http://www.navcen.uscg.gov/pubs/gps/gpsuser/gpsuser.pdf

It's when the DOP, typically the HDOP, is translated into a
corresponding EPE (Estimated Position Error) or accuracy value that it
is converted into a measure with units of distance and these measures
are not standardized across manufacturers or even sometimes between
models of a given manufacturer.

I agree with some previous comments that the VDOP is generally somewhat
larger than HDOP due to the inherent geometry of the situation. You
usually have satellites on all sides of your position, but you can only
get signals from satellites above you, none below, and that reduces the
accuracy of the altitude calculation. An additional consideration in
the OP's case is that when the unit first gets enough satellite signals
to report a position it may only have enough for a '2D' calculation.
With three satellite signals, the unit can only determine the exact
time and two space coordinates, so it assumes the altitude based on the
last value previously seen and only calculates the lat. and long.
coordinates. Then when it gets one or more additional satellite
signals it can switch to '3D' mode and calculate the actual altitude as
well.

[Thomas Borchert]

Marco,

> When we look at a GPS error value in a so-equipped GPS, does that value
> reflect *both* the vertical (i.e. altitude) and the horizontal
> component? Can a separate value for each be determined?
>

The error values commonly given are:

VDOP: Vertical Dilution of Position
HDOP: Horizontal Dilution of Position
PDOP: 3-D dilution of position, i.e. the above two combined

Most GPS receivers show these values.

Google and Wikipedia are your friends...

--
Thomas Borchert (EDDH)

[[email protected]]

On 13 Dec 2006 11:40:00 -0800, "Marco Leon" > wrote:

>When we look at a GPS error value in a so-equipped GPS, does that value
>reflect *both* the vertical (i.e. altitude) and the horizontal
>component? Can a separate value for each be determined?
>
>The reason I ask is that in firing up my Garmin 496 (an early Christmas
>present, comments to follow in another post :)), I noticed the altitude
>initially register at 100 ft below actual then right itself after a few
>minutes. All of this happened while the horizontal location did not
>noticeably move (had it fully-zoom by the way) which led me to believe
>that the error can be different from each other at least initially.
>
>Marco

You can try to improve vertical accuracy using calculations from one
of the following websites but you need to know if your GPS is already
correcting or not!

http://sps.unavco.org/geoid/

http://earth-info.nga.mil/GandG/wgs84/gravitymod/wgs84_180/intptW.html?

[peter]

wrote:

> You can try to improve vertical accuracy using calculations from one
> of the following websites but you need to know if your GPS is already
> correcting or not!

AFAIK, all the Garmin portable GPS units include geoid vs. ellipsoid
corrections.
Set the unit to output NMEA data from its serial port and connect it to
a PC running Hyperterminal or similar program that captures the output
stream. The $GPGGA sentence will show the value being used by the GPS
for the geoid-ellipsoid correction as the 8th data field.
>
> http://sps.unavco.org/geoid/
>
> http://earth-info.nga.mil/GandG/wgs84/gravitymod/wgs84_180/intptW.html?

[Travis Marlatte]

"peter" > wrote in message
oups.com...
> Travis Marlatte wrote:
>> "Ron Lee" > wrote in message
>> ...
>
>> > Actually the DOPs are related to satellite geometry and are unitless.
>> > It is only when you multiply the DOP by the signal/user equipment
>> > error that you get some sort of error in feet/meters.
>
>> The DOPs represent a statistical confidence. They do have a unit but
>> you're
>> right, it's not in units of distance and it is not standardized.
>
> No, Ron is correct that the various DOPs (PDOP, VDOP, HDOP, TDOP, etc.)
> are unit-less measures based on the satellite geometry. They are
> well-defined and standardized as specified in Section 3.1.3 of the
> Navstar User Equipment guide:
> http://www.navcen.uscg.gov/pubs/gps/gpsuser/gpsuser.pdf


I agree. I overstated my point which was only that the DOPs did not
represent something as simple as an error radius around the true position.
-------------------------------
Travis
Lake N3094P
PWK