Incorrect Altitudes

[Isaac]

Hello all,

I have been trying to use the OsmAnd+ app on several of my android devices and am getting a little frustrated with the altitude readings varying between devices. I use the app with multiple devices and also have a Garmin Glo bluetooth receiver that I use for increased accuracy. To give an example of the variations, here are the altitudes the Osmand app will display on different devices:

Google Earth (for reference) - 820'

Motorola Xoom internal GPS - 814'

Motorola Xoom using Garmin Glo - 942'

Moto Z Play internal GPS - 715'

Moto Z play using Garmin Glo - 827'

The Garmin Glo reads an altitude of 829' on both devices via the "Bluetooth GPS" app that I use to deliver a mock GPS location.

Is there any way that I can force the Osmand altitude to match my Garmin reading? It seems to be very close on the Z play, but pretty far off with the Xoom for some reason.

[Harry van der Wolf]

Did you also download the "worldwide altitude correction" map in OsmAnd?

Your Garmin is maybe internally corrected for that, butOsmAnd does a "clean" calculation based on the GPS info, and uses that map for altitude correction.

--
You received this message because you are subscribed to the Google Groups "Osmand" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Poutnik]

It may be issue of NO -vs CORRECT vs DOUBLE correction.

As 2 values are about correct and the other 2 values
are having offset about +120'  -120'

Dne 17/02/2018 v 19:05 Harry van der Wolf napsal(a):

Did you also download the "worldwide altitude correction" map in OsmAnd?

Your Garmin is maybe internally corrected for that, butOsmAnd does a "clean" calculation based on the GPS info, and uses that map for altitude correction.

--
You received this message because you are subscribed to the Google Groups "Osmand" group.

To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

-- 
Poutnik ( The Wanderer )

My Brouter profiles 
https://github.com/poutnikl/Brouter-profiles/wiki

[Harry van der Wolf]

What I mean is that the GPS calculations are based on a globe. We all know that the earth is not a perfect globe, but a flattened globe.

That is reason one for the altitide correction.

Secondl reason: when using it on this "perfect globe", we also have mountains or higher plains, which again require an altitude correction.

For a quite simple tutorial see here: https://www.unavco.org/education/resources/tutorials-and-handouts/tutorials/elevation-and-geoid.html

Harry

2018-02-17 19:09 GMT+01:00 Poutnik <poutni...@gmail.com>:

It may be issue of NO -vs CORRECT vs DOUBLE correction.

As 2 values are about correct and the other 2 values
are having offset about +120'  -120'

Dne 17/02/2018 v 19:05 Harry van der Wolf napsal(a):

Did you also download the "worldwide altitude correction" map in OsmAnd?

Your Garmin is maybe internally corrected for that, butOsmAnd does a "clean" calculation based on the GPS info, and uses that map for altitude correction.

--
You received this message because you are subscribed to the Google Groups "Osmand" group.

To unsubscribe from this group and stop receiving emails from it, send an email to osmand+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

-- 
Poutnik ( The Wanderer )

My Brouter profiles 
https://github.com/poutnikl/Brouter-profiles/wiki

--

You received this message because you are subscribed to the Google Groups "Osmand" group.

To unsubscribe from this group and stop receiving emails from it, send an email to osmand+unsubscribe@googlegroups.com.

[Isaac]

Yes, I installed the correction map but it did not seem to make any difference with the altitude in my location. The Garmin seems to get a good altitude reading, but the Osmand app does not reflect this on the Xoom for some reason. That is what I am most interested in fixing right now.

To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

[Poutnik]

GPS used WGS84 geoid IS a flattened globe.
If it was not, corrections would be much bigger.

But Earth is more like "potatoid".
What that means that due irregularities of Earth shape and densities,
the sea level has in WGS84 geoid location dependent altitude.

The corrections are irregular.

Dne 17/02/2018 v 19:14 Harry van der Wolf napsal(a):

To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Poutnik]

See e.g.
http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm2008/egm08_wgs84.html
together with the local correction map.

Dne 17/02/2018 v 19:22 Poutnik napsal(a):

> GPS used WGS84 geoid IS a flattened globe.
> If it was not, corrections would be much bigger.
>
> But Earth is more like "potatoid".
> What that means that due irregularities of Earth shape and densities,
> the sea level has in WGS84 geoid location dependent altitude.
>
> The corrections are irregular.

[Harry van der Wolf]

Yes, you are completely right.

I wanted to keep it as simple as possible without referring to WSG84 (or other correction models), but for some experienced users that's way too simple.  :)

[Poutnik]

Dne 17/02/2018 v 18:45 Isaac napsal(a):

> Hello all,
>
> I have been trying to use the OsmAnd+ app on several of my android
> devices and am getting a little frustrated with the altitude readings
> varying between devices. I use the app with multiple devices and also
> have a Garmin Glo bluetooth receiver that I use for increased
> accuracy. To give an example of the variations, here are the altitudes
> the Osmand app will display on different devices:
>
> Google Earth (for reference) - 820'

It is better to use a topographic map or direct local altitude info.

While Europe subtracts altitude by the correction,
US adds altitude, so the higher value is the (over?)corrected one.

[Harry van der Wolf]

2018-02-17 19:41 GMT+01:00 Poutnik <poutni...@gmail.com>:

While Europe subtracts altitude by the correction,
US adds altitude, so the higher value is the (over?)corrected one.

--
Poutnik ( The Wanderer )

I didn't know that :) 

[Robert Grant]

I'm pretty sure that none of the devices listed by the op use barometric altimeters; even the Garmin Glo is GPS altimetry.  GPS is known for its lack of precision in determining altitude.  In aviation, old school barometric altimeters are still the gold standard, but they require periodic barometric pressure adjustment.  While GPS is great for navigating around the earth, it would be very foolish to use GPS altitude for landing an aircraft.  Bottom line:  don't expect any phone with only GPS altitude to agree precisely with a database supplied measurement.

Virus-free. www.avast.com

--

You received this message because you are subscribed to the Google Groups "Osmand" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Robert Grant

[Poutnik]

As being trained in past as the military meteorologist, in pre-GPS era,  I am aware of that. But the offset value is bigger than GPS accuracy of the static value averaged.   BTW, the most handy way how to calibrate the barometric altimeter at unknown altitude is the GPS device. While barometric altimeters have superior short-term accuracy and stability, GPS devices have superior long-term accuracy and stability.  Fortunately, for most personal usage, absolute altitudes are not that important, rather the relative changes.

Dne 17/02/2018 v 23:35 Robert Grant napsal(a):

I'm pretty sure that none of the devices listed by the op use barometric altimeters; even the Garmin Glo is GPS altimetry.  GPS is known for its lack of precision in determining altitude.  In aviation, old school barometric altimeters are still the gold standard, but they require periodic barometric pressure adjustment.  While GPS is great for navigating around the earth, it would be very foolish to use GPS altitude for landing an aircraft.  Bottom line:  don't expect any phone with only GPS altitude to agree precisely with a database supplied measurement.

Virus-free. www.avast.com

On Sat, Feb 17, 2018 at 1:41 PM, Harry van der Wolf <hvd...@gmail.com> wrote:

2018-02-17 19:41 GMT+01:00 Poutnik <poutni...@gmail.com>:

While Europe subtracts altitude by the correction,
US adds altitude, so the higher value is the (over?)corrected one.

I didn't know that :) 

 

-- 
Poutnik ( The Wanderer )

My Brouter profiles 
https://github.com/poutnikl/Brouter-profiles/wiki

[Robert Grant]

While my experience agrees with you regarding accuracy and stability, it's still better to know the local pressure setting, especially if landing an aircraft without a radio altimeter. Setting an altimeter based on GPS sounds quite rare to me.

--

[Poutnik]

Sure, but I was not speaking in context of aircraft, but e.g. of multi-day mountain trekking.  Neither I have heard about OSMAnd to be used for a precise aircraft altitude control.

Dne 18/02/2018 v 00:39 Robert Grant napsal(a):

[Kevin Kenny]

For what it's worth, a smartphone WITH a barometer (and an altitude correction model on board, which I think is wired into Location Services on Android) is quite a robust altitude indicator. The GPS altitude can be compared with the barometer, integrated over a very long time - relative to altitude changes, but short relative to the weather. That can yield the sea-level pressure reading that will calibrate the barometer for short-term variability.

My phone doesn't generally do quite as well as my wrist altimeter (which often nails a known elevation within 5 m if the weather is stable), but many, many times better than an unassisted GPS.

--

[Poutnik the Wanderer]

Exactly. This is about the same what I described on the LocusMap forum as possibly ideal way of progressive calibration of barometric altimeter.

Dne 18. února 2018 3:36:55 Kevin Kenny <kevin....@gmail.com> napsal:

To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

[Poutnik]

I.e. to have some exponential filter with a very long time constant, or some kind of a Kalman filter, to minimize the non-zero offset of the average difference of barometric and GPS altitude.

In an ideal case, the GPS altitude would fluctuate around the actually used barometric altitude, while the barometric altitude would very slowly drift to the centre of GPS altitude fluctuations.

Dne 18/02/2018 v 08:58 Poutnik the Wanderer napsal(a):

Exactly. This is about the same what I described on the LocusMap forum as possibly ideal way of progressive calibration of barometric altimeter.

Dne 18. února 2018 3:36:55 Kevin Kenny <kevin....@gmail.com> napsal:

For what it's worth, a smartphone WITH a barometer (and an altitude correction model on board, which I think is wired into Location Services on Android) is quite a robust altitude indicator. The GPS altitude can be compared with the barometer, integrated over a very long time - relative to altitude changes, but short relative to the weather. That can yield the sea-level pressure reading that will calibrate the barometer for short-term variability.

My phone doesn't generally do quite as well as my wrist altimeter (which often nails a known elevation within 5 m if the weather is stable), but many, many times better than an unassisted GPS.

[Greg Troxel]

Harry van der Wolf <hvd...@gmail.com> writes:

Defining terms I'm using (and being not entirely precise, ignoring a lot
that doesn't matter for this discussion)

ellipsoidal height: distance above WGS84 reference ellipsoid, as
geometrically measured by GPS. Not typically used by normal humans.

orthometric height: distance above the standard equipotential (equal
gravity) surface. What people usually call "elevation". In the US,
NAVD88 is the current vertical datum, and fits in this category.

geoid model: coefficients that vary with location to convert between
ellipsoidal height and orthometric height

> Your Garmin is maybe internally corrected for that, butOsmAnd does a
> "clean" calculation based on the GPS info, and uses that map for altitude
> correction.

I think you are saying (and extending) that

OsmAnd is getting ellipsoidal height from the internal GPS receiver
via a standard Android interface (and either not getting orthometric
height or ignoring it)

OsmAnd has a geoid model that might or might not have been downloaded
by the user

without the geoid model, it shows ellipsoidal height as "altitude"

with the geoid model, it applies it and thus shows a GPS-derived
orthometric height as altitude

Is that right? And you didn't address

do the builtin GPS receivers in phones have a geoid model, like a
standalone unit typically does? Is osmand ignoring it, or is it often
not present?

Also, I would not be at all surprised if various phones behaved
differently and that some of them did not follow the spec, returning
orthometric height labelled as ellipsoidal height.


I would recommend to Isaac to run "SatStat", which shows "elevation". I
am not quite sure if that is ellipsoidal height or orthometric height.
However, it's another datapoint.

[Isaac]

From SatStat:

Moto Z Play internal GPS: 712'

Z Play on Garmin Glo: 813'

Xoom internal GPS: 721'

Xoom on Garmin Glo: 813'

[Poutnik]

For me, it looks like both internal GPS units, in agreement with my
experience, do not provide geoid correction to the amplitude. Therefore
is is about 120' lower than the reference, what is about the correction
value for US region.

Neither they make the correction for the GG unit,  but as it seems does
correction on its own, the result is in agreement with the reference.

The original data from OSMAnd in the first post have both phone devices
mutually shifted by another 120'. That seems to me like the SW geoid
correction was ON on the device with higher data using double correction
for the GG, while on the other device it was OFF, using no correction
for the internal GPS.

Dne 18/02/2018 v 16:34 Isaac napsal(a):

> From SatStat:
>
> Moto Z Play internal GPS: 712'
>
> Z Play on Garmin Glo: 813'
>
> Xoom internal GPS: 721'
>
> Xoom on Garmin Glo: 813'

[Poutnik]

After some quick check via an Excel table, I evaluated atmospheric influences on barometric altimeter :

1/ The biggest concern is for the air pressure trend. The difference 2 hPa makes altitude error about 17-18 metres. The pressure change about 2 hPa/h is easily and steadily happening during 12-18 hours before advent of the warm front. A static altimeter would report you are climbing 18 metres in 1 hour.  Near a cold front, pressure trough or ridge, the air pressure can change even faster. Once can easily check this effect during a day trip with the same start and destination. Difference can be tens of meters.
2/ Less concern is change of free air temperature wrt the standard atmosphere, as altitude error is proportional to the altitude difference. 10 deg C temp. change makes error about 33 metres for 1000 m difference. This is more concern for the aircrafts than terrestrials.
3/ The least concern is vertical temperature profile different from standard atmosphere -00065K/m. The error grows with the square of the altitude difference, but is usually small for the difference <1000 m. It is usually biggest during foggy cold seasons. In winter isotermia in lower 2000 m it makes about 11 m / 1000m and 44 m/2000m

Dne 18/02/2018 v 00:39 Robert Grant napsal(a):

While my experience agrees with you regarding accuracy and stability, it's still better to know the local pressure setting, especially if landing an aircraft without a radio altimeter. Setting an altimeter based on GPS sounds quite rare to me.

On Feb 17, 2018 3:12 PM, "Poutnik" <poutni...@gmail.com> wrote:

As being trained in past as the military meteorologist, in pre-GPS era,  I am aware of that. But the offset value is bigger than GPS accuracy of the static value averaged.   BTW, the most handy way how to calibrate the barometric altimeter at unknown altitude is the GPS device. While barometric altimeters have superior short-term accuracy and stability, GPS devices have superior long-term accuracy and stability.  Fortunately, for most personal usage, absolute altitudes are not that important, rather the relative changes.

[Robert Grant]

Guys, I'll do a spreadsheet from my future bicycle rides with various sets of data using bike computers with barometric altimeters and GPS only phones.  This data will get parsed in various ways based on source (GPS, baro), recording app (Osmand, Wahoo fitness) and post ride analysis programs (Strava, RideWithGPS, Garmin Connect).  I may even have my buddies that I ride with help since we're all data junkies.  My theory is that GPS will prove less accurate.

Virus-free. www.avast.com

--

You received this message because you are subscribed to the Google Groups "Osmand" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Robert Grant

[Poutnik]

You may have missed my point. I was not all trying to say what method is
more accurate.
But if you ask, barometric altimeter has superior short-term accuracy,
while GPS altimeter has superior long term accuracy.

Be aware  of the meaning difference of accuracy and precision.
And be aware about difference between short and long term accuracy.

Dne 18/02/2018 v 18:50 Robert Grant napsal(a):

> Guys, I'll do a spreadsheet from my future bicycle rides with various
> sets of data using bike computers with barometric altimeters and GPS
> only phones.  This data will get parsed in various ways based on
> source (GPS, baro), recording app (Osmand, Wahoo fitness) and post
> ride analysis programs (Strava, RideWithGPS, Garmin Connect).  I may
> even have my buddies that I ride with help since we're all data
> junkies.  My theory is that GPS will prove less accurate.
>

[Greg Troxel]

Poutnik <poutni...@gmail.com> writes:

> For me, it looks like both internal GPS units, in agreement with my
> experience, do not provide geoid correction to the amplitude. Therefore
> is is about 120' lower than the reference, what is about the correction
> value for US region.

The Android specification is to return ellipsoidal height:

https://developer.android.com/reference/android/location/Location.html

and one can get the NMEA also and use either geoid height or GPGGA
(orthometric) height:

https://stackoverflow.com/questions/2791927/how-does-getaltitude-of-android-gps-location-works#2797026
https://stackoverflow.com/questions/9361870/android-how-to-get-accurate-altitude

The real question is what's up with the external unit and how that
works. If it's being connected in to replace the internal GPS and
accessed via the same interface, it needs to report ellipsoidal height.
But the fact that it shows the higher (presumably close to correct for
orthometric height) value in satstat on both phones leads me to believe
that the android location/altitude (misnamed) interface is being used to
get the orthometric height from the bluetooth unit. That's the big
mystery.

The other mystery is that one of your phones seems to have the elevation
correction map in OsmAnd and the other does not.

Probably OsmAnd should change the symbol to something other than
mountain, perhaps "EH", to indicate that it's showing ellipsoidal
height, or make it red.

[Isaac]

The discrepancy between devices was my fault - I just realized that I had uninstall the world altitude correction map but the altitude corrections did not revert at that point. Seems a little strange to me, but the higher value only reverted after rebooting the device. It now matches the already corrected value that is output from the Garmin.

The "Bluetooth GPS" app that I use also allows for geoid correction, so it theoretically can be applied as many as three times if that box is checked in addition to the osmand and Garmin corrections.

[Robert Grant]

I'll take a stab at precision vs accuracy.  High precision:  Osmand reports the same value within a very small range for a reference elevation every time.  If reference is 292 feet, I would consider repeated indications of 262 feet to be very precise, though not correct.  High accuracy:  Osmand shows the correct value for the 292 ft reference elevation.  If Osmand reports 292 feet exactly I'd consider that highly accurate.  If an hour later Osmand then shows 414 feet at the same position, that report is not accurate.  If it later shows 210 feet that too is not accurate.  If more readings showed continued variation despite being spot on occasionally, I'd conclude that the device has low precision.

Virus-free. www.avast.com

--
You received this message because you are subscribed to the Google Groups "Osmand" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osmand+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Robert Grant

[Poutnik the Wanderer]

Yes, it is like that. While barometric method beats GPS one in short term accuracy and precision, it is vice versa in long term.

But for the most cases, users do not care about absolute altitude  nor about the profile being tilted along timescale. What they are interested in is the geometry of profile and the smooth lines.

And this is, why the barometric altimeter is a superior method for them.

Dne 19. února 2018 6:36:57 Robert Grant <rgran...@gmail.com> napsal:

To unsubscribe from this group and stop receiving emails from it, send an email to osmand+un...@googlegroups.com.

[Kevin Kenny]

On Mon, Feb 19, 2018 at 1:54 AM, Poutnik the Wanderer <poutni...@gmail.com> wrote:

Yes, it is like that. While barometric method beats GPS one in short term accuracy and precision, it is vice versa in long term.

But for the most cases, users do not care about absolute altitude  nor about the profile being tilted along timescale. What they are interested in is the geometry of profile and the smooth lines.

And this is, why the barometric altimeter is a superior method for them.

I've also seen systems for surface navigation that take the geodetic coordinates and obtain elevation from a digital elevation model. If there's an accurate, high-resolution DEM available, that may be the ideal compromise for drivers, hikers and such - they're seldom anywhere but the surface, after all, and it avoids issues with changing weather (for a barometer) and incompatibility of vertical datums (for GPS).

I love my wrist altimeter when hiking off-trail in mountains - it lets me use any contour line as a backstop. I probably depend on it too much, but the vegetation around here is usually too dense to allow for sight resection - in fact, on off-trail trips, as often as not, a viewpoint is the goal!

[Poutnik]

Similar methods, based on SRTM data ( grid 90x90, resp 30x30 m, or even finer data for some EU regions ) are used for the isolines in some OSM based maps, like the MapsForge compatible maps from OpenAndroMaps.

Some routers, like bicycle focused BRouter, use these data for elevation aware routing, as bikes are much more elevation sensitive then hikers or cars.

Last, but not least, some outdoor GPS applications like LocusMap, optionally replace elevation data in the track log by SRTM elevation data, using bi-cubic  grid interpolation. Due 2D position GPS fluctuation, there is also related SRTM altitude fluctuation on the sloped terrain. E.g. if there is slope 1 vertical m/ 5 horizontal m in the mountains,  position fluctuations 10 m around the place lead to altitude fluctuations up to +/- 2 m of altitude. This is worse than barometric altitude, but better than GPS altitude.

OTOH, these SRTM or the like data contain altitude artefacts due limited resolution. It is typical for cities ( "urban canyons"), wood/field borders, narrow valleys, rivers,  bridges, tunnels. For such places, 2D position fluctuation can cause quite huge SRTM based altitude errors. For such and similar mountain-like cases,  there is also an option to just use SRTM to correct the biggest GPS altitude errors.


Dne 20/02/2018 v 18:45 Kevin Kenny napsal(a):

I've also seen systems for surface navigation that take the geodetic coordinates and obtain elevation from a digital elevation model. If there's an accurate, high-resolution DEM available, that may be the ideal compromise for drivers, hikers and such - they're seldom anywhere but the surface, after all, and it avoids issues with changing weather (for a barometer) and incompatibility of vertical datums (for GPS).

I love my wrist altimeter when hiking off-trail in mountains - it lets me use any contour line as a backstop. I probably depend on it too much, but the vegetation around here is usually too dense to allow for sight resection - in fact, on off-trail trips, as often as not, a viewpoint is the goal!