This object is in archive! 

Unstable altitude

Frédéric Furnelle shared this problem 8 years ago
Not a Problem

Dear,

I recently started with Locus Pro.

It works great!

I have 1 problem: the altitude is unstable, i.e., even without moving it goes up and down.

So basically, anything with altitude is incorrect and unusable.

I haven't had this with another app that I used before (Wikiloc)

I looked for a setting but could not find any.

Can you help?

It would be great if we could fix this, this week as I am leaving for a 4 week trek in Ethiopia next week.

Thank you

Replies (7)

photo
1

Hi Frédéric,


the altitude is basically calculated by GPS unit in your device. If your device is equipped by a pressure sensor, then the GPS values are corrected and optimized by it. All can be set in Altitude manager than can be found in Settings > GPS and location. All about Altitude manager is described here: http://docs.locusmap.eu/doku.php?id=manual:user_guide:tools:altitude

photo
1

Even without pressure sensor, the Accuracy could be always improved by Kalman fitering. It could use for the altitude predictor the height calculated for predicted future position and for the corrector the altitude from the current GPS measurement.

photo
1

Alas I don't have a pressure sensor in my phone. FWIW here is my experience:- By riding over the same (training) track many many times with various altitude manager settings, and comparing to a GPX track that was recorded with a (more accurate) pressure sensor, I have decided the "strong" filter provides the most accurate altitude measurements.

photo
1

Well, it does, but only if real altutude rate change is low enough. Note that in majority of time it is.

photo
1

Yes, without the pressure sensor you can optimize GPS values by offline elevation data (HGT files), engage filtering, you can set manual or automatic altitude offset...

photo
1

But the offset is not applicable to minimize random errors.

SRTM data have their own errors as well, often comparable to GPS altitude error, so the applicability is rather limited. The advantage of SRTM errors is lack of high frequency jumpiness the GPS altitude has. The disadvantage is the local artefacts pushing the altitude values away of real value. sometimes up to several tens of meters. (high buildings, narrow valleys, forests, tunnels, bridges, rivers ... )

Kalman filter minimizes GPS jumpiness, but avoids SRTM artefacts. Simple exponential filter would introduce delay of its time constant, proportionally to its strength.

photo
1

Kalman filter is a good solution but at this moment it's not in our working plan. Maybe later.

Replies have been locked on this page!