Elevation Data optimizations
GPS based elevation information can be so bad
Overview - Improving Elevation data
The app has implemented two independent functionalities to improve the elevation data provided from the GPS signal.
- Use a Mean Sea level altitude reference - correction of WGS84 elevation information to EGM2008
- Make use of Barometer data in order to detect altitude changes (if sensor is available)
WGS84 vs. EGM2008
TLDR; The elevation information provided via GPS is actually not want you want.
I don’t want to waste you time with all the details & reasons why it is what it is - and also I don’t want to start a discussion here & now as to if there are better alternatives to EGM2008 or not (you are invited to start one in the forum).
What’s important to know is, that GPSLogger will record a GeoIdHeight (and if this data is not available, the app will post-calculate it). This GeoIdHeight will be used, to correct the WGS84 elevation data to something that comes quite close to the elevation information we use here on planet earth.
As already described in the How GPSLogger will log your data Section, all GPS data will be recorded as it is, and the MSL correction (when enabled) will be applied when the data is displayed.
If you want to read the full story, use the following Wikipedia link as a good start point:
The use of the Mean Sea level altitude reference is enabled by default - you can disable it in the Application Settings > General Settings > Altitude Reference Model
Using Barometer based Elevation - Theory of operation
With GPSLogger it’s possible to capture barometer data - from the barometer sensor data the app can calculate an internal 500ms average value and then use the avg pressure value as base to calculate your current elevation.
A few things I would like to explain additionally:
-
The app stores the barometer sensor value - from that stored pressure value the elevation value will be calculated during display/export time - To calculate the actual elevation you can base it on the yearly avg pressure at 15 °C and sea level - with this formula you can calculate with a given pressure the current altitude - but of course this will be not quite correct since the pressure is changing weather wise - and we do not have a constant temp of 15 °C (so I decided against that)… What I do is that I calculate the elevation based on the actual pressure change - this generates two small additional challenges: what is the base pressure? and what is your starting elevation?
-
I decided a simple approach to calculate the base pressure and base elevation - When you start the application the app is going to track the current GPS based elevation and the actual pressure - when you start recording I calculate the average GPS elevation and average pressure of the last 30 seconds. Then I use these average values for the actual elevation calculation (BaseGPSElevation + currentPressure/BasePressure) - by doing this, the elevation calculation should be IMHO quite weather and mostly temperature Independent…
Warning some Physics & Maths here - In the actual formula the temperature used is Kelvin - this temp in K will then be divided by 0.0065 and used as the multiplier for the relation of both the pressure values -> so at 0 °C the factor is 42023 and at 30 °C the factor is 46638 - so the difference between both factors is 1,1098 - this will result in a difference of 10 cm when a pressure change happens at 0 °C compared to the same pressure delta at 30 °C
I don’t know how drastically the temperature is going to change during your rides - my initial train of though here is, that a delta of 10 °C might happen… If temp changes more drastically I can add some things in the post-processing that will allow you to also take the temp change also account (therefore it would be cool if you take a device that records the temp over time with you) - I think I have seen some bike computers doing that… [there are really few phones that have an external temp sensor - because the device itself generates quite some heat] -
The app records the gps elevation and the barometer values separate from each other - so in the internal file format there are always two values present - during GPX Export the user has to decide which value should be written into the elevation data field
Preparation before Start Recoding
Before you start with the recording, its recommended being stationary for at least 30sec. The best thing you can do is to put your device on something solid/stable like a table, the ground or the roof of your car. To allow the Barometer sensor to calibrate itself.
When you start the app while moving up or down hill and then start the recording, this will result in non-plausible BaseElevation & BasePressure values. Then the calculated pressure based elevation values will also not be correct. BUT there is no reason to jump out of the window in this case. It’s possible to correct both base values after the recording in order to finally get good elevation values. This adjustment is not an easy task, but you don’t have to be a hacker.
- Export the Path as JSON
- Open the JSON file in a text editor like Notepad
- Search for the fields basealt (, basemslalt) and basepsi and edit them
- Save the file & import the JSON into GPSLogger