Locus reports 650 meters Uphill Elevation on a lake trip!

[Menion]

thanks for link, I'll look at it later.

Support for pressure sensor added. Current system:

- when you turn on GPS, first five locations is used for calibration. After that, altitude is computed from pressure values. Every new location is checked and if difference between measure and computed value is too high, new calibration is computed. Probably too simple, but we'll see

[joeloc]

Ive read that garmin sort of recalibrates gradually with time to counterbalance weather influence. No immediate changes though, its more like a gradual thing moving the two data streams closer together.

What seems to be important:

A) only very good 3d gps fixes with excellent accuracy are taken into consideration for calibration at all.

B) when the difference of gps & barometric data is too big (garmin uses 300m i guess), measuring is considered an error and not used for re-calibration.

Obviously we need to be able to turn off all pressure sensor handling and use raw gps data like now. Otherwise you cannot track heights while flying in pressurized cabins.

[joeloc]

In general, i think automatic re-calibration during track recording should be done very careful, if at all. Its much more likely to receive a bunch of bullshit gps fixes (think canyons, deep wet forest, rock climbing) than it is to receive wrong barometric data.

[joeloc]

Just thinking... it might be worth to consider storing raw data in your data base, ie both gps value and pressure sensor value. You could do the calculations later on export/display. And maybe do funny things too, like a weather forecast .

[Menion]

- yes, re-calibration during record seems like a problem. When recalibration is done, recorded altitude skip to new calibrated values and this cause "altitude jump" ... probably not wanted.

- ability to disable/enable using of pressure sensor is in settings, so no problem to turn if off immediately

- about database storing. Hmm ... I was also thinking about storing raw altitude data and also data from some heart-rate sensors, so probably this will be a good idea

[joeloc]

I think older garmins had the altitude jump problem after recalibration. So they changed it to the gradial thing in newer devices/firmware

Is the barometric locus online somewhere for testing already? I could do a sea journey recording in 30 mins and see if it stays at deck altitude properly

[Menion]

not yet, I'll create one when I'll be able to compile it give me one-two hours

[bebe814]

Hi, I would like to post my opinion in this topic

Basically you have two options when using gps with barometric values,
1. automatic calibration (like garmin mentioned above)
2. manual - you have manully enter pressure at sea level or altitude

To get ideas you can check SportyPal software where it is already implemented. But I prefer manual setting where recalibration takes effect only when user wants. In hiking it is not changed too often as area and period where pressure changes are not too big. But in aviation terms, there are changes more common

Basically there are 3 values with 1 user input, 1 measured and 1 computed values
1. local pressure (measured value taken by sensor) (QFE in aviation terminology)
2. pressure at sea level (QNH)
3. altitude
(either 2. or 3. serve as an input the other one is then calculated)

In hiking most common approach is to set 3. as input then you are able to compute QNH (2) and then every change in pressure correspond to offset change in altitude. (e.g. suunto watches are using this method, altitude is taken from known points/landmarks etc)

In aviation most common approach is to set 2. as input then you are able to compute present altitude and then changes in pressure correspond to offset change in altitude ( QNH value is available in weather METAR messages for every airport or by tower when flying to different area - woudl be usefull to change during flight/track easily. Btw. flying above so called "transition altitude" (often 8000ft) requires all planes/glider to set QNH to 1013 - in aviation, pressure altitude ALWAYS takes precendence over GPS altitude, so manual setting would be welcome.

calculations can be taken from this page as resource http://www.hochwarth.com/misc/AviationCalculator.html (first part regarding QNH/QFE) or if exact equation is necessary let me know, but it is inside JavaScript on that page

[Menion]

thanks bebe, it was much worst (mainly on time) then I expect, so I spend almost whole day on this ... nevermind

test version here: viewtopic.php?f=25&t=1404#p8322

[joeloc]

Is there any way to tell if locus is actually using the pressure sensor?

Just thinking... for testing and maybe also for geeky details, locus could add a waypoint whenever it does some calibration thingies to the sensor. Add a special "automated" poi category so users can enable at will. This could also be used as a general "geologging" thing. You could create "automated" waypoints for many different events (like track export, minimize/maximize, anything of interest).

[Menion]

no there isn't. I was thinking about some notification when you enabled GPS and using of pressure sensor is enabled. This is next step ...

[bebe814]

just quick tested in my flat and seems fine, change in 1 milibar ~ change of 10 meters (approx), when weather get's better will check with airplane devices.

if anyone want calibrate (pressure based) you can use http://www.aviador.es/Airports/Nearby/ search any airport near you and in METAR (meteo report) find Qxxxx where xxxx is pressure at sea level e.g. Prague LKPR -  METAR LKPR 052130Z 26021KT 9999 BKN042 03/M02 Q0988 NOSIG= pressure is 988 hPa

Maybe another idea but not so important: mainly for hiking - to display height even when no gps fix available, can be good for someone in canyon or dense area. But for aviation it's not necessary as we shouldn't have any obstacle nearby ;-)

Big thanks, I'm also GIS developer and haven't seen anybody to react so fast on user requests ;-)

[Maksym]

First 5 GPS locations using for calibrating pressure sensor are very inaccurate. In about 1 minute after GPS start hight is more accurate. But if I start GPS inside the building I'll get inaccurate height until go to open air. So it's difficult to calibrate pressure altimeter by first GPS locations. Maybe better calibrate using SRTM and GPS. Now I can calibrate by SRTM but only manually in preferences - it's difficult to make calibration on every program start.
So really now I compare dynamic height by SRTM, GPS height and when the difference is small I turn GPS off/on and pressure sensor calibrated correct automatically.
Of course re-calibrate during writing track is bad idea.
I want to see function to quick calibration of pressure sensor. Maybe manual calibration when I see that GPS height is correct, but not in deep preferences menu. Maybe by SRTM automatically. Mabe by GPS height but compare with SRTM and when the difference is too big don't calibrate.

edit
I test different methods to calibrate pressure sensor. The best and most accurate way to calibrate pressure sensor is set up pressure at sea level by METAR (thanks to bebe814). Maybe can add automatic calibration using nearest METAR data? If not now, please, can you add a button for quick enter sea level pressure (now I must do it deep in preferences).

edit
I still test Locus in different situations. Today I see dramatically fast preassure fall during day from 1023 to 1010 mmHg. So it's very important to re-calibrate altimeter periodically. But how to make it better? I don't know . Of course it will be great if Locus in future can calibrate using nearest METAR data and check pressure every hour. But in hight mountines no Internet connection for check METAR. But I can download SRTM data and put it to my phone. So for calibration compare height of SRTM data and GPS height and calibrate when both height are almost the same. And repeat calibration time to time.
Does anybody know how solve problem with altimeter calibration in GPS devices like Garmin, for example?

[Klonutzer]

Hi, despite the topic is quite old I still see huge jumps in recorded track altitudes. Is there a build in solution now?

[Menion]

Hi Klonutzer,
  there were quite a lot of discussions about this topic

  I anyway suggest to check Altitude manager screen in Locus (more here https://asammsoftware.zendesk.com/hc/en-us/articles/200028131-Altitude-manager ). There are some techniques that may help. Mainly if you set stronger filter in Altitude filter section

  Second possibility is to Fill recorded track with computed altitudes. Locus allow to compute completely offline altitude for a whole track, which is very very fast and result should be very good.