Altitude accuracy problem

Hello everyone,

During the last flights, I have been experiencing problems with the altitude when I fly predefined flightplans. This has caused up to 30 meters error between the desired altitude of the waypoints and the flown trajectory in some missions.

I made the following plots comparing the barometer and GPS altitude readings to the actual flown trajectory (extracted from the APX-15 readings), trying to find out if the issue was related to one of the sensors, but both of them report a similar flight pattern.



As the waypoints do not have timestamps, I have also included images of kmz files, showing the difference between the flown trajectory (yellow) and the desired trajectory (green).

Apparently the altitude deviation appears at the first waypoint and remains constant during the whole flight.

For these flights the vehicle was using the barometer as the primary height sensor. EK2_ALT_SOURCE = 0
EK2_OGN_HGT_MASK = 0

I have taken into account many other parameters like GPS/baro health, weather conditions, altitude of the launch point, flight patterns of the programmed missions, PID controller, and much more but I cannot identify the source of this error.

Any idea which can be the source of this altitude difference?

All is about reference and local altimetry data (QNH).

https://www.esri.com/news/arcuser/0703/geoid1of3.html

If you start your flight at sea level with a standard pressure (1013Hp) all is at zero.

Hi, thanks for the reply. Do you know if there is in missionplanner any parameters or configuration that can fix this issue?

There is also a more technical explanation:
a) some barometers require to be covered in order to reduce effects of wind and in some cases UV light affecting the sensor.
b) changes in air pressure as found whilst passing through hot and cold air streams can have an impact - Those can be often found near mountain ranges / tall hills
c) the actual passing of air past the sensor itself can have an impact - The faster the speed the more likely it is if sensor is exposed
d) There have been various types of barometers been used in FC’s over the years. - Some much more precise and reliable than others.

Hence in my experience one has to decide what should take priority, the GPS values or the barometer reading. - Or in your case perhaps even the LIDAR.
Often such results as you describe can provide insights as to what will work best for you.

Thanks for your insight Karl.

I had already done some calculations comparing the GPS and the barometer drifts. The barometer readings are way more reliable when it comes to altitude changes during the missions and matches quite well the programmed trajectory. The main issue is that when the mission is launched and the drone is flying to the first waypoint, it does not reach the desired altitude, it is usually below or above it.

If it was possible to get the right altitude on the first waypoint of each mission, the problem would be probably solved.

I’m attaching another image, where one of the missions reached the desired altitude on the 1st waypoint and the obtained flightpath was almost identical to the planned one.

It is interesting from @mlebret that the example given at the ESRI headquarters in Redlands, CA has about the same deviation as witnessed in this problem. Indeed this looks like a GPS bias issue.

Certainly is an interesting issue.
May I ask / double check what version Arducopter are you using ?
How did you create mission ? …using height values above sea level or using height levels above ground?