Bitcraze Forums

After observing the log data of the magnetometer a considerable noise is noticed. When I rotate the CF and come back to the previous heading, it gives different headings in each time. Because of that I have following questions.

1) Does kalman estimator use the onboard magnetometer readings ?
2) Will it be helpful if we cover the magnetometer using a ferrite cover?

Hi thasu,

As far as I understand, currently the magnetometer doesn't play a part in the estimation [1]. If you want to simply find the yaw value, you can use the value from the state variable. It seems to be pretty good. (It doesn't have the sense of north, that'll produce the yaw relative to the initial position at the beginning of each session).

[1] https://github.com/bitcraze/crazyflie-f ... man.c#L386

Does the algorithm use magnetometer readings if we pass values from an external magnetometer ?

Unfortunatly the kalman filter does not currently implement support for a magnetometer: https://github.com/bitcraze/crazyflie-f ... #L381-L383.