ekf2: check mag field with mag bias removed

[dagar]

• this is to allow the estimated mag bias to clear mag_field_disturbed and unlock mag_3D/mag_hdg

Opening for review, discussion, testing.

[bresch]

Should we then rather have the checkMagField in common_conditions_passing ? Otherwise the bias can't be learned if "mag" fusion doesn't start.

[dagar]

Yep, doesn't make sense otherwise. Thanks

[dagar]

@bresch are you using the _mag_strength and _mag_inclination logging and if so do you want it to be from the raw mag_sample or bias corrected?

PX4-Autopilot/src/modules/ekf2/EKF/aid_sources/magnetometer/mag_control.cpp

Line 533 in 8dc84c7

_mag_inclination = asinf(mag_earth(2) / fmaxf(mag_earth.norm(), 1e-4f));

If it matters I'll split up checkMagField(), probably make it a pure function, and then set the rest of these logging and status fields separately.

[bresch]

are you using the _mag_strength and _mag_inclination logging and if so do you want it to be from the raw mag_sample or bias corrected?

Yes, I find it really useful to quickly check if there is an issue with the mag. Since we'll do the checks on the bias-corrected data with this PR, I think it makes sense to also log the bias-corrected inclination and strength (as they can be used to tune the thresholds).