MC land detector question

[garfieldG]

Hi,

After looking at the land detector code I have a questions :

1. Why do we use vz in https://github.com/PX4/Firmware/blob/26e85736a52f5dd6cf2395c06a151d17ae0512f1/src/modules/land_detector/MulticopterLandDetector.cpp#L193, but we use derived z in https://github.com/PX4/Firmware/blob/26e85736a52f5dd6cf2395c06a151d17ae0512f1/src/modules/land_detector/MulticopterLandDetector.cpp#L200? shouldn't we use derived in both cases?

Thanks in advanced:)

[RomanBapst]

@garfieldG I'm not too familiar with that code but I think it has to do with the fact that the estimated velocity can have a considerable bias when landing which makes land detection harder. The derivative of the height state might suffer less and I think that's why it's used.

[garfieldG]

@RomanBapst Thank for response!
But I still can't understand why if we landed less than 2 seconds ago we use vz, otherwise we use vertical derivative. Shouldn't we use the same one (estimated vertical velocity or vertical derivative) in both cases?
@Stifael maybe you know how to answer on this one? relates to 26e8573

[Stifael]

if I remember it correctly, that commit entered the firmware against my advice. That commit was part of a test branch to check if it would improve the detection, which I could not verify (maybe it does, I was just not able to do verify it). The fact that the derivative term is not always used is probably because of inconsistency. It probably would make sense to evaluate once more the difference between z-deriv and vz.