Situation Awareness - Free Fall Prevention

[Konstantin]

Sure! The very problem was the implementation of the "High" variable by expressing it as a function. The thesis is, that this function triggered the timeout. The solution was kind of easy: I just modified the original function in position_estimator.c, in this configuration it spits out the "High" variable.

Update concerning sitaw.c: Today I made some outdoor tests (with my CF bound to a small filament), but the results were not very satisfying.. The value fluctuations caused by wind etc. were to high and the CF did not really perform a smooth land. This evening I activated the kalman filter in my firmware and implemented its estimated Z value into the sitaw code. The indoor results look ok, the descent is veeery smooth. I guess this depends on the more accurate estimation given my the kalman filter. During the next days some outdoor tests will follow. If these are satisfying, I'll update the github repo (or put a second verson into). And maybe I'll make some videos, as this threat does not seem to attract much interested people..

What I would need is a small introduction into the python code of the client. Atm the Timer (triggered by freefall detection) can not be interupted by controller inputs, which is not optimal. E.g.: if the CF is about to land on a bad surface I couldn't deny that.

[tobias]

Very interesting stuff!

What kind of introduction are you thinking about? Have you had a look at the documentation on the wiki. It is a bit sparse but maybe a good start?

[Vincent]

Why does the crazyflie perform a flip instead of hovering "smoothly"? Any suggestion how