Servers by jDrones

Changing the behavior of flare to be more like a deep stall


(Pedro Henrique Silva) #1

Hello, guys

I’m new to this dev environment, I just managed to compile my first firmware.
Currently, I’m using Arduplane 3.6 and landing with rev_thrust. But sometimes, depending on the wind conditions, when the flare is triggered, the plane takes too long to land. And depending on the aproximation slope, if the plane doesn’t manage to reach the Land point on a lower altitude, the flare happens too high off the ground, which can be dangerous.
So, I’d like to modify the flare to besides of killing the throttle, it also keep the ailerons with a high value (2000 PWM to the servo 2, for example) to perform a deep stall. In that way, it won’t matter how the LAND point is reached, I will always have a safe and short landing.
Does anyone knows the simplest way to achieve that? I’m afraid of messing with something that is not supposed to be changed.

Thanks!


(Kikislater) #2

In 3.8 there is a note :

  • support for deep stall landings

https://github.com/ArduPilot/ardupilot/commit/e66f30eb90f75e52a3e37e25a65df4a532fdead2

Did you try it ?

http://ardupilot.org/plane/docs/parameters.html?highlight=deepstall#land-type-auto-landing-type


(Kikislater) #3

An update here on a bug :


(lucamax) #4

WIthout any explanations in the Wiki it is almost impossible to use the new deep stall landing feature.


(Kikislater) #5

I wonder if without lidar or other ground detection it will work or not …


(lucamax) #6

I imagine that the deep stall landing is similar to a water landing where a constant descent rate and attitude is maintained till the water is touched.
So the Lidar is not necessary also because I think that the steep angle of the airplane would make the Lidar reading not so accurate even if there is a function that calculates the plane angle to correct the Lidar reading.


(Kikislater) #7

Not sure about it. Cumulus one have it, have to check if they put lidar inside.
Better is to check the code …


(lucamax) #8

I’m sure that the code has the function that calculates the attitude of the plane and correct the reading of the Lidar because this aspect raised in a thread about aborted landings.

If someone gives some directions about this new landing mode I would be happy to run some test.


(Kikislater) #9

Additional reading here :
https://github.com/ArduPilot/ardupilot/issues/3737

I think lidar is needed due to baro drift in long flight …


(lucamax) #10

Baro drift is part of landing problem.

The main problem IMO is that actual landing strategy works only is you use LAND_FLARE_SEC

If you use LAND_FLARE_ALT even with a Lidar and reverse throttle landing do not work.
I have made dozens of test with erratic results.
There are a lot of problems and it is almost impossible to make a real debug because the autopilot just do not do what is supposed to do.