Re: [Paparazzi-devel] which stabilization subsystem is more stable ?

[Prof. Dr.-Ing. Heinrich Warmers]

Hi  Eltonwu ,

we fly  an wing aircraft Twin  Star  (1,8kg)  with an 460g  camera 
without stabilization inside.
The shutter time is fixed to 1ms and the   aperture is variable.
A club member fly this camera also with a mikrokopter  multicopter 
(total mass in the 4kg range) with a backstepping  stabilization  
control system  and good results.


The Paparazzi  kroozSD has a additional  acceleration sensor with a 
bandwidth of  only 17 Hz. like the current mikrokopter electronics.
So this muticopter has no propeller vibrations induced (4000RPM  results 
to 67 Hz)  disturbance like all the other.
On the other side the stabilization  PD control system in paparazzi  is 
not the best to take photos.
The  backstepping control system generates  less  disturbance but is not 
implemented in Paparazzi yet.

In the multikopter forum there are many discussions of the use of 
different cameras:
1. Cameras with CDD sensors need less  light  than cameras with  CMOS 
sensors.
2. Cameras with inside stabilization  have to be avoided since the 
vibrations of the propellers  make the inside control system unstable.
3. Old cameras (150g)  with CCD sensors  4 mpixel  generate better 
results than modern cameras with  16 mpixiel CMOS sensors.

For you  muticopter try the following:
1. use damping balls between the muticopter  frame and the camera mount 
system.
2. test to mount the battery  together with the camera  (mass helps 
again vibrations)
3. balance the propellers very well.
4  test to use the same acceleration 17Hz sensor like the kroozSD
5. for future work implement the backstepping  control system in Paparazzi

If i found a student  for the last point, i will generate the work.

Regards

Heinrich





Eltonwu schrieb:

> In the wiki page, I found 4 stabilization subsystem:
>
> int_quat
> float_quat
> int_euler
> float_euler
>
> In my opinion,performance of float is more powerful than int. Am I right?
> another question is which is more stable between quat and euler? In the wiki
> page I found quat is no gimal lock so it can fly in regimes close or at
> these singularities (e.g. acrobatics or transitioning vehicles). I just want
> fly more stable and take pictures.so It make more sense to use eluer instead
> of quat (search the source code I found the kroozSD use the int_euler, so I
> think euler is more stable than quat?)
>
>
> Thank you very much in advance
>
>
>
> --
> View this message in context: 
> http://lists.paparazziuav.org/which-stabilization-subsystem-is-more-stable-tp16577.html
> Sent from the paparazzi-devel mailing list archive at Nabble.com.
>
> _______________________________________________
> Paparazzi-devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
>