You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was reading this paper and so I am here to know the scope of this technique in PX4 firmware. I want to know if this can be implemented in PX4 and what should be the overall approach and strategy.
Paper : Actuator Fault Detection and Fault-Tolerant Control for HexacopterLINK
Some tips would suffice on the overall implementation and design strategy.
Also please note this/similar question was unanswered on the discuss forum link here.
Looking forward to some guidance and insights into the scope of this paper into becoming an actual working implementation in PX4.
cc: @LorenzMeier
Thanks,
Sayantan
The text was updated successfully, but these errors were encountered:
Hi @ucalyptus
I started working on dynamic mixing to address the fault tolerance part (not the fault detection). You can find it here : #13351
The approach is to dynamically recompute the mixer in case of motor failure.
Here is a demo with an hexacopter loosing one then two motors: https://www.youtube.com/watch?v=yR5okaHITpg
Feel free to help if you will, that would be much appreciated!
That said, if all you want is a fault tolerant hexacopter, there would be easier ways. Because the work linked above aims to be applicable for all kinds of robots, it involves computing pseudoinverse on-board, which is pretty involved.
For a more limited case, say an hexacopter in x config, a simpler alternative would be to precompute the mixers for all failure cases (6 times one motor failure cases + 2-among-6 double motor failure cases) at compile time and avoid the pseudoinverse.
I was reading this paper and so I am here to know the scope of this technique in PX4 firmware. I want to know if this can be implemented in PX4 and what should be the overall approach and strategy.
Paper : Actuator Fault Detection and Fault-Tolerant Control for Hexacopter LINK
Some tips would suffice on the overall implementation and design strategy.
Also please note this/similar question was unanswered on the discuss forum link here.
Looking forward to some guidance and insights into the scope of this paper into becoming an actual working implementation in PX4.
cc: @LorenzMeier
Thanks,
Sayantan
The text was updated successfully, but these errors were encountered: