[airframes] use standard motor mixing types for some quads #1231
Conversation
flixr
force-pushed
the
default_quad_x_mixing
branch
from
8312aea
to
847b4fe
Compare
|
@gautierhattenberger @kevindehecker @fvantienen @EwoudSmeur Also the ardrone2_optitrack.xml (which I didn't change here) has different gains, which I don't quite understand as it is the same controller and airframe. |
flixr
force-pushed
the
default_quad_x_mixing
branch
from
847b4fe
to
79b3f02
Compare
This will change the effect of the gains (depending on the type), but should be finally correct. Especially the yaw commands have about half the effect now...
After changing the motor mixing to the predefined QUAD_X type, multiply the roll/pitch gains by sqrt(2) and yaw by 2. This should result in the same effectiveness of the gains as before.
flixr
force-pushed
the
default_quad_x_mixing
branch
from
79b3f02
to
6ada818
Compare
|
Tested and it seems to be working as expected |
|
You tested both ardrone2 and bebop? |
|
I tested ARDrone2 in ATT and NAV mode with the indi airframe file, which seems to work. |
|
The question is if the updated gains match the new motor mixing scaling. Also I'm still curious on why neither ardrone2 nor bebop have any igains set... |
|
I tested both, it looks the same. I just don't like some of the parameters for the ardrone2 (max yaw rate, gains a bit higher than what I'm used to). Otherwise it is flyable like this. Only did ATT so far. |
|
It does also matter for the INDI controller, as the actuator effectiveness is sqrt(2) lower... Though Kevin was not flying with INDI but with quat_int |
|
Yes, but the indi airframe was not changed in this PR... And the question
|
|
Ups, sorry... I did change the ardrone indi airframe... Wanted to omit that one first... |
|
But now I actually have flown with adaptive indi, and it works great! And as @EwoudSmeur mentioned, apparently I flied with standard stuff op to now, and that also seems to work. (in both cases for ARDrone2 in optitrack enviroment that is) |
|
And what is up with the optitrack airframe and it's totally different gains? |
|
That I don't know. I use the standard indi gains with the optitrack systen, except for the h&v P and D guidance gains, which I have set much higher since there is perfect gps. Using those I get almost no position error, and pretty stable flight otherwise. |
|
Could you please discuss this with @fvantienen, you guys are the only ones using optitrack, if the ardrone2_optitrack.xml airframe file is outdated, then plz update it to what you are using... |
|
@fvantienen I think my setup for optitrack runs pretty smooth by now. So, how about taking this one #1257 ? |
This will change the effect of the gains (depending on the type), but should be finally correct.
Especially the yaw commands have about half the effect now...
Needs testing on ardrone and bebop, and probably some adjustment of the gains.