Fix tailsitter frame conventions

[Jaeyoung-Lim]

Solved Problem

Previously, the tailsitter control axis of the fixed wing and multicopter had to be switched.

This has been changed in #21319, but has not been applied to the differential thrust control for attitude control

Solution

Remove the axis switching when copying torque setpoints for tailsitters

Test coverage

• Tested in SITL Gazebo: https://review.px4.io/plot_app?log=c1f84f4e-5bf8-43c2-9392-e84aaf500bc3

Context

• This popped up when rebasing Add support for quadtailsitter in SITL Gazebo #20558

[sfuhrer]

Looks correct, thanks for the fix! The change that broke it was though #20904 and not the linked PR in the description.

[julianoes]

Wow that's confusing. Are all these different frames somewhere explained?

[sfuhrer]

Wow that's confusing. Are all these different frames somewhere explained?

You mean the fw vs mc resp. torque_sp instance 0 vs 1? The best we have is https://docs.px4.io/main/en/flight_stack/controller_diagrams.html#vtol-flight-controller, which needs an update after switching to the control allocation topics.
Or do you mean the airspeed (FW) vs the body frame?

[julianoes]

Ok, I guess it makes sense looking at the docs. And then yaw goes to roll, pitch to pitch, and roll to yaw, because it's a tailsitter. Got it. Thanks!