[HIL VTOL Offboard] set_attitude() not working

[ThomasRigi]

Describe the bug
Setting offboard attitude setpoints doesn't work for VTOLs in HIL simulation. The last attitude setpoint from before sending atttitude setpoints is kept instead.

This problem persists even with @irsdkv's fix #13141, even though now a few of the setpoints are registered :

To Reproduce
Steps to reproduce the behavior:

1. Switch to branch of PR Check system type for SET_ATTITUDE_TARGET attitude publising #13141 (optional, but recommended)
2. Start HIL simulation with the standard VTOL model (Gazebo)
3. Launch MAVSDK offboard example.
4. Observe the problem

Expected behavior
The vehicle should follow the offboard setpoints, like in this log where the MAVSDK offboard example works with Iris quadcopter in HIL:
https://logs.px4.io/plot_app?log=f032c92d-ffa3-4831-a730-977bbb8e229e

Log Files and Screenshots
Standard VTOL, HIL:
https://logs.px4.io/plot_app?log=1fae08b3-82f3-4a45-b5d9-ff997bb9e7a3

Drone (please complete the following information):

• HIL Standard VTOL QuadPlane, sys_autostart = 1002
• Pixhawk 4 is the used hardware

Additional context
#13141 fixes the problem for SITL, but interestingly enough not for HIL. Without #13141, it makes sense that it's not working, whereas with it I really don't know why the old setpoint persists.

[LorenzMeier]

Thanks for reporting! Given this is a fairly specialized and new usage scenario and not affecting a lot of users this is something that will take a while to get back to for the core maintenance team. It would be appreciated if you could dive into it and send a pull request.

[ThomasRigi]

I agree that it is something that will most likely not affect many users - for as long as this is not a hardware problem that ports over to real drones. (Even then it would be quite a limited user circle, but the consequences would be worse.) If real drones aren't affected and SITL is working, this issue certainly is on low priority - also for myself.

I tried to delve into it and thought to have tracked down every part of the code where something is written into the corresponding memory, but I couldn't find anything. Also, I struggled with using VSC's debugger for HIL simulation. Sometimes it kind of worked, but most of the time the HIL drone would refuse arming and again I couldn't figure out why. So don't count on a PR from me in the close future, sorry. :/

[julianoes]

FYI @Jaeyoung-Lim

[stale]

This issue has been automatically marked as stale because it has not had recent activity. Thank you for your contributions.