Position control problems with typhoon_h480 model in gazebo sitl

[jannsta1]

The bug

Since Firmware Commit 43e3fc7 the typhoon_h480 model does not exhibit good position control in gazebo simulation with mavros. Still the case as at b0176dc. The error message Critical navigation failure! Check sensor calibration is logged in conjunction with the problem.

From release v1.9.0 to the commit prior to the reported issue (4a4d323) seem to work fine.

To Reproduce
Steps to reproduce the behavior:

1. Load commit version and ensure a totally clean build/initialisation (problem may not manifest unless all these steps are followed):

cd <FIRMWARE LOCATION>
git checkout <COMMIT UNDER TEST>
make clean
git submodule update
rm ~/.ros/eeprom/parameters
make px4_sitl_default gazebo_typhoon_h480

2. roslaunch model roslaunch px4 mavros_posix_sitl.launch vehicle:=typhoon_h480
3. Arm / transition to offboard mode
4. Send position target

Expected behavior
Stable position control behaviour (from commit 4a4d323):
https://logs.px4.io/plot_app?log=6c788665-95f7-4b2b-a963-9ad58911024c

Log Files and Screenshots
Unstable position control behaviour:
https://logs.px4.io/plot_app?log=894c8553-18b1-4fc1-ad4a-d1587485637e

Additional context

• Iris model still seems to work OK in this regard
• Have also observed Preflight Fail: Compass #0 uncalibrated but possibly due to not clearing parameter files between different build levels.
• when using commander takeoff in gazebo via the pxh> shell the vehicle seems stable.

[jannsta1]

Issue still present as far as I can see - would be great to see the reason this has been closed?

[FlavioTonelli]

You have probably been misled by the referenced Weekly Digest.
This issue is still open

[jannsta1]

So it is - silly me!

[Jaeyoung-Lim]

@jannsta1 Do you still have this issue? I can fly the H520 in SITL with the current master

[jannsta1]

@Jaeyoung-Lim - nice chatting with you at the dev summit :) and thanks for looking into this. I last checked with the master 5 days ago and still saw the issue - won't have access to my workstation until Monday so will try then and report back. Just to check though - is there a H520 model in the sitl_gazebo repo as well as the H480? Its also worth mentioning that its important to do the clean build steps (step 1 from the initial bug report) to ensure that everything has been rebuilt with desired git checkout.

[jannsta1]

@Jaeyoung-Lim - update as promised - I am still seeing the issue on latest master. Couple of other notes:

1. It is still possible to fly the vehicle - just that following takeoff the z position control is very poor and the
   warning messages reported in the bug report are present.
2. x & y position control seem to work as normal
3. The z position control stabilises eventually, but a new z position setpoint causes high oscillations again (I am using barometer for height control)

Finally - looking at the changes made on the commit when I started seeing this (listed above), there were many changes made to the sitl sensor IDs - I just didn't notice anything that would explain why this would effect the typhoon model and not the iris models.

[Jaeyoung-Lim]

@jannsta1 I will look into this as soon as possible. Currently I am overloaded with other stuff not related to SITL, so won't be able to look at it this week

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[Jaeyoung-Lim]

I have tested this and it can fly without any problems. @jannsta1 Please reopen if the problem persists