AP_DDS: Joystick Support 2

[tizianofiorenzani]

I noticed that this Pull Request has been stuck for one year.
#25334

Tagging @srmainwaring and @Ryanf55.

[Ryanf55]

Please add some tests (automated), or some instructions on how to test this in CLI. Once we've got those and we're happy, the commit history can be cleaned up.

Nice job cleaning up all the previous actions. Looking good.

[Ryanf55]

We also need some idea on how this will interract if you are also sending normal RC input (SBUS from RC transmitter for example). If that gives complete authority to the companion computer, whether the two inputs fight, or something else.

[tizianofiorenzani]

@Ryanf55 I am working on the tests. For the RC, we might check whether the Radio is connected before overriding its channels. This would prevent any conflict and prioritize RC at all time.

[Ryanf55]

Yep. It's worth reviewing this wiki and seeing whether we want to make DDS act in a similar way of "no pulses" and re-use the existing failover logic.
https://ardupilot.org/copter/docs/common-multiple-rx.html#failsafe-and-changeover
It's not clear how ArduPilot ranks the priority of two active receivers if they are both providing pules.

[tizianofiorenzani]

The RcChannel already handles overrides:

• If a valid value is passed (>0 and <UINT16_MAX), the channel is overwritten.
• If 0 is passed, the channel is released (falls back to RC).
• There is a FS_RADIO_RC_OVERRIDE_TIMEOUT_MS=1s timeout for each channel. So if you keep commanding UINT16_MAX, the channel will eventually fall back to RC after 1s.

[Ryanf55]

Looks good, adding in the teleop joy instructions could be nice for this PR. I can dump it into the wiki.

Can you clean up the history and fix your git config user.name to not be =?

[tizianofiorenzani]

@Ryanf55 of course, I forgot to update my laptop configuration.
Also, I have changed the code such that an input joystick value >= UINT16_MAX disables the channel override.

[Ryanf55]

Needs update to libraries/AP_DDS/README.md

[Ryanf55]

I force pushed with clean history. Can you share which email you want to attribute the work to?

I also

• Ran astyle (./Tools/scripts/run_astyle.py
• Edited the commit message to squash to one commit with the AP_DDS: prefix that is required

[Ryanf55]

Functionally seems to work fine. It clamps to -1/1 through the linear interpolate nicely. I think it's OK to silently clamp. NaN discards.

I think it might be nicer in the readme to give a command for a right turn and climb in FBWB.

[tizianofiorenzani]

@Ryanf55 The test is pushed and working now, run it with:
colcon test --packages-select ardupilot_dds_tests --event-handlers=console_cohesion+ --pytest-args -k test_joy_msg_received.

I also updated the Readme, I think we are good to go once CI has passed.

[tizianofiorenzani]

@Ryanf55 I just noticed that when I passed the astyle on the Readme, it really messed it up. I will need to restore it and push it again.

[Ryanf55]

• ./Tools/scripts/run_astyle.py

How did you run astyle? This doesn't run on the README. astyle is a C++ formatter, not a markdown formatter.

[tizianofiorenzani]

As a simple test, you can

• run the ros2_sitl, quadcopter is fine, then takeoff.
• On a new terminal, start sending joystick axes with ros2 topic pub /ap/joy sensor_msgs/msg/Joy "{axes: [0.0, 0.0, 0.0, 0.0]}"
• Set the vehicle in LOITER mode
• change the published joystick to ros2 topic pub /ap/joy sensor_msgs/msg/Joy "{axes: [0.5, 0.0, 0.0, 0.0]}" and you will see the vehicle moving sideways.

[Ryanf55]

Nice, looks good! Remove dead code and squash all your commits to one AP_DDS: ... commit.

[tizianofiorenzani]

sitl-plane test fails. I don't know how to test it locally, here it's what is says:

2024-09-21T21:16:51.5884422Z >>>> FAILED STEP: test.Plane at Sat Sep 21 21:16:51 2024
2024-09-21T21:16:51.5885040Z Failure Summary:
2024-09-21T21:16:51.5885407Z test.Plane:
2024-09-21T21:16:51.5887138Z DeadreckoningNoAirSpeed (Test deadreckoning support with no airspeed sensor) (Failed to get EKF.flags=16) (see /tmp/buildlogs/ArduPlane-DeadreckoningNoAirSpeed.txt) (duration 10.50739574432373 s)

[Ryanf55]

sitl-plane test fails. I don't know how to test it locally, here it's what is says:

2024-09-21T21:16:51.5884422Z >>>> FAILED STEP: test.Plane at Sat Sep 21 21:16:51 2024 2024-09-21T21:16:51.5885040Z Failure Summary: 2024-09-21T21:16:51.5885407Z test.Plane: 2024-09-21T21:16:51.5887138Z DeadreckoningNoAirSpeed (Test deadreckoning support with no airspeed sensor) (Failed to get EKF.flags=16) (see /tmp/buildlogs/ArduPlane-DeadreckoningNoAirSpeed.txt) (duration 10.50739574432373 s)

Some tests are flaky. Don't worry about those passing till you have the commit history clean, and I'll rerun the flaky unrelated tests.

[tizianofiorenzani]

@Ryanf55 I rebased and cleaned up the history, leaving one single commit after your previous approval.

[Ryanf55]

I tagged for dev call for review, but will be late. Feel free to provide any feedback blocking merge.