Manual Control refactor

[DavidPaddy97]

This PR is interlinked with the following SeesInterface2 and MAVSDK PRs:

• SI2
• MAVSDK

Testing for this PR was completed here.

Problem(s)

This PR includes fixes and improvements for the following bugs:

• 4467 - The QGC backup and RC control connection status incorrect
• 4385 - RC Kill switch does not work if in “No Source” state
• 3792 - PX4 doesn't switch to local RC on Taranis mode change.
• 4401 - MAV/RC transition info message incorrect

Solution(s)

4467 - The QGC backup and RC control connection status incorrect

Here, data_link_lost was used by SI2 to monitor whether an instance of QGC (implied to be the Backup GCS) was connected or not.
However, data_link_lost considers any Mavlink connection as a valid data link; for example, it does not distinguish between RTK Sender and QGC.

Instead, we now count the number of manual_control_inputs (manual_control_setpoint candidates) that are live and are from a Mavlink source which is published to a new topic called "sees_manual_control_data".

Mavlink connections must have a joystick connected (can be virtual) to provide a setpoint, and our configuration should only ever have one Mavlink Joystick connected; the Backup GCS.
Subsequently, other 'passive' Mavlink connections are ignored in this count (RTK QGC, Mavlink Router, LDS...).

This also allows us to monitor if there are any additional unexpected Mavlink Joysticks connected that may cause conflict (e.g accidental connection from Van Backup GCS).
The number of RC manual_control_inputs is also logged but is not currently used for anything.

We also now pass the desired control source (RC or Mav) as this is more informative in the way we now toggle between sources.

4385 - RC Kill switch does not work if in “No Source” state

Previously, the RC switches would only be executed if the current selected input was valid and its source was RC.

This PR moves the RC switch execution process to a separate method which can then be called under 2 different conditions:

• If COM_RC_IN_MODE parameter = sees.ai control selector behaviour (5), execute RC switches if a live RC setpoint is present.
• If COM_RC_IN_MODE parameter != sees.ai control selector behaviour, execute vanilla behaviour.

3792 - PX4 doesn't switch to local RC on Taranis mode change.

Now that RC switches always work (see 4385 above), we've added a toggle to RC Control when the RC flight mode switch is changed.

4401 - MAV/RC transition info message incorrect

This was a small bug where a datatype was wrongly set to Bool instead of Int. This resulted in the feedback always equalling RC (1) even when it should be Mav (2).

Additional improvements:

• Removed additional enum in manual_control_setpoint.msg, alternative solution found as wasn't best practice.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

❗ No coverage uploaded for pull request base (v1.13.1_dev@20cec0d). Click here to learn what that means.

❗ Current head 1b7aac0 differs from pull request most recent head 9ed65b8. Consider uploading reports for the commit 9ed65b8 to get more accurate results

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@              Coverage Diff               @@
##             v1.13.1_dev      #62   +/-   ##
==============================================
  Coverage               ?   58.92%           
==============================================
  Files                  ?      958           
  Lines                  ?    83110           
  Branches               ?        0           
==============================================
  Hits                   ?    48975           
  Misses                 ?    34135           
  Partials               ?        0           

Flag	Coverage Δ
mavsdk	38.79% <0.00%> (?)
unittests	43.02% <0.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[RichardHopkirk]

As discussed I haven't reviewed exhaustively but this looks fine and I can't see any issues from a reasonable review of the changes. Also you've tested on the bench and the behaviour seems as intended - including timeouts etc. all working.

Next steps:

• You'll have a quick look at Manual Control tests - see if they can be extended easily to test our functionality.
• And before merging you'll do a test flight on the real drone in the air.

[wsolichin-sees]

Do we need uint64 if the expected count is less than 2? I think uint8, similar to sees_desired_control_source, should be sufficient.

[wsolichin-sees]

Similarly to valid_mavlink_setpoint_count, it might be good to change the type to uint8 if the integer value is not large so the message size is not too large.

[DavidPaddy97]

Good spots, thanks!