Plane: Allow a Lua script to return to a previous mission track

[timtuxworth]

As discussed in #27639 on the dev call on 29 Aug 2024, added a way to get the previous WP location prior to switching out of AUTO mode and then restore it again after (in my case) doing some terrain object avoidance. This allows a plane to resume cross tracking mid mission which is required for magnetic field measurements in wilderness areas.

I changed the names proposed by @tridge in the call because:

1. getting the previous WP could conceivably be used for other purposes not just cross tracking, so I'm proposing vehicle:get_previous_location() - REMOVED based on the dev call on 12 Aug 2024
2. Restoring the previous location for crosstracking also requires setting the internal crosstrack boolean to true, so proposing to call this vehicle:set_crosstrack_start() which both sets the prev_WP_loc and also sets auto_state.crosstrack = true to resume cross tracking for the current mission item.

The sample script includes a LocationTracker() class which can be used to save previous target locations that can be passed to set_crosstrack_start()

Includes a sample lua script which requires you to be running a mission. A switch will save the previous location (high) and switch to loiter, then resume the mission and restore crosstracking (low). This screen shot shows a loiter being triggered, then switching back to crosstracking and resuming the path by crosstracking to the mission track rather than just heading direct for the next WP (WP 10).

[timtuxworth]

Tested on SITL and by Overhead Intelligence on a MFE Fighter VTOL

[tridge]

I think Pete is right that this functionality can be done in the script without a binding

[timtuxworth]

I got a CI fail on sitltest-plane. I downloaded the log but can't see the problem. Can you help @IamPete1 ? Oh found it - I'm pretty sure my code shouldn't affect this:

Failure Summary:
2024-08-14T19:53:35.4094916Z   test.Plane:
2024-08-14T19:53:35.4096328Z     DeadreckoningNoAirSpeed (Test deadreckoning support with no airspeed sensor) (GLOBAL_POSITION_INT diverged from simstate by 300.611553m (max=300.000000m) (see /tmp/buildlogs/ArduPlane-DeadreckoningNoAirSpeed.txt) (duration 7.538629055023193 s)
2024-08-14T19:53:35.4097532Z FAILED 1 tests: ['test.Plane']

[IamPete1]

I got a CI fail on sitltest-plane. I downloaded the log but can't see the problem. Can you help @IamPete1 ?

I have restarted it, that test has been failing occasionally.

[timtuxworth]

I got a CI fail on sitltest-plane. I downloaded the log but can't see the problem. Can you help @IamPete1 ?

I have restarted it, that test has been failing occasionally.

Is that something I can do myself?

[IamPete1]

Is that something I can do myself?

Not directly, maintainers only. However, whenever you push it will restart all of CI, so you can always rebase to re-trigger everything.

[timtuxworth]

Is that something I can do myself?

Not directly, maintainers only. However, whenever you push it will restart all of CI, so you can always rebase to re-trigger everything.

Thanks - looks like I need to do a rebase anyway. Are you good with the changes?

[timtuxworth]

Rebased on master