Have done some testing on random medium gates and this update seems worse. I now have both trucks going through either my aircraft or the one next to me.
That's to be expected, the previous value was reasonable enough, but didn't work with some parkings like that one.
I'd like to know if it is possible to see the logic behind how the various vehicles choose their paths to the doors.
As I've already explained in my previous message, that particular situation is the result of having to throw away AFCAD nodes too close to the parking itself, otherwise the vehicle will try to follow them all, when they are already too close to the airplane to be able to maneuver.
But if we enlarge the threshold too much, the last AFCAD node might be too far instead, and the path between that one and the parked airplane might cross something, like another airplane or another vehicle.
I'd like to hope there could be a better way
When you have to deal with external data such a random AFCAD, it's very difficult to have something that will always work everywhere, that's why the previous threshold value was quite small, in order to remove only the really obvious problems (such as a node *inside* the parking radius).
Perhaps it might be possible to calculate intermediate nodes, but still something might end up in between, because we don't really know what's inside the scenery, only the AFCAD designer really does (or, at least, he's supposed to).