I have a vague understanding of the basics of machine learning, so I can give you some advice.
Firstly, I don't think this is quiiiite enough information for the drone.
If the drone only has an input that's the distance to the goal, the drone can know how far it is from the goal, but not where the goal is.
Second, the drone doesn't know how fast it's going or what its rotation is.
Aside from this though, it very well could be done with a neural network. I've tried to do something similar before, but my programming was really sloppy and it was super laggy. I might try it again now that I know a few ways to optimize my old neural net box though.
Let me know if you want me to (try to) do that as a scene response!
"and no i wasnt making it entire year. i just forgot my login"
Happened to me several times in the first few years I was using Algobox, it's nothing to be ashamed of
I highly doubt this counts but I built a machine to flip it for me.
The machine flipped it 6 times in a row. I know this contest is counting the total amount of flips, but considering I'm cheating by using a machine anyway it doesn't really matter
I informed UnityDogGaming04 of one of the glitches shortly after she released the VC3.0 scene.
Long story short, it's because the manipulator (the red thing) has nonzero restitution.
I told her about this and she has since reuploaded the scene with the proper fixes.
This doesn't affect this scene in particular as far as I can tell, but I do recommend turning the restitution to 0 on the player.
The main glitch I'm seeing is the manipulator getting stuck in the seams between objects.