I am struggling with designing an AI that flies through a maze and fide a "score food"
I first tried immediate ray test based feedback but it seems to not work well, I think I need to somehow store it's detected ray test results and use it as its some sort of state,
Any idea how you guys would approach this?
#Flying AI that navigates through the maze
proper sail
cobalt magnet
sonic rover
I'm not sure ML Agents would be a good fit for maze solving. There is no feedback to give it once it finds a dead end. You are just having it wander around randomly