An environment for OpenAI gym emulating the classic Snake game
Observation state is NxN matrix 0 = empty cell 1 = snake 2 = fruit
+1 reward for every fruit eaten
snake grows in length when fruit is eaten
episode is done when snake hits wall or itself
actions: down = 0 up = 1 right = 2 left = 3
- Fix collision detection system
- should actions be [ left, right, continue] ?
- timestep between actions?
- add multiple snakes