Simple reinforcement learning (Q-Learning) pathfinder for UnicornHatHD.
This programme requires the unicornhathd module for the raspberry pi.
Run the programme by invoking the play() command. Specify verbose=1 to see information on rewards, and epsilon (exploration criteria).
Additional obstacles / mazes etc can be added by amending the build_walls function.
Those familiar with Q-Learning may wish to alter the initial epsilon / epsilon decay values.