Self driving car with Reinforcement learning (Q learning) remix

Instructions

Use turbo mode (Hold Shift and click the green flag) for faster result. The car uses 6 sensors to gather data and updates its Q-table (expressed as red dots on the bottom) based on the immediate reward and the Q-value of the subsequent state. The AI learns to drive straight for the first 20 generations then goes on to more complex maps. After the 20th generation, the map changes by 50% chance when the car crashes or reaches the goal.