Self-driving cars are one of the most happening research areas in Artificial Intelligence. Lots of work is being done in several areas related to self-driving cars. One of those areas is vehicle control. Drifting is one of the hardest parts of autonomous vehicle control. Using traditional motion control equations that are derived from the vehicle dynamics makes it hard to generalize or model it precisely. In this project, we try to model an end-to-end drift controller that utilizes a state-of-the-art model-free reinforcement learning algorithm. This RL problem is devised as a trajectory following problem.
