Deep Reinforcement Learning for Autonomous Driving: A Survey

With the development of deep representation learning, domain reinforcement learning (RL) has become a powerful framework now capable complex policies in high dimensional environments. This review summarises (DRL) algorithms and provides taxonomy automated driving tasks where (D)RL methods have been employed, while addressing key computational challenges real world deployment autonomous agents. It also delineates adjacent domains such as behavior cloning, imitation inverse that are related but not classical RL algorithms. The role simulators training agents, to validate, test robustify existing solutions discussed.