Behavioral decision-making for urban autonomous driving in the presence of pedestrians using Deep Recurrent Q-Network

Decision making for autonomous driving in urban environments is challenging\ndue to the complexity of the road structure and the uncertainty in the behavior\nof diverse road users. Traditional methods consist of manually designed rules\nas the driving policy, which require expert domain knowledge, are difficult to\ngeneralize and might give sub-optimal results as the environment gets complex.\nWhereas, using reinforcement learning, optimal driving policy could be learned\nand improved automatically through several interactions with the environment.\nHowever, current research in the field of reinforcement learning for autonomous\ndriving is mainly focused on highway setup with little to no emphasis on urban\nenvironments. In this work, a deep reinforcement learning based decision-making\napproach for high-level driving behavior is proposed for urban environments in\nthe presence of pedestrians. For this, the use of Deep Recurrent Q-Network\n(DRQN) is explored, a method combining state-of-the art Deep Q-Network (DQN)\nwith a long term short term memory (LSTM) layer helping the agent gain a memory\nof the environment. A 3-D state representation is designed as the input\ncombined with a well defined reward function to train the agent for learning an\nappropriate behavior policy in a real-world like urban simulator. The proposed\nmethod is evaluated for dense urban scenarios and compared with a rule-based\napproach and results show that the proposed DRQN based driving behavior\ndecision maker outperforms the rule-based approach.\n