Q-learning based Reinforcement Learning Approach for Lane Keeping

The paper presents the application of a Q-Iearning reinforcement learning method in the area of vehicle control. The purpose of the research presented is to design an end-to-end behavior control of a kinematic vehicle model placed in a simulated race track environment, by using reinforcement learning approach. The varied trajectory of the track to provide different situations for the agent. The environment sensing model is based on high-level sensor information. Track curvature, lateral position, and relative yaw angle can be reached from the current automotive sensors, such as camera or IMU based systems. The objectives were reached through the definition of a rewarding system, with subrewards and penalties enforcing lane keeping. After the description of the theoretical basis the environment model with the reward function is detailed. Finally the experiments with the learning process are presented and the results of the evaluation are given from some aspects of control quality and safety.