A Reinforcement Learning Framework for Video Frame-Based Autonomous Car-Following

Car-following theory has received considerable attention as a core component of Intelligent Transportation Systems. However, its application to the emerging autonomous vehicles (AVs) remains an unexplored research area. AVs are designed provide convenient and safe driving by avoiding accidents caused human errors. They require advanced levels recognition other drivers' driving-style. With car-following models, can use their built-in technology understand environment surrounding them make real-time decisions follow vehicles. In this paper, we design end-to-end framework for using automated object detection navigation decision modules. The objective is allow AV another vehicle based on Red Green Blue Depth (RGB-D) frames. We propose employ joint solution involving You Look Once version 3 (YOLOv3) detector identify leader obstacles reinforcement learning (RL) algorithm navigate self-driving vehicle. Two RL algorithms, namely Q-learning Deep have been investigated. Simulation results show convergence developed models investigate efficiency in following leader. It shown that, with video frames only, promising achieved that adopt reasonable behavior.