JOURNAL OF CHINA UNIVERSITIES OF POSTS AND TELECOM ›› 2018, Vol. 25 ›› Issue (6): 58-64.doi: 10.19682/j.cnki.1005-8885.2018.1027

• Artificial intelligence • Previous Articles     Next Articles

Steering control in autonomous vehicles using deep reinforcement learning

Xue Chong, Jia Peng, Zhang Xinyu   

  1. 1. Department of Computer Science and Technique, Tsinghua University, Beijing 300401, China
    2. Joint Logistics College, Beijing 100858, China
  • Received:2018-08-07 Revised:2019-01-04 Online:2018-12-30 Published:2019-02-26
  • Contact: Jia Peng, E-mail: jiapeng1018@163.com E-mail:jiapeng1018@163.com
  • About author:Jia Peng, E-mail: jiapeng1018@163.com

Abstract: A novel deep reinforcement learning-based steering control method of autonomous vehicles is proposed. A distortionless compressing method of action space is presented. Convolutional neural networks (CNNs) are designed to serve as an action policy. Driver experience is investigated and modeled to optimize policy of new actions exploration. Experimental results show that the proposed algorithm has better robustness and smoothness. Moreover, it is applicable to different roads, velocities or wire-control systems.

Key words: autonomous vehicles, deep reinforcement learning, steering control, CNNs

CLC Number: