主办:陕西省汽车工程学会
ISSN 1671-7988  CN 61-1394/TH
创刊:1976年

汽车实用技术 ›› 2023, Vol. 48 ›› Issue (24): 189-194.DOI: 10.16638/j.cnki.1671-7988.2023.024.037

• 综述 • 上一篇    

基于学习的自动驾驶换道决策算法研究进展

于士杰,马 冲,陈见哲   

  1. 长安大学 汽车学院
  • 出版日期:2023-12-30 发布日期:2023-12-30
  • 通讯作者: 于士杰
  • 作者简介:于士杰(1999-),男,硕士研究生,研究方向为运载工具运用工程,E-mail:2021122058@chd.edu.cn。

Research Progress of Automatic Driving Lane Change Decision Algorithms Based on Learning

YU Shijie, MA Chong, CHEN Jianzhe   

  1. School of Automobile, Chang'an University
  • Online:2023-12-30 Published:2023-12-30
  • Contact: YU Shijie

摘要: 换道决策是自动驾驶决策规划技术关键一环,研究自动驾驶汽车的换道决策算法,可 以更好的实现车辆的自主换道。文章通过对现有的换道策略文献进行查阅,对近年来基于学 习的自动驾驶换道决策算法进行分类和总结。首先将基于学习的换道决策方法分为两类,即 基于传统机器学习和基于强化学习的换道决策,其次对决策算法进行详细说明与研究综述, 分析各个方法的优劣,最后对自动驾驶换道决策未来研究方向进行展望,为车辆自主换道决 策研究提供参考。

关键词: 换道决策;自动驾驶;机器学习;强化学习

Abstract: Lane change decision is a key part of autonomous driving decision planning technology. Studying the lane change decision algorithms of autonomous vehicles can better realize the autonomous lane change of vehicles. In this paper, by consulting the existing literature on lane change strategies, the learning-based lane change decision-making algorithms for autonomous driving in recent years are classified and summarized. Firstly, the learning-based lane change decision-making methods are divided into two categories, based on traditional machine learning and based on reinforcement learning. Secondly, the decision-making algorithms are described in detail and the research is reviewed, and the advantages and disadvantages of each method are analyzed. Finally, the future research direction of autonomous driving lane change decision-making is prospected, which provides a reference for the research of vehicle autonomous lane change decisionmaking.

Key words: Lane change decision; Automatic driving; Machine learning; Reinforcement learning