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): 25-31.DOI: 10.16638/j.cnki.1671-7988.2023.024.006

• 智能网联汽车 • 上一篇    

基于知识图谱的自动驾驶算法研究热点分析

韩文思 1,孟瑞锋*1,乔 志 2,沈宝泉 2,王 飞 3   

  1. 1.内蒙古工业大学 航空学院;2.内蒙古高速公路集团有限责任公司;3.北奔重汽(北京)汽车研发有限公司
  • 出版日期:2023-12-30 发布日期:2023-12-30
  • 通讯作者: 孟瑞锋
  • 作者简介:韩文思(1997-),女,硕士研究生,研究方向为自动驾驶路径规划,E-mail:1559574874@qq.com。
  • 基金资助:
    内蒙古自治区科技计划项目(2022YFSJ0040);内蒙古自治区直属高校基本科研业务费项目(JY20220170)

Analysis of Hot Spots of Autonomous Driving Algorithm Based on Knowledge Graph

HAN Wensi1 , MENG Ruifeng*1 , QIAO Zhi2 , SHEN Baoquan2 , WANG Fei3   

  1. 1.School of Aeronautics, Inner Mongolia University of Technology;2.Inner Mongolia Expressway Group Company Limited;3.Beiben Heavy Truck (Beijing) Automobile R&D Company Limited
  • Online:2023-12-30 Published:2023-12-30
  • Contact: MENG Ruifeng

摘要: 自动驾驶算法作为感知、决策规划和控制模块中必不可少的一部分,对自动驾驶技术 的发展起着至关重要的作用。文章对自动驾驶算法研究领域的文献进行了梳理和研究,围绕 “自动驾驶+算法”主题,对中国知网(CNKI)数据库和 WOS 核心合集数据库筛选的 1 914 篇文献数据进行处理,分别导入 CiteSpace 软件绘制关键词聚类、共现和突现知识图谱,展现 自动驾驶算法的研究脉络并进行针对性分析。研究结果表明,遗传算法、神经网络、深度强 化学习等启发式算法是重要的研究主题,单目视觉检测、端到端控制策略模型和自动驾驶车 辆在复杂工况下的学习、模仿和选择是未来自动驾驶算法领域研究的重点,自动驾驶多种算 法融合和多传感器融合将会是未来发展的重要方向。

关键词: 自动驾驶算法;CiteSpace;知识图谱;文献计量分析

Abstract: As an indispensable part of perception, decision planning and control module, autonomous driving algorithm plays a vital role in the development of autonomous driving technology. Focusing on the theme of "autonomous driving+algorithm", this paper reviews and studies the literatures in the field of autonomous driving algorithm, and the data of 1 914 literatures screened by China national knowledge Internet (CNKI) database and web of science (WOS) core collection database are processed, and CiteSpace software is introduced to draw keyword clustering, co-occurrence and emergence knowledge maps which shows the research context of automatic driving algorithm and conduct targeted analysis. The results show that heuristic algorithms such as genetic algorithm, neural network and deep reinforcement learning are important research topics,and monocular vision detection, end-to-end control strategy model and learning, imitation and selection of autonomous vehicles under complex working conditions will be the focus of future research in the field of autonomous driving algorithms, and fusion of multiple autonomous driving algorithms and multisensor fusion will be an important direction of future development.

Key words: Autonomous driving algorithm; CiteSpace; Knowledge graph; Ibliometric analysis