现代制造工程 ›› 2025, Vol. 541 ›› Issue (10): 82-88.doi: 10.16731/j.cnki.1671-3133.2025.10.009

• 车辆工程制造技术 • 上一篇    下一篇

基于模糊LQR的自动驾驶车辆路径跟踪优化

程广伟1,2, 赵小康1, 张子扬1, 卢艳阳2, 郭占正1   

  1. 1 河南科技大学车辆与交通工程学院,洛阳 471003;
    2 洛阳理工学院智能制造学院,洛阳 471023
  • 收稿日期:2024-11-06 发布日期:2025-10-29
  • 通讯作者: 程广伟,博士研究生,教授,主要从事车辆自动驾驶技术、现代车辆测试技术及装备研究工作。E-mail:2066814401@qq.com
  • 作者简介:赵小康,硕士研究生,主要从事车辆自动驾驶技术研究工作。
  • 基金资助:
    河南省高等教育机构重点研究项目(24A413006);河南省科技攻关项目(232102240022)

Path tracking optimization for autonomous vehicles based on fuzzy LQR

CHENG Guangwei1,2, ZHAO Xiaokang1, ZHANG Ziyang1, LU Yanyang2, GUO Zhanzheng1   

  1. 1 School of Vehicle and Traffic Engineering, Henan University of Science and Technology, Luoyang 471003, China;
    2 School of Intelligent Manufacturing, Luoyang Institute of Science and Technology, Luoyang 471023, China
  • Received:2024-11-06 Published:2025-10-29

摘要: 为了提高自动驾驶车辆的路径跟踪精度,提出了一种基于模糊线性二次型调节器(Linear Quadratic Regulator,LQR)的自动驾驶车辆路径跟踪控制方法。首先,在建立车辆二自由度动力学模型和路径跟踪误差模型的基础上,设计了LQR路径跟踪控制器;然后,针对误差权重系数固定的LQR路径跟踪控制器对多变行驶工况适应性较差的问题,利用模糊控制算法对LQR路径跟踪控制器的误差权重系数进行自适应调节,以达到对路径的精确跟踪;最后,通过CarSim软件和Simulink软件进行联合仿真,结果表明,与LQR路径跟踪控制相比,模糊LQR路径跟踪控制的最大侧向误差减小了53.3 %,最大航向角误差减小了46.9 %;通过Apollo Advanced试验平台验证了所提方法的有效性,该研究结果可为自动驾驶车辆路径跟踪控制器的设计与优化提供创新思路。

关键词: 自动驾驶车辆, 路径跟踪, 线性二次型调节器控制器, 模糊控制, 误差权重系数

Abstract: In order to improve the path tracking accuracy of autonomous vehicles, a path tracking control method for autonomous vehicles based on fuzzy Linear Quadratic Regulator (LQR) was proposed. Firstly, based on the establishment of a two degree of freedom vehicle dynamics model and a path tracking error model, an LQR path tracking controller was designed. Then, in response to the problem of poor adaptability of the LQR path tracking controller with a fixed error weight coefficient to variable driving conditions, a fuzzy control algorithm was used to adaptively adjust the error weight coefficient of the LQR path tracking controller to achieve accurate path tracking. Finally, through the joint simulation of CarSim software and Simulink software, the results showed that compared with LQR path tracking control, the maximum lateral error of the fuzzy LQR path tracking control was reduced by 53.3 %, and the maximum heading angle error was reduced by 46.9 %. The effectiveness of the proposed method was verified through the Apollo Advanced test platform, and the research results can provide innovative ideas for the design and optimization of path tracking controllers for autonomous vehicles.

Key words: autonomous vehicles, path tracking, linear quadratic regulator controller, fuzzy control, error weight coefficient

中图分类号: 

版权所有 © 《现代制造工程》编辑部 
地址:北京市东城区东四块玉南街28号 邮编:100061 电话:010-67126028 电子信箱:2645173083@qq.com
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn