China Mechanical Engineering ›› 2026, Vol. 37 ›› Issue (2): 498-507.DOI: 10.3969/j.issn.1004-132X.2026.02.024

Previous Articles    

Intelligent Vehicle Road Recognition Considering System Noises and Unknown Load Weights

WANG Jiantao(), YANG Chao, LIU Shuaishuai, ZHANG Lipeng(), WANG Qijun   

  1. School of Vehicle and Energy,Yanshan University,Qinhuangdao,Hebei,066004
  • Received:2024-06-13 Revised:2026-01-14 Online:2026-02-25 Published:2026-03-13
  • Contact: ZHANG Lipeng

考虑系统噪声与载重未知的智能车路况辨识

王建涛(), 杨超, 刘帅帅, 张利鹏(), 王启军   

  1. 燕山大学车辆与能源学院, 秦皇岛, 066004
  • 通讯作者: 张利鹏
  • 作者简介:王建涛,男,1988年生,讲师。研究方向为智能车底盘协同控制。E-mail: wjt@ysu.edu.cn
    张利鹏*(通信作者),男,1979年生,教授、博士研究生导师。研究方向为智能车辆动力学与控制。E-mail: evic2024@163.com
  • 基金资助:
    国家自然科学基金(52272407);国家自然科学基金(U20A20332);中央引导地方科技发展资金(226Z2202G);河北省高等学校科学研究重点项目(ZD2022029);燕山大学基础创新科研培育项目(2023LGQN009)

Abstract:

Based on the existing technology for road recognition derived from vehicle dynamics response, an ARTSF algorithm was introduced to improve the accuracy of road recognition. By incorporating a dynamic statistical estimation component, the algorithm adjusted system model parameters and noise statistical parameters in real time, resolved the issues of reduced model accuracy due to system noises and unknown load weights effectively. Through offline simulations and vehicle testing, the algorithm’s recognition performances on class C roads, pothole roads, and bumpy roads were validated, with recognition accuracy surpassing 90%. The results show that the algorithm has greater adaptability and precision in situations with system noises and unknown load weights. This paper provides a reference for intelligent electric vehicles to recognize road elevation information when driving on unstructured roads.

Key words: intelligent electric vehicle, dynamics response, road recognition, adaptive recursive three-step filtering(ARTSF)

摘要:

在现有基于车辆动力学响应的路面高程信息识别技术的基础上,提出了一种自适应递归三步滤波算法以提高路面高程信息的识别精度。该算法通过引入动态统计估计环节实时调整系统模型参数和噪声统计参数,有效解决了因系统噪声和车辆载重未知导致的模型精度下降问题。通过离线仿真和实车试验,验证了所提算法在C级路面、凹坑路面和凸块路面上的识别效果,识别精度超过90%。结果表明,在存在系统噪声和车辆载重未知的情况下,所提算法具有更强的适应性和准确性,可为智能电动汽车在非结构道路行驶时的路面高程信息识别提供参考。

关键词: 智能电动汽车, 动力学响应, 路面识别, 自适应递归三步滤波

CLC Number: