中国机械工程 ›› 2022, Vol. 33 ›› Issue (19): 2381-2387.DOI: 10.3969/j.issn.1004-132X.2022.19.013

• 工程前沿 • 上一篇    下一篇

基于多传感器信息融合的车辆高精度定位技术

师小波1,4; 赵丁选2;孔志飞5;倪涛3;赵小龙2;郭庆贺2   

  1. 1.燕山大学电气工程学院,秦皇岛,066004
    2.燕山大学机械工程学院,秦皇岛,066004
    3.燕山大学车辆与能源学院,秦皇岛,066004
    4.河北建材职业技术学院,秦皇岛,066004
    5.吉林大学机械与航空航天工程学院,长春,130022
  • 出版日期:2022-10-10 发布日期:2022-10-20
  • 通讯作者: 倪涛(通信作者),男,1978年生,教授。研究方向为智能驾驶、液压悬架控制。E-mail:nitao@jlu.edu.cn。
  • 作者简介:师小波,男,1982年生,博士研究生。研究方向为车辆悬架控制、智能机器人。
  • 基金资助:
    河北省创新群体项目(E2020203174);国家自然科学基金(U20A20332)

Vehicle High-precision Positioning Technique Based on Multi-sensors Information Fusion

SHI Xiaobo1,4;ZHAO Dingxuan2;KONG Zhifei5;NI Tao3;ZHAO Xiaolong2;GUO Qinghe2   

  1. 1.School of Electrical Engineering,Yanshan University,Qinhuangdao,Hebei,066004
    2.School of Mechanical Engineering,Yanshan University,Qinhuangdao,Hebei,066004
    3.School of Vehicle and Energy,Yanshan University,Qinhuangdao,Hebei,066004
    4.Hebei Construction Material Vocational and Technical College,Qinhuangdao,Hebei,066004
    5.School of Mechanical and Aerospace Engineering,Jilin University,Changchun,130022
  • Online:2022-10-10 Published:2022-10-20

摘要: 为解决多轴应急救援车辆在复杂环境行驶时的定位误差问题,设计了一种基于多信息融合的组合定位系统,该定位系统主要由前端里程计和后端优化模型组成。首先完成三轴车运动学建模,推导出基于惯性测量单元的运动学预测方程和基于全球定位系统的运动学观测方程,通过Kalman滤波器进行状态更新,完成前端里程计的构建;然后设计了基于迭代最近点匹配算法的外参标定工具,并构建了特征点云匹配模型,完成后端模型优化;最后进行了仿真与实验,研究结果表明,该系统定位误差为±3.45 cm,角度误差为±0.10°,准确性、稳定性均得到了很大的提高。

关键词: 组合定位系统, 前端里程计, 点云匹配, 信息融合

Abstract:  In order to solve the positioning error problems of multi-axle emergency rescue vehicles driving in complex environments, a combined positioning system was designed based on multi-information fusion. The positioning system was mainly composed of a front-end odometer and a back-end optimization model. The three-axis vehicle kinematics modeling was completed, the inertial measurement unit-based kinematics prediction equation and global positional system-based kinematics observation equation were derived, the state was updated through the Kalman filter, and the construction of the front-end odometer was completed. The external reference calibration tool was designed based on iterative closest point matching algorithm, and a feature point cloud matching model  was constructed to complete the back-end model optimization. Finally, after simulation and experiments were carried out, the results show that the positioning errors of the system are less than 3.45 cm, the angle errors are less than 0.10 degrees, the accuracy and stability are greatly improved. 

Key words: combined positioning system, front-end odometer, point cloud matching, information fusion

中图分类号: