• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2019, Vol. 55 ›› Issue (18): 86-94.doi: 10.3901/JME.2019.18.086

• 运载工程 • 上一篇    

基于纵向力伪测量的分布式驱动电动汽车行驶状态估计

陈特1, 陈龙1,2, 蔡英凤1,2, 徐兴1,2, 江浩斌1,2   

  1. 1. 江苏大学汽车与交通工程学院 镇江 212013;
    2. 江苏大学汽车工程研究院 镇江 212013
  • 收稿日期:2018-08-25 修回日期:2019-02-20 发布日期:2020-01-07
  • 通讯作者: 陈龙(通信作者),男,1958年出生,教授,博士研究生导师。主要研究方向为车辆动态性能模拟与控制。E-mail:chenlong@ujs.edu.cn
  • 作者简介:陈特,男,1992年出生,博士研究生。主要研究方向为车辆行驶状态估计与动力学控制。E-mail:ujschente@163.com
  • 基金资助:
    国家自然科学基金(U1564201,U1664258,51875255)、国家重点研发计划(2017YFB0102603)、江苏省“六大人才高峰”(2018-TD-GDZB-022)、江苏省自然科学基金(BK20160525)和江苏省重点研发计划竞争(BE2017129)资助项目。

Estimation of Driving States Based on Pseudo-measurements of Longitudinal Force for Distributed Drive Electric Vehicles

CHEN Te1, CHEN Long1,2, CAI Yingfeng1,2, XU Xing1,2, JIANG Haobin1,2   

  1. 1. School of Automotive and Traffic Engineering, Jiangsu University, Zhenjiang 212013;
    2. Automotive Engineering Research Institute, Jiangsu University, Zhenjiang 212013
  • Received:2018-08-25 Revised:2019-02-20 Published:2020-01-07

摘要: 精确可靠的状态估计是车辆主动安全控制的必要因素之一,提出一种纵、横向加速度传感器信息缺失情况下的车辆状态补偿估计方法。建立了3自由度车辆模型与轮胎模型,提出电驱动轮模型并将其应用到纵向力估计中,考虑电驱动轮模型含有噪声和未知输入,通过模型解耦的方式得到了纵向力重构方程,并基于伦伯格观测器和高阶滑模观测器实现纵向力估计。将纵向力估计作为伪量测值,设计了一种传感器信息不足情况下的补偿估计方法,并基于强跟踪滤波实现车辆状态估计。联合仿真结果表明,所设计的纵向力观测器针对含未知输入和干扰的系统能够实时估计纵向力,采用补偿和强跟踪结合的方式能够有效提高估计精度。考虑估计方法的实车表现,进行了台架和道路测试,台架试验结果表明纵向力观测器估计精度达到了91.3%,道路试验结果表明STF相比EKF对纵向车速、侧向车速以及横摆角速度的估计精度分别提高了14.03%,15.02%和16.58%。

关键词: 电动汽车, 分布式驱动, 纵向力估计, 车辆状态

Abstract: Accurate and reliable estimation of vehicle state is one of the essential factors of vehicle active safety control, a method of vehicle state compensation estimation for the case of longitudinal and lateral acceleration sensor information shortage is proposed. The vehicle model with three degree of freedom and the tire model are established, an electric drive wheel model is proposed and applied to longitudinal force estimation, the longitudinal force reconstruction equation is obtained by model decoupling with the consideration of the noise and unknown input in electric drive wheel model, and then the longitudinal force estimation is achieved based on the combination of Luenberger observer and high order sliding mode observer. The longitudinal force estimation is regarded as the pseudo-measurement, a compensation estimation method under the condition of insufficient sensor information is designed, and the vehicle state estimation is obtained on the basis of strong tracking filter. The joint simulation results show that the designed longitudinal force observer can estimate the longitudinal force in real time for the system with unknown inputs and disturbances, the combination of compensation method and strong tracking filter can effectively improve the estimation accuracy. Considering the real vehicle performance of the estimation method, the bench and road tests are carried out, the results of bench test show that the estimation accuracy of the proposed longitudinal force observer is 91.3%, and the results of the road test indicate that, compared with EKF, the accuracy of STF in the estimation of longitudinal vehicle speed, lateral vehicle speed and yaw rate increase by 14.03%, 15.02% and 16.58% respectively.

Key words: electric vehicle, distributed drive, longitudinal force estimation, vehicle state

中图分类号: