欢迎访问《汽车安全与节能学报》，
汽车安全与节能学报

JASE ›› 2017, Vol. 08 ›› Issue (01): 38-45.DOI: 10.3969/j.issn.1674-8484.2017.01.004

• 汽车安全 • 上一篇    下一篇

基于行程时间的人车特征动态辨识方法

王晓原1,2，刘亚奇1，张敬磊1   

  1. 1. 山东理工大学 智能交通研究所，淄博 255049，中国；
    2. 清华大学 汽车安全与节能国家重点实验室，北京 100084，中国
  • 收稿日期:2016-07-08 出版日期:2017-03-23 发布日期:2017-03-23
  • 作者简介:王晓原(1970—)，男（汉），山东，教授，E-mail: wangxiaoyuan@sdut.edu.cn
  • 基金资助:

    ： 国家自然科学基金(61074140, 61573009, 51508315, 51608313) ；山东省自然科学基金(ZR2014FM027, ZR2016EL19) ；
    山东省社会科学规划研究项目(14CGLJ27)；汽车安全与节能国家重点实验室开放基金(KF16232)；山东省高等学校科技计划(J15LB07)。

Dynamic identification method of driver and vehicle features based on travel time

WANG Xiaoyuan1,2, LIU Yaqi1, ZHANG Jinglei1   

  1. 1. Institute of Intelligent Transportation, Shandong University of Technology, Zibo 255049, China ；
    2. State Key Laboratory of Automotive Safety and Energy, Tsinghua University, Beijing 100084, China
  • Received:2016-07-08 Online:2017-03-23 Published:2017-03-23

摘要：

实现对驾驶倾向性、汽车类型的动态识别对构建以人为中心的汽车安全驾驶辅助系统具有重要意义。本文以驾驶员隐私保护为基础，利用车载全球定位系统（GPS）捕获的行程时间，建立基于Bayes 决策树的人车特征动态辨识模型，识别汽车不同类型及其驾驶员倾向性。通过设计实车和虚拟驾驶实验分别验证在不同渗透率条件下的人车特征辨识效果，验证结果表明，本文建立的人车特征辨识模型准确率在80%以上，明显优于传统决策树模型；通过设计仿真实验验证了考虑驾驶倾向性的微观仿真和实际情况具有较好相合性，仿真验证结果间接证明本文研究成果的合理性。

关键词: 汽车安全, 人车特征, 行程时间, 隐私保护, 他控技术, Bayes 决策树

Abstract:

It is of great significance for the people-centered safe driving assistant system to realize the dynamic identification of driver’s propensity and vehicle type. Considering the privacy protection, a dynamic identification model for driver and vehicle characteristics was established using Bayesian decision tree. The ravel time obtained by global position system (GPS) data when cars went through an intersection was used to identify the vehicle type and driver’s propensity in the model. The identification effectiveness of human-vehicle characteristics under different permeability conditions was verified by real and virtual driving experiments. The results show that the accuracy rate of the established recognition model was above 80% and the established model was significantly better than the traditional decision tree model. The good consistency between the microscopic simulation considering driving tendency and the actual situation was verified by simulation experiment, and the rationality of the research results was proved indirectly.

Key words: vehicle safety, drivers and vehicles features, travel time, privacy protection, other-control technology, Bayesian decision tree