交通运输系统工程与信息 ›› 2020, Vol. 20 ›› Issue (4): 77-82.

• 智能交通系统与信息技术 • 上一篇    下一篇

基于车辆运行数据的疲劳驾驶状态检测

蔡素贤1 ,杜超坎1 ,周思毅1 ,王雅斐*2   

  1. 1. 汉纳森(厦门)数据股份有限公司, 福建 厦门 361000:2. 上海财经大学 信息管理与工程学院, 上海 200433
  • 收稿日期:2020-02-28 修回日期:2020-05-12 出版日期:2020-08-25 发布日期:2020-08-25
  • 作者简介:蔡素贤(1988-),女,福建厦门人,工程师.
  • 基金资助:

    国家自然科学基金/National Natural Science Foundation of China(61773248);国家社科基金重大项目/ Major Program of the National Social Science Foundation of China(18ZDA088).

Fatigue Driving State Detection Based on Vehicle Running Data

CAI Su-xian1 , DU Chao-kan1 , ZHOU Si-yi1 , WANG Ya-fei2   

  1. 1. Honorsun(Xiamen) Data Company Limited by Shares, Xiamen 361000, Fujian, China; 2. School of Information Management & Engineering, Shanghai University of Finance and Economics, Shanghai 200433, China
  • Received:2020-02-28 Revised:2020-05-12 Online:2020-08-25 Published:2020-08-25

摘要:

疲劳驾驶是导致交通事故的主要原因之一,及时检测疲劳驾驶,并提醒驾驶员集中注意力,对保证安全行车具有重要意义.本文基于CAN(Controller Area Network)总线采集的车辆运行状态数据,提取了18项与驾驶行为相关的特征,并采用随机森林算法对疲劳驾驶进行识别,结果表明整体的识别准确率为0.785,其中召回率为0.61,即61%的疲劳驾驶状态可被识别出来.实验表明,基于车辆运行状态的疲劳驾驶检测具有一定的效果,且与其他客观的疲劳驾驶检测方法(基于驾驶员生理指标和图像面部特征)相比,具有简单方便,不影响驾驶,且成本低的优势.

关键词: 智能交通, 疲劳检测, 随机森林, CAN数据

Abstract:

Fatigue driving is one of the main causes of traffic accidents. It is of great importance to detect fatigue driving dynamically and remind drivers to concentrate on driving safely. Based on the vehicle running data collected by Controller Area Network (CAN) bus, this paper extracts 18 features relevant to driving behaviors and uses random forest algorithm to identify fatigue driving. The results show that the overall recognition accuracy is 0.785, and the recall rate is 0.61 which means 61% of fatigue driving conditions can be successfully identified. Experiments show that fatigue driving detection based on vehicle running data is effective. Compared with other fatigue driving detection methods (for example, based on driver physiological indicators and image facial features), the proposed method is simple and convenient, without affecting driver's operations and the cost is relatively low

Key words: intelligent transportation, fatigue detection, random forest, CAN data

中图分类号: