汽车工程 ›› 2025, Vol. 47 ›› Issue (5): 951-962.doi: 10.19562/j.chinasae.qcgc.2025.05.015

• • 上一篇    下一篇

基于腿部sEMG的驾驶疲劳状态判别方法

俞宁1(),罗晓茗1,舒梓荣1,李博远2,张颜3   

  1. 1.重庆理工大学机械工程学院,重庆 400054
    2.高机动防暴车辆技术国家工程研究中心,北京 100072
    3.重庆邮电大学计算机科学与技术学院,重庆 400065
  • 收稿日期:2024-11-20 修回日期:2025-01-09 出版日期:2025-05-25 发布日期:2025-05-20
  • 通讯作者: 俞宁 E-mail:yuning@cqut.edu.cn
  • 基金资助:
    高机动防暴车辆技术国家工程研究中心开放基金(2023NELEV003)

A Discriminative Method for Driving Fatigue State Based on Leg sEMG

Ning Yu1(),Xiaoming Luo1,Zirong Shu1,Boyuan Li2,Yan Zhang3   

  1. 1.School of Mechanical Engineering,Chongqing University of Technology,Chongqing 400054
    2.National Engineering Research Center for High Mobility Anti-riot Vehicle Technology,Beijing 100072
    3.School of Computer Science and Technology,Chongqing University of Posts and Telecommunications,Chongqing 400065
  • Received:2024-11-20 Revised:2025-01-09 Online:2025-05-25 Published:2025-05-20
  • Contact: Ning Yu E-mail:yuning@cqut.edu.cn

摘要:

基于驾驶员腿部的表面肌电信号提出一种非侵入式的驾驶疲劳状态判别方法。首先,通过模拟驾驶疲劳实验采集驾驶员右腿胫骨前肌的肌电信号,并通过主观评价量表进行疲劳状态的标注。其次,采用变分模态分解算法对表面肌电信号进行噪声滤除,并从分解得到的5个IMF分量中提取12个时频域特征值。最后,构建基于鲸鱼算法优化支持向量机的驾驶疲劳状态判别模型。结果表明:该方法对于3种疲劳状态具有较好的判别效果,其准确率可达84%以上。

关键词: 驾驶疲劳, 表面肌电信号, 疲劳状态判别

Abstract:

A non-invasive driving fatigue state identification method based on the surface electromyographic signals of the driver's legs is proposed. Firstly, the electromyographic signal of the tibialis anterior muscle of the driver's right leg is collected through a simulated driving fatigue experiment, and the fatigue status is marked through a subjective evaluation scale. Secondly, a variational mode decomposition algorithm is used to filter out noise on the surface electromyographic signal, and 12 time-frequency domain eigenvalues ??are extracted from the five IMF components obtained by decomposition. Finally, a driving fatigue state discrimination model based on whale algorithm optimized support vector machine is constructed. The results show that this method has a good discrimination effect on three fatigue states, with an accuracy of more than 84%.

Key words: driving fatigue, surface electromyographic signals, fatigue state classification