基于心电的铁路列车驾驶压力检测研究

doi:10.16265/j.cnki.issn1003-3033.2022.06.2406

中国安全科学学报 ›› 2022, Vol. 32 ›› Issue (6): 31-37.doi: 10.16265/j.cnki.issn1003-3033.2022.06.2406

• 安全社会科学与安全管理 • 上一篇下一篇

基于心电的铁路列车驾驶压力检测研究

刘坤(), 焦钰博, 张晓明, 陈晓宇, 蒋朝哲^**()

西南交通大学交通运输与物流学院,四川成都 610031

收稿日期:2022-01-10 修回日期:2022-04-10 出版日期:2022-06-28 发布日期:2022-12-28
通讯作者: 蒋朝哲
作者简介:
刘坤 (1996—),男,河南焦作人,硕士研究生,研究方向为交通领域人因安全等。E-mail: 1325289243@qq.com。
蒋朝哲,副教授
基金资助:
国家自然科学基金资助(71871188); 国家社会科学基金资助(15GBL143)

Test of railway train drivers' stress by using ECG signal

LIU Kun(), JIAO Yubo, ZHANG Xiaoming, CHEN Xiaoyu, JIANG Chaozhe^**()

School of Transportation and Logistics, Southwest Jiaotong University, Chengdu Sichuan 610031, China

Received:2022-01-10 Revised:2022-04-10 Online:2022-06-28 Published:2022-12-28
Contact: JIANG Chaozhe

摘要/Abstract

摘要：

为实时检测列车司机压力水平,基于高铁驾驶行为与安全仿真平台,获取16名被试不同驾驶速度下的主观压力量表和心电(ECG)信号。首先,通过分析主观压力量表,探究不同速度下司机的压力水平;其次,统计分析不同压力水平下的心率变异性(HRV)特征;最后,比较最邻近算法(KNN),支持向量机(SVM),随机森林(RF)3种机器学习算法在压力检测方面的表现,并分析不同输入特征对分类器性能的影响。研究表明:随着速度的增加,司机的压力增大。连续R波之间的时间差间隔大于50 ms的数量(NN50),连续R波之间的时间差间隔大于50 ms的数量占比(PNN50),低频段功率值与高频段功率值之比(LF/HF),心脏交感神经指数(CSI)在不同的压力条件下具有显著性差异。在高速条件下,PNN50、HF、LF/HF减小,NN50增加,其中LF/HF值显著降低。此外,特征选择和特征标准化均有助于提高模型的表现;相比KNN和RF,径向基为核函数(RBF)的SVM分类器模型在检测驾驶压力方面的效果最好,准确度为71.2%。

关键词: 心电(ECG)信号, 驾驶压力, 列车司机, 驾驶速度, 机器学习

Abstract:

In order to test stress level of railway train drivers in real time, stress response inventory and ECG signal data of 16 subjects under different train speeds were collected by using high-speed rail driving behavior and safety simulation platform. Firstly, stress response inventory was analyzed to investigate drivers' stress level along with increasing train speed. Then, HRV features were statistically analyzed at different stress levels. Finally, K nearest neighbor (KNN) algorithm, support vector machine (SVM), and random forest (RF) were compared in testing stress level, and influence of different input characteristics on these classifiers' performance was analyzed. The results show that drivers' pressure will rise along with the increase of speed, and there are significant differences for number of successive normal to normal intervals pairs that differ more than 50 ms(NN50), proportion of number of successive normal to normal intervals more than 50ms(PNN50), ratio of low frequency and high frequency(LF/HF) and cardiac sympathetic index(CSI) between different driving speeds. Besides, PNN50, HF, and LF/HF decrease with increased driving speed, while NN50 increase. In particular, LF/HF decrease significantly. Moreover, feature selection and feature normalization could improve the model's accuracy, and best performance is achieved by SVM classifier model with radial basis function(RBF) as kernel function,with an accuracy of 71.2%.

Key words: electrocardiogram (ECG) signal, drivers'stress, train driver, driving speed, machine learning

刘坤, 焦钰博, 张晓明, 陈晓宇, 蒋朝哲. 基于心电的铁路列车驾驶压力检测研究[J]. 中国安全科学学报, 2022, 32(6): 31-37.

LIU Kun, JIAO Yubo, ZHANG Xiaoming, CHEN Xiaoyu, JIANG Chaozhe. Test of railway train drivers' stress by using ECG signal[J]. China Safety Science Journal, 2022, 32(6): 31-37.

图/表 7

图1

图2

图3

表1

图4

表2

表3

参考文献 21

[1]	RASTGOO M N, NAKISA B, RAKOTONIRAINY A, et al. A critical review of proactive detection of driver stress levels based on multimodal measurements[J]. ACM Computing Surveys (CSUR), 2018, 51(5): 1-35.
[2]	WIBERG H, NILSSON E, LINDÉN P, et al. Physiological responses related to moderate mental load during car driving in field conditions[J]. Biological Psychology, 2015, 108: 115-125. doi: 10.1016/j.biopsycho.2015.03.017
[3]	张琼, 付锐, 秦加合, 等. 基于知信行理论的驾驶人风险行为研究[J]. 中国安全科学学报, 2012, 22(4): 8-13.
	ZHANG Qiong, FU Rui, QIN Jiahe, et al. Study into drivers' risk behavior based on theory of knowledge attitude practice[J]. China Safety Science Journal, 2012, 22(4): 8-13.
[4]	EVANS A W. Fatal train accidents on Europe's railways: 1980-2009[J]. Accident Analysis & Prevention, 2011, 43(1): 391-401. doi: 10.1016/j.aap.2010.09.009
[5]	SANKIM D, YOON W C. An accident causation model for the railway industry: application of the model to 80 rail accident investigation reports from the UK[J]. Safety Science, 2013, 60: 57-68. doi: 10.1016/j.ssci.2013.06.010
[6]	SCHMIDT P, REISS A, DÜRICHEN R, et al. Wearable-based affect recognition: a review[J]. Sensors, 2019, 19(19): Doi: 10.3390/s19194079. doi: 10.3390/s19194079
[7]	ZOER I, SLUITER J K, FRINGS-DRESEN M H W. Psychological work characteristics, psychological workload and associated psychological and cognitive requirements of train drivers[J]. Ergonomics, 2014, 57(10): 1473-1487. doi: 10.1080/00140139.2014.938130
[8]	KAYE S A, LEWIS I, FREEMAN J. Comparison of self-report and objective measures of driving behavior and road safety: a systematic review[J]. Journal of Safety Research, 2018, 65: 141-151. doi: 10.1016/j.jsr.2018.02.012
[9]	HEALEY J A, PICARD R W. Detecting stress during real-world driving tasks using physiological sensors[J]. Transactions on Intelligent Transportation Systems, 2005, 6(2): 156-166.
[10]	郭孜政, 刘仙, 陈瑞雅, 等. 基于ERP技术的疲劳对驾驶人行为监控能力影响研究[J]. 中国公路学报, 2018, 31(8):173-180.
	GUO Zizheng, LIU Xian, CHEN Ruiya, et al. ERP-based study on the influence of fatigue on drivers' action monitoring ability[J]. China Journal of Highway and Transport, 2018, 31(8): 173-180.
[11]	莫秋云, 李荣敬, 李军, 等. 基于ECG指标的山区公路线形对驾驶员特性的影响研究[J]. 中国安全科学学报, 2013, 23(12):16-20.
	MO Qiuyun, LI Rongjing, LI Jun, et al. Research on effect of alignments of mountainous highways on driver characteristics based on ECG index[J]. China Safety Science Journal, 2013, 23(12):16-20.
[12]	KAJIWARA S. Evaluation of driver's mental workload by facial temperature and electrodermal activity under simulated driving conditions[J]. International Journal of Automotive Technology, 2014, 15(1): 65-70. doi: 10.1007/s12239-014-0007-9
[13]	CARDONE D, PERPETUINI D, FILIPPINI C, et al. Driver stress state evaluation by means of thermal imaging: a supervised machine learning approach based on ECG signal[J]. Applied Sciences, 2020, 10(16): DOI: 10.3390/app10165673.
[14]	GILGEN R, SCHWEIZER T, WYSS T. RR interval signal quality of a heart rate monitor and an ECG holter at rest and during exercise[J]. European Journal of Applied Physiology, 2019, 119(7): 1525-1532. doi: 10.1007/s00421-019-04142-5
[15]	KOH K B, PARK J K, KIM C H, et al. Development of the stress response inventory and its application in clinical practice[J]. Psychosomatic Medicine, 2001, 63(4): 668-678. doi: 10.1097/00006842-200107000-00020
[16]	UPGANLAWAR I V, CHOWHAN H. Pre-processing of ECG signals using filters[J]. International Journal of Computer Trends and Technology (IJCTT), 2014, 11(4): 166-168.
[17]	PAN J, TOMPKINS W J. A real-time QRS detection algorithm[J]. Transactions on Biomedical Engineering, 1985, 3: 230-236.
[18]	TIWARI A, CASSANI R, GAGNON J F, et al. Prediction of stress and mental workload during police academy training using ultra-short-term heart rate variability and breathing analysis[C]. Engineering in Medicine & Biology Society (EMBS), 2020: 4530-4533.
[19]	CINAZ B, ARNRICH B, MARCA R, et al. Monitoring of mental workload levels during an everyday life office-work scenario[J]. Personal and Ubiquitous Computing, 2013, 17(2): 229-239. doi: 10.1007/s00779-011-0466-1
[20]	MAGANA V C, SCHERZ W D, SEEPOLD R, et al. The effects of the driver's mental state and passenger compartment conditions on driving performance and driving stress[J]. Sensors, 2020, 20(18): Doi: 10.3390/s20185274. doi: 10.3390/s20185274
[21]	BELGIU M, DRǍGUŢ L. Random forest in remote sensing: a review of applications and future directions[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016, 114: 24-31. doi: 10.1016/j.isprsjprs.2016.01.011

HRV指标	原假设	P值
SDNN	X{stress = 0} = X{stress = 1}	0.831
RMSSD	X{stress = 0} = X{stress = 1}	0.067
NN50	X{stress = 0} = X{stress = 1}	0.048
PNN50	X{stress = 0} = X{stress = 1}	0.044
VLF	X{stress = 0} = X{stress = 1}	0.231
LF	X{stress = 0} = X{stress = 1}	0.230
HF	X{stress = 0} = X{stress = 1}	0.062
LF/HF	X{stress = 0} = X{stress = 1}	0.005
SD1	X{stress = 0} = X{stress = 1}	0.060
SD2	X{stress = 0} = X{stress = 1}	0.978
CSI	X{stress = 0} = X{stress = 1}	0.048
CVI	X{stress = 0} = X{stress = 1}	0.165

分类器算法	HRV全部特征	特征选择	标准化的HRV全部特征	标准化的特征选择
KNN(K=1)	46.5±1.2(43.2~50.7)	57.8±1.1(52.3~60.1)	51.4±1.6(48.8~54.0)	54.9±2.0(50.0~58.3)
KNN(K=3)	48.1±1.9(46.7~53.3)	63.6±2.0(59.1~67.4)	52.1±3.8(49.8~63.5)	66.4±3.9(59.4~72.1)
KNN(K=5)	49.8±1.5(48.1~52.9)	62.5±1.8(55.1~63.2)	53.3±2.3(50.8~61.1)	59.3±1.5(57.5~61.8)
RF	65.9±2.6(60.3~71.3)	68.8±2.4(65.3~73.9)	64.8±1.5(61.3~66.7)	70.4±2.0(66.3~72.1)
SVM(LKF)	58.1±2.2(55.4~62.4)	63.3±3.0(60.4~72.8)	61.1±3.3(56.4~69.3)	65.8±2.4(61.3~70.5)
SVM(RBF)	63.1±1.2(60.7~66.3)	68.9±1.9(66.5~74.4)	65.2±2.4(62.5~71.8)	71.2±4.0(64.1~78.5)

基于心电的铁路列车驾驶压力检测研究

Test of railway train drivers' stress by using ECG signal

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 21

相关文章 5

编辑推荐

Metrics

本文评价

[1]	刘全义, 朱博, 邓力, 石航, 梁光华. 基于机器学习的双参数火灾探测方法[J]. 中国安全科学学报, 2022, 32(5): 90-96.
[2]	杨文臣, 周燕宁, 田毕江, 郭凤香, 胡澄宇. 基于聚类分析和SVM的二级公路交通事故严重度预测[J]. 中国安全科学学报, 2022, 32(5): 163-169.
[3]	王玉化, 戚春华, 朱守林, 解松芳, 赵婷. 驾驶疲劳恢复时间的心电信号分析[J]. 中国安全科学学报, 2017, 27(8): 7-12.
[4]	张海云，王阿川. 基于BP神经网络在森林防火专家系统不确定推理中的应用研究[J]. 中国安全科学学报, 2010, 20(7): 38-.
[5]	胡毅亭，乔小玲，陈网桦，彭金华，刘荣海. 火炸药企业重大事故隐患危险性评估的计算机自学习实现[J]. 中国安全科学学报, 1999, 9(6): 35-.