A smart bracelet-based method for dynamic detection of angry driving behaviors

doi:10.16265/j.cnki.issn1003-3033.2022.S1.0644

Abstract

Abstract:

In order to prevent traffic accidents caused by angry driving, 18 professional drivers were employed for real-world tests where they their electrocardiogram indicators were collected through the smart bracelets they wore. Then, it was found in statistical tests that electrocardiogram indicators heart rate (HR), mean of RR intervals (RRmean), standard deviation of RR intervals (SDNN), root mean square of continuous difference (RMSSD), number of RR intervals greater than 50ms (PNN50), high frequency (HF) and non-linear indicators (SD1, SD2, SD2/SD1) were significantly different. Finally, with tertiary angry driving behaviors (normal, mildly angry, and strongly angry) and secondary ones (normal and angry) as dependent variables and significantly different electrocardiogram indicators as independent ones respectively, a dynamic detection model of drivers' angry driving behavior was developed based on support vector machine (SVM), K-nearest neighbor (KNN) and linear analysis (LD). The results show that the model's recognition effect for secondary angry driving behaviors was greatly better than that for third-level behavior recognition, and the SVM presents the best performance in the second-level angry driving behavior recognition model, while LD model is the best in the third-level recognition one.

Key words: smart bracelet, angry driving, testing index, electrocardiogram signal, emotion recognition

NIU Shifeng, MA Bintao, LIU Yanjun, HAO Shuaijie. A smart bracelet-based method for dynamic detection of angry driving behaviors[J]. China Safety Science Journal, 2022, 32(S1): 178-183.

Figures/Tables 7

Fig.1

Fig.2

Tab.1

Tab.2

Tab.3

Tab.4

Tab.5

References 15

[1]	DEFFENBACHER J L, OETTING E R, LYNCH R S. Development of a driving angry scale[J]. Psychological Report, 1994, 74(1): 83-91. doi: 10.2466/pr0.1994.74.1.83
[2]	SULLMAN M J, STEPHENS A N, YONG M. Driving anger in malaysia[J]. Accident Analysis and Prevention, 2014, 10(7): 1-9. doi: 10.1016/0001-4575(78)90002-7
[3]	吴超仲, 雷虎. 汽车驾驶愤怒情绪研究现状与展望[J]. 中国安全科学学报, 2010, 20(7): 3-8.
	WU Chaozhong, LEI Hu. Current status and outlook of automobile driving anger research[J]. China Safety Science Journal, 2010, 20(7): 3-8.
[4]	郑华荣, 吴超仲, 马晓凤. 考虑驾驶愤怒的元胞自动机交通流模型[J]. 武汉理工大学学报:交通科学与工程版, 2013, 37(3): 617-621.
	ZHENG Huarong, WU Chaozhong, MA Xiaofeng. Cellular automaton model for urban traffic flow considering driving anger[J]. Journal of Wuhan University of Technology: Transportation Science&Engineering, 2013, 37(3): 617-621.
[5]	PASCHERO M, VESCOVO G, BENUCCI L, et al. Real time classifier for emotion and stress recognition in a vehicle driver[C]. IEEE International Symposium on Industrial Electronics, 2012: 1 690-1 695.
[6]	KAMARUDDIN N, WAHAB A. Driver behavior analysis through speech emotion understanding[C]. IEEE Intelligent Vehicles Symposium University of California, 2010: 238-243.
[7]	COULSON M. Attributing emotion to static body postures: recognition accuracy,confusions,and viewpoint dependence[J]. Journal of Nonverbal Behavior, 2004, 28(2): 117-139. doi: 10.1023/B:JONB.0000023655.25550.be
[8]	张迪, 万柏坤, 明东. 基于生理信号的情绪识别研究进展[J]. 生物医学工程学杂志, 2015, 32(1): 229-234.
	ZHANG Di, WAN Baikun, MING Dong. Advances in research on emotion recognition based on physiological signals[J]. Journal of Biomedical Engineering, 2015, 32(1): 229-234.
[9]	赵昱音. 面向可穿戴式设备的情绪识别研究[D]. 哈尔滨: 哈尔滨工业大学, 2021.
	ZHAO Yuyin. Research on emotion recognition for wearable devices[D]. Harbin: Harbin Institute of Technology, 2021.
[10]	李英春, 尤磊, 贺靖康. 基于生理信号的情绪识别腕戴设备[J]. 电子技术应用, 2017, 43(2): 69-72,76.
	LI Yingchun, YOU Lei, HE Jingkang. Emotion recognition wrist-worn device based on physiological signals[J]. Electronic Technology Applications, 2017, 43(2): 69-72,76.
[11]	吴坤, 陈向东. 基于可穿戴设备的人体情绪识别系统研究[J]. 单片机与嵌入式系统应用, 2019, 19(5): 3-6,10.
	WU Kun, CHEN Xiangdong. Research on human emotion recognition system based on wearable devices[J]. Microcontroller and Embedded System Applications, 2019, 19(5): 3-6,10.
[12]	PICARD R W. Toward computers that recognize and respond to user emotion[J]. IBM Systems Journal, 2000, 39(3): 705-719. doi: 10.1147/sj.393.0705
[13]	KATSIS C D, KATERTSIDIS N, GANIATSAS G, et al. Toward emotion recognition in car-racing drivers: a biosignal processing approach[J]. IEEE Transactions on Systems, Man, and Cybernetics-Part A: Systems and Humans, 2008, 38(3): 502-512. doi: 10.1109/TSMCA.2008.918624
[14]	KREIBIG S D. Autonomic nervous system activity in emotion: a review[J]. Biological Psychology, 2010, 84(3): 394-421. doi: 10.1016/j.biopsycho.2010.03.010 pmid: 20371374
[15]	CAMM A J, MALIK M, BIGGER J T, et al. Heart rate variability·standards of measurement, physiological interpretation and clinical use. Task force of the European society of cardiology and the North American society of pacing and electro physiology[J]. Circulation, 1996, 93: 1 043-1 065. doi: 10.1161/01.CIR.93.5.1043

特征类型	特征名称与定义	单位
心率	HR:心率	次/min
时域特征	RRmean:RR间期均值	ms
	SDNN:RR间期标准差	ms
	RMSSD:RR间期差值均方根	ms
	PNN50:RR间期大于50ms的个数	次
频域特征	LF(低频功率):功率谱密度曲线函数在0.04~0.15 Hz范围内的积分值	ms²
频域特征	HF(高频功率):功率谱密度曲线函数在0.15~0.4 Hz范围内的积分值	ms²
非线性指标	SD1:垂直于标识线点的标准偏差	—
	SD2:同一条直线的标准偏差	—
	SD2/SD1:横纵标准偏差比值	—

指标	卡方	自由度	显著性
HR/(次·min^-1)	216.093	2	0.000
RRmean/ms	216.891	2	0.000
SDNN/ms	28.395	2	0.000
RMSSD/ms	73.569	2	0.000
PNN50/次	133.635	2	0.000
LF	1.979	2	0.372
HF	89.099	2	0.000
SD1	68.64	2	0.000
SD2	40.772	2	0.000
SD2/SD1	36.407	2	0.000

指标	正常情绪- 轻微愤怒	正常情绪- 强烈愤怒	轻微愤怒- 强烈愤怒
HR/(次·min^-1)	0.000	0.000	0.000
RRmean/ms	0.000	0.000	0.000
SDNN/ms	0.000	0.000	0.829
RMSSD/ms	0.000	0.000	0.890
PNN50/次	0.000	0.000	0.673
HF	0.000	0.000	0.195
SD1	0.000	0.000	0.056
SD2	0.000	0.000	0.697
SD2/SD1	0.000	0.000	0.256

指标	SVM			KNN			LD
指标	轻微	正常	强烈	轻微	正常	强烈	轻微	正常	强烈
TPR	0.563	0.974	0.954	0.563	0.944	0.943	0.531	0.969	0.977
PR	0.667	0.946	0.965	0.563	0.939	0.854	0.773	0.945	0.955
F1	0.610	0.960	0.960	0.563	0.941	0.896	0.630	0.957	0.966
FPR	0.032	0.098	0.014	0.050	0.107	0.065	0.018	0.097	0.019

指标	SVM		KNN		LD
指标	二级愤怒	三级愤怒	二级愤怒	三级愤怒	二级愤怒	三级愤怒
TPR	0.990	0.830	0.995	0.816	0.969	0.826
FPR	0.118	0.048	0.185	0.074	0.101	0.045
PR	0.933	0.859	0.899	0.785	0.941	0.891
F1	0.960	0.843	0.944	0.800	0.955	0.851
AUC	0.980	0.970	0.980	0.920	0.990	0.970
ACC	0.949	0.827	0.927	0.905	0.943	0.924