基于多分类Adaboost算法的驾驶人风险感知倾向研究

doi:10.16265/j.cnki.issn1003-3033.2022.04.021

摘要/Abstract

摘要：

为预防和减少道路交通事故,利用多分类Adaboost SAMME算法辨识不同驾驶人的风险感知倾向。首先基于交通冲突分析方法,量化56位驾驶人的风险感知效用值;然后通过KMRTDS驾驶模拟器获取驾驶人在6个风险驾驶情境中的行为表征参数;最后运用线性判别分析(LDA)、Adaboost SAMME算法逐步构建基于驾驶行为数据的驾驶人风险感知倾向分类预测模型,并采用 k折交叉验证法评估该模型的有效性。研究结果表明:所提模型预测准确率达92.9%,可以有效辨识不同驾驶人的风险感知水平,将驾驶人分为安全、激进、复合3种类型。

关键词: Adaboost SAMME算法, 驾驶人, 风险感知倾向, 驾驶行为, 驾驶模拟器

Abstract:

In order to prevent and reduce road traffic accidents, multi-class Adaboost SAMME algorithm was applied to identify risk perception tendency of different drivers. Firstly, perception utility of 56 drivers was quantified based on traffic conflict analysis method. Then, drivers' behavior characterization parameters in 6 risky situations were obtained through KMRTDS driving simulator. Finally, linear discriminant analysis (LDA) and Adaboost SAMME algorithm were employed to gradually construct a classification and prediction model for drivers' risk perception tendency based on driving behavior data, and k-fold cross-validation method was adopted to evaluate the model's effectiveness. The results show that the accuracy of proposed model is up to 92.9%, and it can effectively identify risk perception tendency of different drivers which are divided into three risk perception types, namely safe type, radical type, and compound type.

Key words: Adaboost SAMME algorithm, drivers, risk perception tendency, driving behavior, driving simulator

秦雅琴, 李秋谷, 赵鹏燕, 宝发旺, 谢济铭. 基于多分类Adaboost算法的驾驶人风险感知倾向研究[J]. 中国安全科学学报, 2022, 32(4): 141-147.

QIN Yaqin, LI Qiugu, ZHAO Pengyan, BAO Fawang, XIE Jiming. Research on risk perception tendency of drivers based on multi-class Adaboost algorithm[J]. China Safety Science Journal, 2022, 32(4): 141-147.

图/表 15

表1

图1

表2

图2

表3

图3

图4

表4

驾驶行为数据及特征量

类别	驾驶行为数据名称	特征量
驾驶操作数据	油门踏板位移A/mm	$A ¯ 、 S A 、 A m a x 、 A m i n$
	制动踏板位移B/mm	$B ¯ 、 S B 、 B m a x$
	方向盘转角 $α$ /(°)	$α ¯ 、 S α 、 α m a x 、 α m i n$
纵向运行数据	纵向加速度 $β$ / (m·s^-2)	$β ¯ 、 S β 、 β m a x$ 、β_min
	纵向速度 $λ$ / (km·h^-1)	$λ ¯$ 、S_λ、λ_max、 $λ m i n$
	侧向加速度μ/ (m·s^-2)	$μ ¯$ 、S_μ、μ_max、 $μ m i n$
	侧向速度 $θ$ / (km·h^-1)	$θ ¯$ 、S_θ、θ_max、 $θ m i n$
横向运行数据	横向位移 $τ$ /m	$τ ¯$ 、S_ω、τ_max、 $τ m i n$

表4

图5

图6

图7

表5

表6

表7

图8

参考文献 12

[1]	MIYAJI M, DANNO M, OGURI K. Analysis of driver behavior based on traffic incidents for driver monitor systems[C]. Intelligent Vehicles Symposium, IEEE, 2008:930-935.
[2]	MARTINEZ C M, HEUCKE M, WANG Feiyue, et al. Driving style recognition for intelligent vehicle control and advanced driver assistance: a survey[J]. IEEE Transactions on Intelligent Transportation Systems, 2018, 19(3):666-676. doi: 10.1109/TITS.2017.2706978
[3]	MARTINUSSEN L M, MOLLER M, PRATO C G. Assessing the relationship between the driver behavior questionnaire and the driver skill inventory: revealing sub-groups of drivers[J]. Transportation Research Part F, 2014, 26(26):82-91. doi: 10.1016/j.trf.2014.06.008
[4]	秦雅琴, 梁锺月, 贾现广, 等. 驾驶人风险感知类型预测模型研究[J]. 安全与环境学报, 2019, 19(4):1266-1273.
	QIN Yaqin, LIANG Zhongyue, JIA Xianguang, et al. A predictive and forecast model of the risk perception type of the vehicle drivers[J]. Journal of Safety and Environment, 2019, 19(4):1266-1273.
[5]	CRUNDALL D, CHAPMAN P, TRAWLEY S, et al. Some hazards are more attractive than others: drivers of varying experience respond differently to a different type of hazard[J]. Accident Analysis and Prevention, 2012, 45(1):600-609. doi: 10.1016/j.aap.2011.09.049
[6]	SONG Xiaolin, YIN Yangang, CAO Haotian, et al. The mediating effect of driver characteristics on risky driving behaviors moderated by gender, and the classification model of driver's driving risk[J]. Accident Analysis and Prevention, 2021, 153:1-17.
[7]	艾倩楠. 基于驾驶员指标的驾驶员风险感知水平评价[J]. 中国安全科学学报, 2018, 28(12):144-149. doi: 10.16265/j.cnki.issn1003-3033.2018.12.023
	AI Qiannan. Driver risk perception level evaluation based on driver's index[J]. China Safety Science Journal, 2018, 28(12):144-149. doi: 10.16265/j.cnki.issn1003-3033.2018.12.023
[8]	徐丹, 代勇, 纪军红. 基于卷积神经网络的驾驶人行为识别方法研究[J]. 中国安全科学学报, 2019, 29(10):12-17. doi: 10.16265/j.cnki.issn1003-3033.2019.10.003
	XU Dan, DAI Yong, JI Junhong. Research on driver behavior recognition method based on convolutional neural network[J]. China Safety Science Journal, 2019, 29(10):12-17. doi: 10.16265/j.cnki.issn1003-3033.2019.10.003
[9]	张雅丽, 袁伟, 付锐, 等. 驾驶人应激反应能力强化培训方法[J]. 中国安全科学学报, 2018, 28(5):24-29. doi: 10.16265/j.cnki.issn1003-3033.2018.05.005
	ZHANG Yali, YUAN Wei, FU Rui, et al. Research on training method for enhancing driver's stress response ability[J]. China Safety Science Journal, 2018, 28(5):24-29. doi: 10.16265/j.cnki.issn1003-3033.2018.05.005
[10]	肖将. 驾驶人风险感知效用研究[D]. 昆明: 昆明理工大学, 2015.
	XIAO Jiang. Research on the utility of driver's risk perception[D]. Kunming: Kunming University of Science and Technology, 2015.
[11]	JIN Xiaohang, ZHAO Mingbo, CHOW T W S, et al. Motor bearing fault diagnosis using trace ratio linear discriminant analysis[J]. IEEE Transactions on Industrial Electronics, 2014, 61(5):2441-2451. doi: 10.1109/TIE.2013.2273471
[12]	YAN Yan, RICCI E, RAMANATHAN S, et al. Multitask linear discriminant analysis for view invariant action recognition[J]. IEEE Transactions on Image Processing, 2014, 23(12):5599-5611. pmid: 25361507

个人信息	驾驶背景	人数	占比/%	均值	标准差
性别	男	40	65.6	—	—
性别	女	21	34.4	—	—
年龄/ 岁	18~24	13	21.3	29.84	7.80
	25~40	40	65.6
	>40	8	13.1
驾龄/ a	<2	17	27.9	5.22	4.65
	2~10	37	60.7
	>10	7	11.4

情境编号	情境描述	冲突类型
1	主车准备直行通过无信号控制十字平面交叉口(以下简称平交口)时,有一行人准备横穿马路	机-人
2	主车准备右转通过平交口时,主车右后侧有一摩托车准备超越主车直行通过平交口,同时主车左前方有一辆卡车准备直行通过平交口	机-非机、机- 机
3	对向车道发生堵塞,主车正要直行通过平交口时,在主车左前方有一辆自行车突然从车流中窜出	机- 非机
4	主车准备右转通过平交口时,前方突然滚出一个皮球,然后在主车右前方有一小孩窜出捡球	机-人
5	主车准备左转通过平交口时,主车左前方有一辆卡车直行通过平交口,同时对向车道有一辆小汽车直行通过平交口	机-机
6	主车准备直行通过无信号平交口时,右前方有一辆车右转通过平交口	机-机

情境编号	1	2	3	4	5	6
客观情境风险度	0.10	0.38	1.30	0.90	0.30	0.74
风险等级	低	低	高	中	低	中

函数	特征值	方差占比/%	累计占比/%	典型相关性
F₁	1.988	63.7	63.7	0.816
F₂	1.135	36.3	100.0	0.729

函数检验	威尔克Lambda	卡方	自由度	显著性
F₁-F₂	0.157	70.419	62	0.217
F₂	0.468	28.826	30	0.527