驾驶人认知分心识别随机森林模型研究

doi:10.16265/j.cnki.issn1003-3033.2018.01.004

中国安全科学学报 ›› 2018, Vol. 28 ›› Issue (1): 20-25.doi: 10.16265/j.cnki.issn1003-3033.2018.01.004

驾驶人认知分心识别随机森林模型研究

周扬^1,2 讲师, 付锐¹ 教授, 袁伟¹ 教授, 王栋^2,3 讲师, 张瑞宾¹ 讲师

1 长安大学汽车学院,陕西西安 710064
2 西安航空学院车辆工程学院,陕西西安 710077
3 长安大学汽车运输安全保障技术交通行业重点实验室,陕西西安 710064

收稿日期:2017-10-23 出版日期:2018-01-28 发布日期:2020-09-28
作者简介:周扬 (1989—),男,陕西汉中人,博士研究生,讲师,研究方向为驾驶人行为、车辆安全技术。E-mail:297399014@qq.com。
基金资助:
国家自然科学基金资助(61473046,51775053);中央高校基本科研业务费资金资助(310822171116)。

Research on drivers' cognitive distracted recognition model based on random forest

ZHOU Yang^1,2, FU Rui¹, YUAN Wei¹, WANG Dong^2,3, ZHANG Ruibin¹

1 School of Automobile, Chang'an University, Xi'an Shaanxi 710064, China
2 School of Vehicle Engineering, Xi'an Aeronautical University, Xi'an Shaanxi 710077, China
3 Key Laboratory of Automotive Transportation Safety Technology, Ministry of Transport, Chang'an University, Xi'an Shaanxi 710064, China

Received:2017-10-23 Online:2018-01-28 Published:2020-09-28

摘要/Abstract

摘要： 为识别驾驶人认知分心状态,招募13名驾驶人参与驾驶模拟器试验。通过眼动仪采集被试正常驾驶及认知分心状态下的眼动数据,提取5 s 时间窗口内的眼动特征。运用随机森林方法构建认知分心识别模型,应用网格搜索确定最优模型参数,并采用100次留出法评估模型性能。根据随机森林模型特征重要性度量结果,进一步分析认知负荷对驾驶人注视及眨眼持续时间的影响。结果表明:当决策树数量为125、最大特征数为5时,模型识别平均准确率为83.69%;注视持续时间及噪声持续时间是认知分心识别的2个关键特征,随着认知负荷的提高,驾驶人注视持续时间减少、眨眼时间增加。

关键词: 交通安全, 认知分心, 眼动特征, 随机森林, 模拟器试验

Abstract: In order to recognize the cognitive distracted state of drivers, a driving simulator test was carried out with 13 drivers recruited as testees. Data on eye movement of the testees were collected by an eye tracker when they performed a normal driving task or a cognitive secondary tasks. Eye movement features were extracted through 5 s time window. A cognitive distracted recognition model was built using random forest. The optimal model parameters were determined by grid searching, and the model performance were evaluated by using the 100 times hold out tests method. The effects of cognitive load on drivers' fixation duration and blink duration were analyzed according to feature importance measurements generated by the random forest model. The results show that when the number of decision trees is 125 and max features are 5, the mean accuracy can achieve 83.69%, that fixation duration and noise data duration are the two features having a key role in recognizing cognitive distraction, and that an increase in cognitive load will result in longer fixation duration and shorter blink duration.

Key words: transportation safety, cognitive distraction, eye movement feature, random forest, driving simulator test

中图分类号:

周扬, 付锐, 袁伟, 王栋, 张瑞宾. 驾驶人认知分心识别随机森林模型研究[J]. 中国安全科学学报, 2018, 28(1): 20-25.

ZHOU Yang, FU Rui, YUAN Wei, WANG Dong, ZHANG Ruibin. Research on drivers' cognitive distracted recognition model based on random forest[J]. China Safety Science Journal, 2018, 28(1): 20-25.

参考文献

[1] NHTSA. Distracted driving 2015[EB/OL]. [2017-04-10]. https://www.nhtsa.gov/risky-driving/distracted-driving#topic-consequences.
[2] HUISINGH C, GRIFFIN R, MCGWIN J G. The prevalence of distraction among passenger vehicle drivers: a roadside observational approach [J]. Traffic Injury Prevention, 2015, 16(2): 140-146.
[3] NHTSA. What is distracted driving[EB/OL]. [2014-05-23]. http://www.distraction.gov/content/get-the-facts/facts-and-statistics.html.
[4] HAIGNEY D, WESTERMAN S J. Mobile (cellular) phone use and driving: a critical review of research methodology[J]. Ergonomics, 2001, 44(2):132.
[5] 马艳丽, 顾高峰, 高月娥, 等. 基于驾驶绩效的车载信息系统操作分心判定模型[J]. 中国公路学报, 2016, 29(4): 124-129.
MA Yanli, GU Gaofeng, GAO Yuee, et al. Driver distraction judging model under in-vehicle information system operation based on driving performance [J]. China Journal of Highway and Transport, 2016, 29(4): 124-129.
[6] AHLSTORM U, FRIEDMAN-BERG F J. Using eye movement activity as a correlate of cognitive workload[J]. International Journal of Industrial Ergonomics, 2006, 36(7): 623-636.
[7] HARBLUK J L, NOY Y I, EIZENMAN M. The impact of cognitive distraction on driver visual behavior and vehicle control[R], 2002.
[8] LIANG Yulan, LEE J D. Combining cognitive and visual distraction: less than the sum of its parts[J]. Accident Analysis & Prevention, 2010, 42(3): 881-890.
[9] LIANG Yulan, REYES M L, LEE J D. Real time detection of driver cognitive distraction using support vector machines[J].IEEE Transactions on Intelligent Transportation Systems, 2007, 8(2): 340 -350.
[10] LIU Tianchi, YANG Yan, HUANG Guangbin. Driver distraction detection using semi supervised machine learning [J]. IEEE Transactions on Intelligent Transportation Systems, 2016, 17(4): 1 108-1 120.
[11] 方匡南, 吴见彬, 朱建平, 等. 随机森林方法研究综述[J]. 统计与信息论坛, 2011, 26(3): 32-36.
FANG Kuangnan, WU Jianbin, ZHU Jianping, el al. A review of technologies on random forest[J]. Statistics &Information Forum, 2011, 26(3): 32-36.
[12] 马勇, 付锐, 袁钰, 等. 基于驾驶人分心影响程度的驾驶分心任务评级[J]. 交通信息与安全, 2014, 32(5): 47-57.
MA Yong, FU Rui, YUAN Yu, el al. Rating driving distraction tasks based on distraction impact on drivers[J]. Journal of Transport Information and Safety, 2014, 32(5): 47-57.
[13] 张存保, 彭汉辉, 张珊,等. 雾天高速公路实时交通安全状态评价方法[J]. 中国安全科学学报, 2017, 27(4): 110-114.
ZHANG Cunbao, PENG Hanhui, ZHANG Shan, el al. Real-time traffic safety evaluation method for freeway in fog[J]. China Safety Science Journal, 2017, 27(4): 110-114.
[14] LEE H I, PARK S, LIM J, et al. Influence of driver's career and secondary cognitive task on visual search behavior in driving: a dual task paradigm[J]. Advances in Physical Education, 2015, 5(4): 245-254.

[1]	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.
[2]	卢颖, 赵志攀, 姜学鹏, 吴锦东, 范小鹏. 大数据视域下体育场馆动态火灾风险指标研究[J]. 中国安全科学学报, 2022, 32(4): 155-162.
[3]	李超, 邓小宝, 史运涛, 孙德辉, 焦彦宗. 社区户内燃气泄漏动态预警模型[J]. 中国安全科学学报, 2022, 32(3): 90-97.
[4]	丁茜, 赵晓东, 吴鑫俊, 张泰丽, 徐振涛. 基于RBF核的多分类SVM滑塌易发性评价模型[J]. 中国安全科学学报, 2022, 32(3): 194-200.
[5]	陈圣迪, 葛晓婉, 赵晓晨, 陆键, 邢莹莹. 山区二级公路运营期常态交通风险评价^*[J]. 中国安全科学学报, 2022, 32(2): 176-183.
[6]	丘积, 王红, 蒋若曦, 肖文彬, 常红光, 朱顺应. 高速公路转换区严重冲突识别与分析^*[J]. 中国安全科学学报, 2022, 32(2): 184-191.
[7]	张新生, 蔡宝泉. 基于改进随机森林模型的海底管道腐蚀预测^*[J]. 中国安全科学学报, 2021, 31(8): 69-74.
[8]	李杰, 曾叙砜, 孙领, 刘伟, 李平. 全国道路交通安全水平的时空布局演化研究^*[J]. 中国安全科学学报, 2021, 31(12): 136-143.
[9]	吴仁彪, 何宇翔, 贾云飞. 基于改进层次分析法的重点人员风险评价方法^*[J]. 中国安全科学学报, 2021, 31(10): 112-118.
[10]	马丹, 裴玉龙, 田昆. 基于文献计量分析的行人交通安全研究综述[J]. 中国安全科学学报, 2020, 30(8): 101-108.
[11]	田睿, 孟海东, 陈世江, 王创业, 石磊. RF-AHP-云模型下岩爆烈度分级预测模型[J]. 中国安全科学学报, 2020, 30(7): 166-172.
[12]	刘文歌, 杨晶. 基于混合Logit模型的轨道交通安全统计生命价值[J]. 中国安全科学学报, 2020, 30(4): 186-191.
[13]	江欣国, 任瀚堃, 范英飞, 李想, 付川云. 信号协调控制干线交通安全仿真分析[J]. 中国安全科学学报, 2020, 30(3): 143-149.
[14]	胡瑞, 徐传玲, 冯永泰, 文超, 王全泉. 广深高速铁路列车分类型晚点预测^*[J]. 中国安全科学学报, 2019, 29(S2): 181-186.
[15]	陆欢, 戢晓峰, 杨文臣, 胡澄宇. 高原山区公路环境下交通事故形态致因分析[J]. 中国安全科学学报, 2019, 29(5): 44-49.

驾驶人认知分心识别随机森林模型研究

Research on drivers' cognitive distracted recognition model based on random forest

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价