面向自动驾驶虚拟测试的变道切入场景库构建方法

doi:10.13229/j.cnki.jdxbgxb.20220029

摘要/Abstract

摘要：

基于智能汽车虚拟测试技术对驾驶场景的构建需求，采用阈值法和人工验证法获得了59例变道切入行为数据；分析了差异化场景类型元素和连续场景元素对风险感知系数的影响，降维后得到与变道切入场景危险程度显著相关的4类场景要素：主车纵向速度、相对纵向速度、纵向距离和前车切入持续时间；采用层次聚类优化K-means算法聚类得到了4类城市道路变道切入场景，借助PreScan软件构建了基于自然驾驶数据的自动驾驶虚拟测试场景库。

关键词: 汽车工程, 自动驾驶测试, 虚拟场景库, 切入场景, 聚类分析

Abstract:

Aiming at the requirements of intelligent vehicle virtual test technology for the construction of driving scenario， 59 cases of lane change behavior data were obtained by threshold method and manual verification method. The effects of differentiated scenario type elements and continuous scenario elements on risk perception coefficient were analyzed. After dimensionality reduction， 4 types of scenario elements significantly related to the risk degree of lane change scene were obtained： ego-vehicle longitudinal speed， relative longitudinal speed， longitudinal distance and front vehicle cut in duration. The k-means algorithm based on hierarchical clustering optimization was used to cluster and obtain 4 types of urban road cut-in scenarios. With the help of PreScan， the automatic driving virtual test scene database based on natural driving data was constructed.

Key words: automotive engineering, automatic driving test, virtual scenario library, cut-in scenario, cluster analysis

中图分类号:

U461.91

郭柏苍,雒国凤,金立生,谢宪毅,孙栋先. 面向自动驾驶虚拟测试的变道切入场景库构建方法[J]. 吉林大学学报(工学版), 2023, 53(11): 3130-3140.

Bai-cang GUO,Guo-feng LUO,Li-sheng JIN,Xian-yi XIE,Dong-xian SUN. Construction method of cut-in scenario library for automatic driving virtual tests[J]. Journal of Jilin University(Engineering and Technology Edition), 2023, 53(11): 3130-3140.

图/表 25

表1

图1

图2

图3

图4

表2

图5

表3

图6

表4

图7

表5

图8

图9

图10

图11

图12

图13

图14

图15

表6

表7

图16

表8

图17

参考文献 26

1	徐向阳, 胡文浩, 董红磊,等. 自动驾驶汽车测试场景构建关键技术综述[J]. 汽车工程, 2021, 43(4): 610-619.
	Xu Xiang-yang, Hu Wen-hao, Dong Hong-lei, et al. Review of key technologies for autonomous vehicle test scenario construction[J]. Automotive Engineering, 2021, 43(4): 610-619.
2	朱冰, 张培兴, 赵健,等. 基于场景的自动驾驶汽车虚拟测试研究进展[J]. 中国公路学报, 2019, 32(6): 1-19.
	Zhu Bing, Zhang Pei-xing, Zhao Jian, et al. Advances in scenario-based virtual testing of self-driving vehicles[J]. Chinese Journal of Highways, 2019, 32(6): 1-19.
3	朱冰,范天昕,赵健,等. 基于危险边界搜索的自动驾驶系统加速测试方法[J]. 吉林大学学报: 工学版, 2022, 52(1): 1-8.
	Zhu Bing, Fan Tian-xin, Zhao Jian, et al. Accelerated test method of automatic driving system based on dangerous boundary search[J]. Journal of Jilin University (Engineering Edition), 2022, 52(1): 1-8.
4	Zhao D, Lam H, Peng H, et al. Accelerated evaluation of automated vehicles safety in lane-change scenarios based on importance sampling techniques[J]. IEEE Transactions on Intelligent Transportation Systems, 2016, 18(3): 595-607.
5	Hou Y, Edara P, Sun C. Situation assessment and decision making for lane change assistance using ensemble learning methods[J]. Expert Systems with Applications, 2015, 42(8): 2015-3875.
6	中华人民共和国公安部交通管理局. 2015中华人民共和国道路交通事故统计年报[R]. 北京: 中华人民共和国公安部交通管理局, 2016.
7	Feng Z, Ma X, Zhu X, et al. Analysis of driver brake behavior under critical cut-in scenarios[C]∥ IEEE Intelligent Vehicles Symposium, New York, USA, 2018: 2054-2059.
8	王雪松, 杨敏明. 基于自然驾驶数据的变道切入行为分析[J]. 同济大学学报: 自然科学版, 2018, 46(8): 1057-1063.
	Wang Xue-song, Yang Min-ming. Analysis of lane change cut-in behavior based on natural driving data[J]. Journal of Tongji University (Natural Science Edition), 2018, 46(8): 1057-1063.
9	朱西产, 张佳瑞, 马志雄. 安全切入场景下的驾驶人初始制动时刻分析[J]. 中国公路学报, 2019, 32(6): 262-273.
	Zhu Xi-chan, Zhang Jia-rui, Ma Zhi-xiong. Analysis of driver's initial braking moment under safety cut-in scenario[J]. Chinese Journal of Highways, 2019, 32(6): 262-273.
10	Menzel T, Bagschik G, Maurer M. Scenarios for development, test and validation of automated vehicles[C]∥IEEE Intelligent Vehicles Symposium, New York, USA, 2018: 1-7.
11	Menzel T, Bagschik G, Isensee L, et al. From functional to logical scenarios: detailing a keyword-based scenario description for execution in a simulation environment[C]∥IEEE Intelligent Vehicles Symposium, Paris, France, 2019: 2383-2390.
12	胡林, 易平, 黄晶, 等. 基于真实事故案例的自动紧急制动系统两轮车测试场景研究[J]. 汽车工程, 2018, 40(12): 1436-1446.
	Hu Lin, Yi Ping, Huang Jing, et al. Research on two-wheeled vehicle test scenarios of automatic emergency braking system based on real accident cases[J]. Automotive Engineering, 2018, 40(12): 1436-1446.
13	郭景华, 李克强, 王进, 等. 基于危险场景聚类分析的前车随机运动状态预测研究[J]. 汽车工程, 2020, 42(7): 847-853, 859.
	Guo Jing-hua, Li Ke-qiang, Wang Jin, et al. Research on stochastic motion state prediction of front vehicle based on hazard scenario clustering analysis[J]. Automotive Engineering, 2020, 42(7): 847-853, 859.
14	Kondoh T, Yamamura T, Kitazaki S, et al. Identification of visual cues and quantification of drivers' perception of proximity risk to the lead vehicle in car-following situations[J]. Journal of Mechanical Systems for Transportation and Logistics, 2008, 1(2): 170-180.
15	党睿娜. 具有换道辅助功能的车辆自适应巡航控制[D].北京: 清华大学车辆与运载学院, 2013.
	Dang Rui-na. Adaptive cruise control for vehicles with lane change assist[D]. Beijing: School of Vehicles and Vehicles, Tsinghua University, 2013.
16	Wu J, Du Y, Qi G, et al. Leveraging longitudinal driving behaviour data with data mining techniques for driving style analysis[J]. Iet Intelligent Transport Systems, 2015, 9(8): 792-801.
17	金立生, 谢宪毅, 司法, 等. 考虑驾驶人特性的智能驾驶路径跟踪算法[J]. 汽车工程, 2021, 43(4): 553-561.
	Jin Li-sheng, Xie Xian-yi, Si Fa, et al. Intelligent driving path tracking algorithm considering driver characteristics[J]. Automotive Engineering, 2021, 43(4): 553-561.
18	Lv N, Wen J, Duan Z, et al. Vehicle trajectory prediction and cut-in collision warning model in a connected vehicle environment[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(2): 966-981.
19	Wang X, Yang M, Hurwitz D. Analysis of cut-in behavior based on naturalistic driving data[J]. Accident Analysis and Prevention, 2019, 124: 127-137.
20	王雪松, 徐晓妍. 基于自然驾驶数据的危险事件识别方法[J]. 同济大学学报: 自然科学版, 2020, 48(1): 51-59.
	Wang Xue-song, Xu Xiao-yan. Hazardous event identification method based on natural driving data[J]. Journal of Tongji University (Natural Science Edition), 2020, 48(1): 51-59.
21	Lv N, Duan Z, Ma C, et al. Safety margins-a novel approach from risk homeostasis theory for evaluating the impact of advanced driver assistance systems on driving behavior in near-crash events[J]. Journal of Intelligent Transportation Systems, 2020(1): 1-14.
22	丁洁云, 党睿娜, 王建强, 等. 驾驶人换道决策分析及意图识别算法设计[J]. 清华大学学报: 自然科学版, 2015, 55(7): 769-774.
	Ding Jie-yun, Dang Rui-na, Wang Jian-qiang, et al. Driver lane change decision analysis and intention recognition algorithm design[J]. Journal of Tsinghua University (Natural Science Edition), 2015, 55(7): 769-774.
23	李洪雪, 李世武, 孙文财, 等. 重型危险品半挂列车行驶工况的构建[J]. 吉林大学学报: 工学版, 2021, 51(5): 1700-1707.
	Li Hong-xue, Li Shi-wu, Sun Wen-cai, et al. Construction of working conditions of heavy dangerous goods semi-trailer train[J].Journal of Jilin University(Engineering and Technology Edition), 2021, 51(5): 1700-1707.
24	李航. 统计学习方法[M]. 2版. 北京: 清华大学出版社, 2019.
25	孙栋先. 基于自然驾驶数据的变道切入场景库构建研究[D]. 长春: 吉林大学交通学院, 2021.
	Sun Dong-xian. Research on construction of cut-in scenario library based on naturalistic driving data[D]. Changchun: College of Transportation, Jilin University, 2021.
26	彭涛, 方锐, 刘兴亮, 等. 基于典型事故场景的雪天高速换道自动驾驶策略[J]. 吉林大学学报: 工学版, 2022, 52(11): 2558-2567.
	Peng Tao, Fang Rui, Liu Xing-liang, et al. Automatic lane-changing driving strategy based on typical accident scenarios in snowy days[J]. Journal of Jilin University (Engineering and Technology Edition), 2022, 52(11): 2558-2567.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

数据来源	场景要素
前向毫米波雷达	相对纵向速度，相对横向距离，相对纵向距离
OBD接口	主车纵向速度，主车纵向加速度
车载摄像机	切入车辆类型，切入方向，切入持续时间，转向灯是否开启，主车运动是否受限

车辆类型	均值	标准差	标准误差	平均值的95%置信区间		最小值	最大值
车辆类型	均值	标准差	标准误差	下限	上限	最小值	最大值
Ⅰ	-1.38	1.91	0.31	-2.01	-0.76	-5.98	3.04
Ⅱ	-0.97	2.14	0.467	-1.94	0.00	-5.73	2.05

转向灯使用	均值	标准差	标准误差	平均值的95%置信区间		最小值	最大值
转向灯使用	均值	标准差	标准误差	下限	上限	最小值	最大值
未开始	-1.49	2.00	0.32	-2.13	-0.86	-5.98	2.05
开启	-0.71	1.90	0.44	-1.62	-0.21	-5.00	3.04

主车运动是否受限	均值	标准差	标准误差	平均值的95%置信区间		最小值	最大值
主车运动是否受限	均值	标准差	标准误差	下限	上限	最小值	最大值
未受限	-0.73	1.93	0.58	-2.03	-0.56	-5.73	1.39
左侧受限	-1.66	2.18	0.40	-2.47	-0.84	-5.98	3.04
右侧受限	-0.86	1.59	0.37	-1.65	-0.07	-5.00	1.91

方法	相关系数	方法	相关系数
类平均法	0.8858	中间距离法	0.7888
重心法	0.8805	最短距离法	0.8159
最长距离法	0.7065	离差平方和	0.6013