To investigate the effectiveness of open roads in the ICV driving safety test, a rating algorithm for open road in ICV test according to the theory of random forest was proposed. A rating index system for open testing roads of intelligent connected vehicles was constructed from five aspects: road complexity, safety assurance, scene diversity, facility comprehensiveness, and scene typicality. Then a random forest model for the rating of open road is built and calibrated with the data of the existing open roads in ICV test. The rating algorithm is then applied in rating of open roads in Changchun City, the results of which demonstrates the feasibility and effectiveness of the model. The results indicate that the proposed intelligent connected vehicle(ICV) open test road rating algorithm can accurately perform road rating, providing model support for the selection, planning, and construction of open test roads for ICV.
目前,以美、日、欧等为代表的发达国家和地区正在大力发展智能网联汽车测试技术,并出台相关的测试道路评价指标以促进智能网联汽车技术的研发。例如,美国的SAE J3161标准提出3个测试道路等级,分别为:Level 1路侧标志数字化(如电子限速牌),Level 2支持SPAT(信号灯相位时序)广播,Level 3路侧参与车辆轨迹规划(如匝道合流辅助)。欧洲道路运输研究咨询委员会(European road transport research advisory council,ERTRAC)根据是否具有静态道路标识的数字化地图、事故、天气与微观交通情况等因素提出智能网联道路分级标准,将测试道路按照数字化与协同感知程度由高到低分为A、B、C、D、E 5个等级[4]。欧洲的ITS-G5标准提出基础V2I(如限速提醒)、协作式感知(如施工区域预警)和全自动化协同(如优先通行权分配)3个等级。
国内各省市近年大力推进智能网联汽车开放道路测试进程。北京、上海、重庆各地先后发布了智能网联汽车开放道路测实施细则或者管理办法,指导路测工作,并开放了部分道路进行测试[5,6]。针对某一拟建设为智能网联汽车开放测试道路的路段,需根据其道路环境、交通设施、测试场景等情况分析其能够满足哪一等级智能网联汽车的测试要求。北京市交通委员会从道路类型、交通设施以及交通密度等方面首次提出了系统的测试道路评价指标体系,明确了初步的交通管理设施要求[7]。并且针对北京高级别自动驾驶示范区的开放道路,提出3个设施能力建设等级,包括1.0阶段(覆盖5G+V2X),支持信号灯信息推送;2.0阶段(雷视一体机全域覆盖,实现弱势交通参与者预警);3.0阶段(引入“智慧城市专网”,支持车路云一体化控制)。上海奉贤测试区要求路侧单元(Road side unit,RSU)支持DSRC/C-V2X双模通信,且延迟在100 ms以内。我国也正在推动《智能网联汽车车路协同系统系列标准》的制定进程。
YangTao, MaYu-qin, LiuMeng, et al. Vehicle trajectory reconstruction model of signalized intersection in connected automated environments[J]. Journal of Southwest Jiaotong University, 2024, 59(5): 1148-1157.
JiangYang-sheng, LiuMeng, WangSi-chen, et al. Trajectory reconstruction for traffic flow mixed with connected automated vehicles based on car-following characteristics[J]. Journal of Southwest Jiaotong University, 2021, 56(6): 1135-1142.
LiXiao-chi, ZhaoXiang-mo, XuZhi-gang, et al. Modular flexible test bed for intelligent and connected transportation system[J]. China Journal of Highway and Transport, 2019, 32(6): 137-146.
[7]
Connected automated driving roadmap [R]. Brussels:European Road Transport Research Advisory Council, 2019.
HeFu-jian, ShiJuan, QinKong-jian. Research on the current situation and future development of domestic and foreign intelligent and connected vehicle public road testing[J]. China Auto, 2018(12): 28-33.
YuSheng-bo, ChenGui-hua, LiQiao, et al. Comparative research on intelligent and connected vehicle road tests at home and abroad[J]. Automotive Digest, 2020(2): 29-36.
ZongChen-hong, MaJian-xiao, LuTao. Reliability evaluation method of closed test fields for intelligent connected vehicles[J]. Journal of Chongqing University of Technology (Natural Science), 2023(4): 105-114.
LiXiao-qing, LiuQi, LiBing-lin, et al. Research on selection and classification of intelligent connected vehicle public testing roads[J]. Journal of Transportation Engineering, 2019(5): 74-78.
[17]
黄丽. 部分自动驾驶汽车场地测试与评价研究[D]. 重庆: 重庆大学车辆工程学院, 2018.
[18]
HuangLi. Study on the field test and evaluation methods of partial automated vehicles[D]. Chongqing: College of Vehicle Engineering, Chongqing University, 2018.
[19]
滕金润. 智能网联汽车开放测试道路交通安全评价研究[D]. 成都: 西华大学交通学院, 2022.
[20]
TengJin-run. Research on traffic safety evaluation of open test roads for intelligent connected vehicles[D]. Chengdu: College of Transportation, Xihua University, 2022.
[21]
AveryD J. System and method for testing the integrity of a vehicle testing/ diagnostic system[P].United States Patent: 20110185115, 2011-07-28.
[22]
DanielL. Research of the stopping distance for different road conditions[J]. Transport Problems, 2011, 6(4): 119-126.
MaYa-fang, WangWen-jie, GuKe, et al. Test and analysis of autonomous driving function of intelligent and connected vehicles[J]. Auto Time, 2021 (23): 38-40, 46.
[25]
BreimanL. Random forests[J]. Netherlands: Machine Learning, 2001, 45(1): 5-32.
[26]
StroblC, BoulesteixA L, KneibT, et al. Conditional variable importance for random forests[J]. England: BMC Bioinformatics, 2008, 9(1): l-11.
[27]
ReifD M, MotsingerA A, McKinneyB A, et al. Feature selection using a random forests classifier for the integrated analysis of multiple data types[C]∥Proceedings of the 2006 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology, Toronto, Canada, 2006: 171-178.
[28]
MohammedK, SounakC, MihailP. Predicting disease risks from highly imbalanced data using random forest[J]. England: BMC Medical Informatics and Decision Making, 2011, 11(7): 51-58.
[29]
VerikasA, GelzinisA, BacauskieneM. Mining data with random forests: a survey and results of new tests[J]. England: Pattern Recognition, 2011, 44(2): 330-349.
ZhangLi-jian. A new interpretation of people's pursuit of utility maximization[J]. Journal of Guangzhou University (Social Science Edition), 2018, 17(4): 50-56.