Lane-change risk warning in interweaving area considering information from intelligent connected near-neighboring vehicles
-
摘要: 面向车辆换道风险预测时特征差异大、样本不均衡、参数调优时间久的问题，将高精度微观车辆轨迹数据与超参数优化机器学习方法相结合，提出了一种可应用于智能网联车辆(ICV)的交织区换道风险识别与预警方法；基于无人机航拍视频，从广域视角提取了城市快速路交织区时间精度为0.1 s、空间精度为每像素0.1 m的换道轨迹，测算了车辆间距、矢量速度、加速度、接近率、速度角度等换道风险感知信息；引入考虑近邻车辆信息的换道TTC模型，以反映车辆汇入或汇出主线的迫切需求，描述其在不同位置的换道行为差异性；结合15分位数法和四分位差法，划分了换道风险预警等级；基于准确率、真阳性率、灵敏度等多项评价指标，遴选并对比了线性分类器、支持向量机、K近邻以及RUSBoost模型换道风险预测结果，得出交织区换道风险实时预警优选模型，针对优选模型进行了超参数优化与验证。研究结果表明：RUSBoost模型为优选模型；超参数优化机器学习方法迭代至第24次时，RUSBoost具有最小误差与最佳点超参数；RUSBoost、BRUSBoost优化模型预测准确率分别为91.40%、99.80%，AUC分别为0.96、0.99；BRUSBoost优化模型对于Ⅰ级、Ⅲ级换道预警精准率分别提升了50.9%、41.2%，有效改善了极端风险换道条件更复杂也更不易预测的缺陷。研究成果有助于智能网联车辆换道决策与轨迹优化，指导交管部门制定ICV动态预警方案。Abstract: In view of the problems of large feature differences, unbalanced samples, and long parameter tuning time in predicting the lane-change risk of vehicles, the high-precision microscopic vehicle trajectory data were combined with the hyperparametric optimization machine learning method, and a method for identifying and warning of the lane-change risk in interweaving areas was proposed, which could be applied to intelligent connected vehicles (ICV). According to the unmanned aerial vehicle video, lane-change trajectories with time accuracy of 0.1 s and spatial accuracy of 0.1 m per pixel in the urban expressway interweaving area were extracted from a wide-area view, and the lane-change risk perception information, such as the vehicle spacing, vector velocity, acceleration, proximity rate, and velocity angle, was measured. A lane-change time-to-collision (TTC) model that took into account information from near-neighboring vehicles was introduced, and the urgent need for vehicles to merge into or leave arteries was reflected. In addition, the difference in the lane-change behaviors at different locations was described. The 15th quartile method and the interquartile range method were used, and the warning levels for lane-change risks were classified. With several evaluation metrics, such as the accuracy, true positive rate, and sensitivity, the linear classifier, support vector machine, K-nearest neighbor, and RUSBoost model were selected and compared in terms of the lane-change risk prediction results, and the optimal model for the real-time warning of lane-change risks in interweaving areas was selected. For the optimal model, the hyperparameters were optimized and validated. Research results show that the RUSBoost model is the optimal model. When the hyperparametric optimization machine learning method iterates 24th times, the minimum error and optimal point hyperparameter of RUSBoost appear, and the prediction accuracies of RUSBoost and optimized BRUSBoost models are 91.40% and 99.80%, respectively. The AUC values are 0.96 and 0.99, respectively, and the warning accuracy of the optimized BRUSBoost model for level Ⅰ and Ⅲ lane change risks improves by 50.9% and 41.2%, respectively. As a result, the defects of more complex and less predictable lane-change conditions of extreme risks improves effectively. The research results can help the ICV lane-change decisions and trajectory optimization and guide traffic management departments to develop ICV dynamic warning schemes.
-
图 4 左右换道方向TTC累计频率
Figure 4. TTC cumulative frequencies in left and right lane changing directions
图 11 左右换道方向混淆矩阵
Figure 11. Confusion matrixes of left and right lane changing direction
表 1 最优参数配置
Table 1. Setting of optimal parameters
参数 参数调整范围 调参结果 树的数量 10~500 23 学习率 0.001~1.000 0.007 最大特征数 1~21 20 最大分裂数 1~222 064 96 
下载: 导出CSV
-
[1] 李熙莹, 梁靖茹, 张伟斌, 等. 基于航拍视频构建风险指数的交织区拥堵识别方法[J]. 铁道科学与工程学报, 2023, 20(2): 494-505. https://www.cnki.com.cn/Article/CJFDTOTAL-CSTD202302013.htmLI Xi-ying, LIANG Jing-ru, ZHANG Wei-bin, et al. Congestion recognition method in weaving sections by constructing risk index based on aerial video[J]. Journal of Railway Science and Engineering, 2023, 20(2): 494-505. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-CSTD202302013.htm [2] VAN BEINUM A, FARAH H, WEGMAN F, et al. Driving behaviour at motorway ramps and weaving segments based on empirical trajectory data[J]. Transportation Research Part C: Emerging Technologies, 2018, 92: 426-441. doi: 10.1016/j.trc.2018.05.018 [3] HAO Wei, ZHANG Zhao-lei, GAO Zhi-bo, et al. Research on mandatory lane-changing behavior in highway weaving sections[J]. Journal of Advanced Transportation, 2020, 2020: 1-9. [4] 张卫华, 刘嘉茗, 解立鹏, 等. 网联混合环境快速路交织区自动驾驶车辆换道模型[J]. 吉林大学学报(工学版), 2022, DOI: 10.13229/j.cnki.jdxbgxb20220990.ZHANG Wei-hua, LIU Jia-ming, XIE Li-peng, et al. Lane-changing model of autonomous vehicle in weaving area of expressway in intelligent and connected mixed environment[J]. Journal of Jilin University (Engineering and Technology Edition), 2022, DOI: 10.13229/j.cnki.jdxbgxb20220990.ZHANG(inChinese) [5] 陈吉清, 翁楚滨, 兰凤崇. 智能车辆换道潜在冲突分析与风险量化方法[J]. 汽车工程, 2021, 43(11): 1565-1576, 1586. https://www.cnki.com.cn/Article/CJFDTOTAL-QCGC202111001.htmCHEN Ji-qing, WENG Chu-bin, LAN Feng-chong. Potential conflict analysis and risk quantification method of intelligent vehicle lane change[J]. Automotive Engineering, 2021, 43(11): 1565-1576, 1586. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-QCGC202111001.htm [6] BHA M, GHULAM H. A simple lane change model for microscopic traffic flow simulation in weaving sections[J]. Transportation Letters, 2011, 3(4): 231-251. doi: 10.3328/TL.2011.03.04.231-251 [7] PAN T L, LAM W H K, SUMALEE A, et al. Modeling the impacts of mandatory and discretionary lane-changing maneuvers[J]. Transportation Research Part C: Emerging Technologies, 2016, 68: 403-424. doi: 10.1016/j.trc.2016.05.002 [8] YUAN Jing-hui, ABDEL A M, CAI Qing, et al. Investigating drivers' mandatory lane change behavior on the weaving section of freeway with managed lanes: a driving simulator study[J]. Transportation Research Part F: Traffic Psychology and Behaviour, 2019, 62: 11-32. doi: 10.1016/j.trf.2018.12.007 [9] WAN Xia, JIN P J, GU Hai-yan, et al. Modeling freeway merging in a weaving section as a sequential decision-making process[J]. Journal of Transportation Engineering, Part A: Systems, 2017, 143(5): 05017002. doi: 10.1061/JTEPBS.0000048 [10] 蒋渊德, 朱冰, 赵祥模, 等. 面向自动驾驶汽车测试的交通车辆交互过程建模[J]. 汽车工程, 2022, 44(12): 1825-1833. doi: 10.19562/j.chinasae.qcgc.2022.12.004JIANG Yuan-de, ZHU Bing, ZHAO Xiang-mo, et al. Modeling of traffic vehicle interaction for autonomous vehicle testing[J]. Automotive Engineering, 2022, 44(12): 1825-1833. (in Chinese) doi: 10.19562/j.chinasae.qcgc.2022.12.004 [11] 张荣辉, 游峰, 初鑫男, 等. 车-车协同下无人驾驶车辆的换道汇入控制方法[J]. 中国公路学报, 2018, 31(4): 180-191. doi: 10.19721/j.cnki.1001-7372.2018.04.022ZHANG Rong-hui, YOU Feng, CHU Xin-nan, et al. Lane change merging control method for unmanned vehicle under V2V cooperative environment[J]. China Journal of Highway and Transport, 2018, 31(4): 180-191. (in Chinese) doi: 10.19721/j.cnki.1001-7372.2018.04.022 [12] 陈荔, 张驰, 王博, 等. 隧道与互通出口小净距路段换道可靠性研究[J]. 中国公路学报, 2022, 35(9): 51-65. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGGL202209005.htmCHEN Li, ZHANG Chi, WANG Bo, et al. Research of lane changing reliability on short distance section between tunnel and interchange exit[J]. China Journal of Highway and Transport, 2022, 35(9): 51-65. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZGGL202209005.htm [13] 单肖年, 万长薪, 李志斌, 等. 智能网联环境下多车道异质交通流建模与仿真[J]. 交通运输系统工程与信息, 2022, 22(6): 74-84. https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202206008.htmSHAN Xiao-nian, WAN Chang-xin, LI Zhi-bin, et al. Modeling and simulation of multi-lane heterogeneous traffic flow in intelligent and connected vehicle environment[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(6): 74-84. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202206008.htm [14] ZHOU Hui-ping, ITOH M. How does a driver perceive risk when making decision of lane-changing?[J]. IFAC—PapersOnLine, 2016, 49(19): 60-65. doi: 10.1016/j.ifacol.2016.10.462 [15] 吕能超, 杜子君, 吴超仲, 等. 多车道高速公路出口开口段安全特性分析[J]. 交通运输系统工程与信息, 2021, 21(3): 120-130. https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202103015.htmLYU Neng-chao, DU Zi-jun, WU Chao-zhong, et al. Safety characteristics of exit section of multi-lane expressway[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(3): 120-130. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202103015.htm [16] 漆巍巍, 马思维, 周南杰. 高速公路分合流区车流元胞模型的构建及应用[J]. 华南理工大学学报(自然科学版), 2022, 50(10): 11-18.QI Wei-wei, MA Si-wei, ZHOU Nan-jie. Construction and application of cellular automaton model of traffic flow in freeway diverging and merging areas[J]. Journal of South China University of Technology (Natural Science Edition), 2022, 50(10): 11-18. (in Chinese) [17] THEOFILATOS A, YANNIS G, KOPELIAS P, et al. Impact of real-time traffic characteristics on crash occurrence: preliminary results of the case of rare events[J]. Accident Analysis and Prevention, 2019, 130: 151-159. doi: 10.1016/j.aap.2017.12.018 [18] 刘唐志, 毕辉云, 杨卓思, 等. 基于操纵量指标的合流区危险驾驶行为谱研究[J/OL]. 交通运输系统工程与信息, (2023-01-18)[2023-04-01]. https://kns.cnki.net/kcms/detail//11.4520.U.20230118.1025.004.html. LIU Tang-zhi, BI Hui-yun, YANG Zhuo-si, et al. Research on dangerous driving behavior spectrum in merging area based on maneuver indicators[J/OL]. Journal of Transportation Systems Engineering and Information Technology, (2023-01-18)[2023-04-01].https://kns.cnki.net/kcms/detail//11.4520.U.20230118.1025.004.html. (in Chinese)[19] FORMOSA N, QUDDUS M, ISON S, et al. Predicting real-time traffic conflicts using deep learning[J]. Accident Analysis and Prevention, 2020, 136: 105429. doi: 10.1016/j.aap.2019.105429 [20] FERREIRA S, COUTO A. A probabilistic approach towards a crash risk assessment of urban segments[J]. Transportation Research Part C: Emerging Technologies, 2015, 50: 97-105. [21] ZHENG Lai, SAYED T, MANNERING F. Modeling traffic conflicts for use in road safety analysis: a review of analytic methods and future directions[J]. Analytic Methods in Accident Research, 2021, 29: 100142. [22] WARD J R, AGAMENNONI G, WORRALL S, et al. Extending time to collision for probabilistic reasoning in general traffic scenarios[J]. Transportation Research Part C: Emerging Technologies, 2015, 51: 66-82. [23] 刘兵, 王锦锐, 赵荣达, 等. 面向换道冲突互信息的复杂交织区车辆风险识别[J]. 安全与环境学报, 2022, 22(4): 1768-1775. https://www.cnki.com.cn/Article/CJFDTOTAL-AQHJ202204010.htmLIU Bing, WANG Jin-rui, ZHAO Rong-da, et al. Vehicle risk identification in complex weaving areas for conflicting mutual information on lane change[J]. Journal of Safety and Environment, 2022, 22(4): 1768-1775. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-AQHJ202204010.htm [24] YUAN Chen, LI Ye, HUANG He-lai, et al. Using traffic flow characteristics to predict real-time conflict risk: a novel method for trajectory data analysis[J]. Analytic Methods in Accident Research, 2022, 35: 100217. [25] 温惠英, 李秋灵, 赵胜. 快速路合流区大型车换道时空特征及风险研究[J]. 华南理工大学学报(自然科学版), 2022, 50(5): 11-21. https://www.cnki.com.cn/Article/CJFDTOTAL-HNLG202205002.htmWEN Hui-ying, LI Qiu-ling, ZHAO Sheng. Research on spatiotemporal characteristic and risk of lane-changing behaviors of large vehicles in expressway merging area[J]. Journal of South China University of Technology (Natural Science Edition), 2022, 50(5): 11-21. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-HNLG202205002.htm [26] 刘兵, 王锦锐, 谢济铭, 等. 微观轨迹数据驱动的交织区换道概率分布模型[J]. 汽车安全与节能学报, 2022, 13(2): 333-340. https://www.cnki.com.cn/Article/CJFDTOTAL-QCAN202202014.htmLIU Bing, WANG Jin-rui, XIE Ji-ming, et al. Microscopic trajectory data-driven probability distribution model for weaving area of channel change[J]. Journal of Automotive Safety and Energy, 2022, 13(2): 333-340. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-QCAN202202014.htm [27] 谢济铭. 考虑车辆行驶行为的山地城市干线复杂交织区交通流建模与分析[D]. 重庆: 重庆交通大学, 2020.XIE Ji-ming. Traffic flow modeling and analysis on complex weaving area of mountain city trunk road considering vehicle driving behaviors[D]. Chongqing: Chongqing Jiaotong University, 2020. (in Chinese) [28] XING Lu, HE Jie, ABDEL-ATY M, et al. Examining traffic conflicts of up stream toll plaza area using vehicles trajectory data[J]. Accident Analysis and Prevention, 2019, 125: 174-187. [29] 谢济铭, 秦雅琴, 彭博, 等. 多车道交织区车辆跟驰行为风险判别与冲突预测[J]. 交通运输系统工程与信息, 2021, 21(3): 131-139. https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202103016.htmXIE Ji-ming, QIN Ya-qin, PENG Bo, et al. Risk discrimination and conflict prediction of vehicle-following behavior in multi-lane weaving sections[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(3): 131-139. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YSXT202103016.htm [30] 吴晓建, 燕冬, 王爱春, 等. 融合前车轨迹预测的改进人工势场轨迹规划研究[J]. 汽车工程, 2021, 43(12): 1752-1761,1779. https://www.cnki.com.cn/Article/CJFDTOTAL-QCGC202112003.htmWU Xiao-jian, YAN Dong, WANG Ai-chun, et al. Research on improved artificial potential field path planning integrating prediction of preceding vehicle trajectory[J]. Automotive Engineering, 2021, 43(12): 1752-1761, 1779. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-QCGC202112003.htm [31] 王可, 陆键, 蒋愚明. 基于车辆行驶轨迹的道路不良驾驶行为谱构建与特征值计算方法[J]. 交通运输工程学报, 2020, 20(6): 236-249. doi: 10.19818/j.cnki.1671-1637.2020.06.021WANG Ke, LU Jian, JIANG Yu-ming. Abnormal road driving behavior spectrum establishment and characteristic value calculation method based on vehicle driving trajectory[J]. Journal of Traffic and Transportation Engineering, 2020, 20(6): 236-249. (in Chinese) doi: 10.19818/j.cnki.1671-1637.2020.06.021 [32] 曾德宇, 梁泽逍, 吴宗泽. 基于加权核范数和L2, 1范数的最优均值线性分类器[J]. 电子与信息学报, 2022, 44(5): 1602-1609. https://www.cnki.com.cn/Article/CJFDTOTAL-DZYX202205011.htmZENG De-yu, LIANG Ze-xiao, WU Zong-ze. Optimal mean linear classifier via weighted nuclear norm and L2, 1 norm[J]. Journal of Electronics and Information Technology, 2022, 44(5): 1602-1609. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-DZYX202205011.htm [33] 杨飞, 郭煜东, JIN J P, 等. 基于手机传感器数据的交通出行调查实证评估[J]. 交通运输工程学报, 2020, 20(1): 226-238. doi: 10.19818/j.cnki.1671-1637.2020.01.019YANG Fei, GUO Yu-dong, JIN J P, et al. Empirical evaluation of travel survey based on mobile phone sensor data[J]. Journal of Traffic and Transportation Engineering, 2020, 20(1): 226-238. (in Chinese) doi: 10.19818/j.cnki.1671-1637.2020.01.019 [34] XUE Qing-wen, XING Ying-ying, LU Jian. An integrated lane change prediction model incorporating traffic context based on trajectory data[J]. Transportation Research Part C: Emerging Technologies, 2022, 141: 103738. [35] YI De-wei, SU Jin-ya, LIU Cun-jia, et al. A machine learning based personalized system for driving state recognition[J]. Transportation Research Part C: Emerging Technologies, 2019, 105: 241-261. [36] 李秋谷. 隧道出口近邻区车辆换道意图辨识与轨迹预测[D]. 昆明: 昆明理工大学, 2022.LI Qiu-gu. Vehicle lane change intention recognition and trajectory prediction in tunnel exit near-neighborhood[D]. Kunming: Kunming University of Science and Technology, 2022. (in Chinese) [37] SINGH P, CHAUDHURY S, PANIGRAHI B K. Hybrid MPSO-CNN: multi-level particle swarm optimized hyperparameters of convolutional neural network[J]. Swarm and Evolutionary Computation, 2021, 63: 100863. [38] 谢济铭, 夏玉兰, 秦雅琴, 等. 基于双向长短期记忆网络的城市快速路合流区车速预测[J/OL]. 西南交通大学学报, (2022-07-06)[2022-10-19]. https://kns.cnki.net/kcms/detail/51.1277.U.20220705.2038.020.html. XIE Ji-ming, XIA Yu-lan, QIN Ya-qin, et al. Traffic speed prediction in merging zone of urban expressway based on bidirectional long short-term memory network[J/OL]. Journal of Southwest Jiaotong University, (2022-07-06)[2022-10-19].https://kns.cnki.net/kcms/detail/51.1277.U.20220705.2038.020.html. (in Chinese)[39] ZHONG Jing-jing, TSE P W, WANG Dong. Novel Bayesian inference on optimal parameters of support vector machines and its application to industrial survey data classification[J]. Neurocomputing, 2016, 211: 159-171. [40] 魏佳恒, 郭惠勇. 基于贝叶斯优化BiLSTM模型的输电塔损伤识别[J]. 振动与冲击, 2023, 42(1): 238-248. https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ202301028.htmWEI Jia-heng, GUO Hui-yong. Damage identification of transmission tower based on BO-BiLSTM model[J]. Journal of Vibration and Shock, 2023, 42(1): 238-248. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ202301028.htm [41] 于靖宇, 魏海平, 郭宏伟, 等. 贝叶斯优化模糊C均值的城市交通状态判别方法[J]. 测绘科学技术学报, 2021, 38(6): 653-658. https://www.cnki.com.cn/Article/CJFDTOTAL-JFJC202106016.htmYU Jing-yu, WEI Hai-ping, GUO Hong-wei, et al. Urban traffic state identification based on BO-FCM[J]. Journal of Geomatics Science and Technology, 2021, 38(6): 653-658. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JFJC202106016.htm [42] 秦雅琴, 李秋谷, 赵鹏燕, 等. 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. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK202204021.htmQIN Ya-qin, LI Qiu-gu, ZHAO Peng-yan, et al. Research on risk perception tendency of drivers based on multi-class Adaboost algorithm[J]. China Safety Science Journal, 2022, 32(4): 141-147. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK202204021.htm -
点击查看大图
图(12) / 表(1)
计量
- 文章访问数: 479
- HTML全文浏览量: 146
- PDF下载量: 117
- 被引次数: 0
