Real-time identification method of abnormal road driving behavior based on vehicle driving trajectory
-
摘要: 为了提高道路交通安全主动防控能力, 以小汽车行驶轨迹数据为研究对象, 研究了不良驾驶行为的实时辨识问题; 基于无人机拍摄交通流视频提取海量车辆行驶轨迹数据; 提出了应用风险度量方法量化典型不良驾驶行为的理论; 使用大样本统计分布方法确定不良驾驶行为的特征参数阈值; 建立了结合交通环境信息的不良驾驶行为谱, 计算了不良驾驶行为谱特征值; 以车辆不良驾驶行为谱特征值为依据标定不良车辆样本; 以部分驾驶行为谱参数为输入, 使用不平衡类提升的人工智能算法建立了不良驾驶行为辨识模型; 为了验证方法的有效性, 使用无人机交通视频采集了上海市的车辆行驶轨迹数据, 分析了小汽车不良跟驰行为特征。分析结果表明: 使用四分位差法得到不良跟驰特征参数的阈值为0.19 s-1, 大部分样本处于正常跟驰状态, 约2%样本处于不良跟驰状态; 基于每辆车行驶轨迹中正常跟驰状态和不良跟驰状态的比例, 使用95%分位数将8 917 veh小汽车样本划分为不良跟驰车辆445 veh与正常跟驰车辆8 472 veh; 不平衡类提升算法CUSBoost辨识不良跟驰车辆达到了94.4%的召回率和85.9%的精确率, 平衡分数和精确率-召回率曲线下的面积为所有算法中最高。可见, 不良驾驶行为谱作为一种客观的不良驾驶行为量化表达方法, 与人工智能方法结合可以生成海量的不良驾驶行为谱库; 不平衡类提升算法可以解决不良驾驶行为数据的不平衡问题, 与常规算法相比具有更好的不良驾驶行为辨识能力。Abstract: To improve the active prevention and control of road traffic safety, the driving trajectory data of cars were used as the research object and the real-time identification problem of abnormal driving behaviors was studied. A massive amount of vehicle driving trajectory data was extracted based on an unmanned aerial vehicle recorded traffic video. The theory of applying the measurement of risk to quantify the typical abnormal driving behaviors was proposed. A large-sample statistical distribution method was used to determine the characteristic value thresholds for abnormal driving behaviors. The abnormal driving behavior spectrum was established and combined with the traffic environment information. The characteristic value of the abnormal driving behavior spectrum was calculated. Based on the characteristic value of the vehicle's abnormal driving behavior spectrum, the abnormal vehicle samples were labeled. The driving behavior spectrum parameters were partially used as input and imbalanced class boosting artificial intelligence algorithms were used to establish an abnormal driving behavior identification model. To verify the effectiveness of the method, vehicle driving trajectory data were collected using an unmanned aerial vehicle traffic video in Shanghai, and the characteristics of car abnormal car-following behavior were analyzed. Research result shows that the threshold of the abnormal car-following characteristic parameter is 0.19 s-1 using the interquartile range method, most of the samples are in the normal car-following state, and about 2% of the samples are in the abnormal car-following state. Based on the ratio of the normal car-following state and the abnormal car-following state in the driving trajectory of each car, a 95% percentile is used to divide the examples of 8 917 cars into 445 abnormal car-following cars and 8 472 normal car-following cars. The imbalanced class boosting algorithm CUSBoost achieves a recall rate of 94.4% and a precision rate of 85.9% in identifying abnormal car-following vehicles. Its balanced score and the area under the precision-recall curve are the highest among all algorithms. As an objective and quantitative expression method of abnormal driving behavior, the abnormal driving behavior spectrum can be combined with the artificial intelligence method to generate a massive library of abnormal driving behaviors. The imbalanced class boosting algorithm can address the imbalance problem of abnormal driving behavior data and has a better ability to identify abnormal driving behaviors compared to conventional algorithms.
-
图 1 道路不良驾驶行为实时辨识流程
Figure 1. Procedure of real-time identification of abnormal road driving behavior
图 2 无人机交通视频中提取车辆行驶轨迹
Figure 2. Vehicle driving trajectories extracted from unmanned aerial vehicle traffic video
图 3 不良跟驰特征参数分布
Figure 3. Distribution of abnormal car-following characteristic parameters
表 1 车辆行驶轨迹数据示例
Table 1. Example of vehicle driving trajectory data
帧数 纵向位置/m 横向位置/m 车长/m 车宽/m 纵向速度/(m·s-1) 横向速度/(m·s-1) 纵向加速度/(m·s-2) 横向加速度/(m·s-2) 前车编号 后车编号 车道编号 1 466 366.39 25.83 4.60 1.82 16.58 0.11 0.17 -0.03 182 197 5 1 467 366.98 25.83 4.60 1.82 16.59 0.11 0.18 -0.03 182 197 5 1 468 367.61 25.83 4.60 1.82 16.59 0.11 0.18 -0.03 182 197 5 1 469 368.25 25.84 4.60 1.82 16.60 0.10 0.18 -0.03 182 197 5 1 470 368.90 25.84 4.60 1.82 16.61 0.10 0.18 -0.03 182 197 5 1 471 369.55 25.85 4.60 1.82 16.62 0.10 0.19 -0.04 182 197 5 1 472 370.21 25.85 4.60 1.82 16.62 0.10 0.19 -0.04 182 197 5 
下载: 导出CSV
表 2 算法性能评价
Table 2. Performance evaluation of algorithms
算法 精确率/% 召回率/% 平衡分数 AUPRC AdaBoost 83.2 76.8 0.786 0.852 SMOTEAdaBoost 84.5 82.4 0.825 0.869 RUSAdaBoost 68.1 90.1 0.774 0.820 SMOTEBoost 79.9 85.6 0.818 0.895 RUSBoost 58.8 96.2 0.722 0.851 CUSBoost 85.9 94.4 0.861 0.930
下载: 导出CSV
-
[1] 丁雨蕾. 重特大交通事故特征及影响因素分析[D]. 南京: 东南大学, 2016.DING Yu-lei. Analyzing characteristics and influence factors of serious casualty traffic crashes in China[D]. Nanjing: Southeast University, 2016. (in Chinese). [2] WANG Ke, XUE Qing-wen, XING Ying-ying, et al. Improve aggressive driver recognition using collision surrogate measurement and imbalanced class boosting[J]. International Journal of Environmental Research and Public Health, 2020, 17(7): 2375-1-17. [3] SUN Wei, ZHANG Xiao-rui, PEETA S, et al. A real-time fatigue driving recognition method incorporating contextual features and two fusion levels[J]. IEEE Transactions on Intelligent Transportation Systems, 2017, 18(12): 3408-3420. doi: 10.1109/TITS.2017.2690914 [4] LIU Tian-chi, YANG Yan, HUANG Guang-bin, et al. Driver distraction detection using semi-supervised machine learning[J]. IEEE Transactions on Intelligent Transportation Systems, 2016, 17(4): 1108-1120. doi: 10.1109/TITS.2015.2496157 [5] CHANDRASIRI N P, NAWA K, ISHII A. Driving skill classification in curve driving scenes using machine learning[J]. Journal of Modern Transportation, 2016, 24(3): 196-206. doi: 10.1007/s40534-016-0098-2 [6] MOLCHANOV P, GUPTA S, KIM K, et al. Multi-sensor system for driver's hand-gesture recognition[C]//IEEE. 11th International Conference and Workshops on Automatic Face and Gesture Recognition. New York: IEEE, 2015: 1-8. [7] YAN Chao, COENEN F, YUE Yong, et al. Video-based classification of driving behavior using a hierarchal classification system with multiple features[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2016, 30(5): 1-33. [8] WANG Hong, ZHANG Chi, SHI Tian-wei, et al. Real-time EEG-based detection of fatigue driving danger for accident prediction[J]. International Journal of Neural Systems, 2015, 25(2): 1550002-1-14. [9] DENG Chao, WU Chao-zhong, LYU Neng-chao, et al. Driving style recognition method using braking characteristics based on hidden Markov model[J]. PLoS ONE, 2017, 12(8): e0182419-1-15. [10] XUE Qing-wen, WANG Ke, LU Jian-john, et al. Rapid driving style recognition in car-following using machine learning and vehicle trajectory data[J]. Journal of Advanced Transportation, 2019, 2019: 9085238-1-15. [11] LY M V, MARTIN S, TRIVEDI M M. Driver classification and driving style recognition using inertial sensors[C]//IEEE. 2013 IEEE Intelligent Vehicles Symposium. New York: IEEE, 2013: 1040-1045. [12] WU Ming-lin, ZHANG Sheng, DONG Yu-han. A novel model-based driving behavior recognition system using motion sensors[J]. Sensors, 2016, 16(10): 1-23. doi: 10.1109/JSEN.2016.2532220 [13] FERNANDEZ S, ITO T. Driver classification for intelligent transportation systems using fuzzy logic[C]//IEEE. 19th International Conference on Intelligent Transportation Systems. New York: IEEE, 2016: 1212-1216. [14] WANG Wen-shuo, XI Jun-qiang. A rapid pattern-recognition method for driving styles using clustering-based support vector machines[C]//IEEE. 2016 American Control Conference. New York: IEEE, 2016: 5270-5275. [15] AOUDE G S, DESARAJU V R, STEPHENS L H, et al. Driver behavior classification at intersections and validation on large naturalistic data set[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(2): 724-736. doi: 10.1109/TITS.2011.2179537 [16] KLUGER R, SMITH B L, PARK H, et al. Identification of safety-critical events using kinematic vehicle data and the discrete fourier transform[J]. Accident Analysis and Prevention, 2016, 96: 162-168. doi: 10.1016/j.aap.2016.08.006 [17] BEJANI M M, GHATEE M. A context aware system for driving style evaluation by an ensemble learning on smartphone sensors data[J]. Transportation Research Part C: Emerging Technologies, 2018, 89: 303-320. doi: 10.1016/j.trc.2018.02.009 [18] ZHAO Nan, MEHLER B, REIMER B, et al. An investigation of the relationship between the driving behavior questionnaire and objective measures of highway driving behavior[J]. Transportation Research Part F: Traffic Psychology and Behaviour, 2012, 15(6): 676-685. doi: 10.1016/j.trf.2012.08.001 [19] 田飞. 基于驾驶风格的换道行为谱分析研究[D]. 武汉: 武汉理工大学, 2016.TIAN Fei. Research on lane change behavioral spectral analysis based on driving style classification[D]. Wuhan: Wuhan University of Technology, 2016. (in Chinese). [20] 陈镜任, 吴业福, 吴冰. 基于车辆行驶数据的驾驶人行为谱分析方法[J]. 计算机应用, 2018, 38(7): 1916-1922, 1928. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY201807017.htmCHEN Jing-ren, WU Ye-fu, WU Bing. Driver behavior spectrum analysis method based on vehicle driving data[J]. Journal of Computer Applications, 2018, 38(7): 1916-1922, 1928. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY201807017.htm [21] WANG Wen-shuo, XI Jun-qiang, ZHAO Ding. Driving style analysis using primitive driving patterns with Bayesian nonparametric approaches[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(8): 2986-2998. doi: 10.1109/TITS.2018.2870525 [22] AKAI N, HIRAYAMA T, MORALES L Y, et al. Driving behavior modeling based on hidden markov models with driver's eye-gaze measurement and ego-vehicle localization[C]//IEEE. 2019 IEEE Intelligent Vehicles Symposium. New York: IEEE, 2019: 949-956. [23] LIANG Yu-lan, 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. doi: 10.1109/TITS.2007.895298 [24] LIU Yong-gang, WANG Ji-ming, ZHAO Pan, et al. Research on classification and recognition of driving styles based on feature engineering[J]. IEEE Access, 2019, 7: 89245-89255. doi: 10.1109/ACCESS.2019.2926593 [25] CHEN Zhi-jun, WU Chao-zhong, HUANG Zhen, et al. Dangerous driving behavior detection using video-extracted vehicle trajectory histograms[J]. Journal of Intelligent Transportation Systems, 2017, 21(5): 409-421. doi: 10.1080/15472450.2017.1305271 [26] HASHEMI A, SABA V, RESALAT S N. Real time driver's drowsiness detection by processing the EEG signals stimulated with external flickering light[J]. Basic and Clinical Neuroscience, 2014, 5(1): 22-27. [27] BEJANI M M, GHATEE M. Convolutional neural network with adaptive regularization to classify driving styles on smartphones[J]. IEEE Transactions on Intelligent Transportation Systems, 2020, 21(2): 543-552. doi: 10.1109/TITS.2019.2896672 [28] MOUKAFIH Y, HAFIDI H, GHOGHO M. Aggressive driving detection using deep learning-based time series classification[C]//IEEE. IEEE International Symposium on Innovations in Intelligent Systems and Applications. New York: IEEE, 2019: 1-5. [29] SHI Xiu-peng, WONG Y D, LI M Z F, et al. A feature learning approach based on XGBoost for driving assessment and risk prediction[J]. Accident Analysis and Prevention, 2019, 129: 170-179. doi: 10.1016/j.aap.2019.05.005 [30] CHAWLA N V, BOWYER K W, HALL L O, et al. SMOTE: synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research, 2002, 16: 321-357. doi: 10.1613/jair.953 [31] FREUND Y, SCHAPIRE R E. A decision-theoretic generalization of on-line learning and an application to boosting[J]. Journal of Computer and System Sciences, 1997, 55(1): 119-139. doi: 10.1006/jcss.1997.1504 [32] CHAWLA N V, LAZAREVIC A, HALL L O, et al. SMOTEBoost: improving prediction of the minority class in boosting[C]∥Springer. 7th European Conference on Principles of Data Mining and Knowledge Discovery in Databases. Berlin: Springer, 2003: 107-119. [33] SEIFFERT C, KHOSHGOFTAAR T M, HULSE J V, et al. Rusboost: a hybrid approach to alleviating class imbalance[J]. IEEE Transactions on Systems, Man, and Cybernetics—Part A: Systems and Humans, 2010, 40(1): 185-197. doi: 10.1109/TSMCA.2009.2029559 [34] RAYHAN F, AHMED S, MAHBUB A, et al. CUSBoost: cluster-based under-sampling with boosting for imbalanced classification[C]∥IEEE. 2nd International Conference on Computational Systems and Information Technology for Sustainable Solution. New York: IEEE, 2017: 1-6. [35] SAITO T, REHMSMEIER M. The precision-recall plot is more informative than the roc plot when evaluating binary classifiers on imbalanced datasets[J]. PloS ONE, 2015, 10(3): e0118432-1-21. -
点击查看大图
图(3) / 表(2)
计量
- 文章访问数: 1152
- HTML全文浏览量: 268
- PDF下载量: 516
- 被引次数: 0
