Drivers' Travel Pattern Mining Based on OBD Data
-
摘要: 传统的出行模式研究通常依靠问卷调查分析驾驶人出行特征，所得结果易受调查数据主观性影响，针对此问题基于北京市域范围内2个月共计3 570辆私家车的车载诊断数据，对驾驶人的不同出行模式进行分析并建模。通过长期采集的车辆各项参数，采用基于密度峰值的聚类算法进行聚类，将不同的驾驶人分为高频出行者、通勤出行者、长距偶发出行者以及危险出行者，并从平均出行距离、出行频次、百公里危险驾驶行为次数和出行时段等多维度进行分析，反映驾驶人行为的变化性和规律性。根据聚类的结果，使用多维离散隐马尔可夫模型进行建模并完成测试。测试表明，所提出的算法对于驾驶人出行模式的识别具有较高的准确性，对于4种类型的出行者，平均识别率超过91%，最高识别率可达94.5%。
-
关键词:
- 交通信息 /
- OBD数据 /
- 出行模式 /
- 聚类分析 /
- 基于密度峰值的聚类算法 /
- 多维离散隐马尔可夫模型
Abstract: The traditional travel pattern research mainly relies on questionnaires to analyze the driver's travel characteristics, the result of which is not objective. In order to solve the problem, the study analyzed and identifieddifferentdrivers' travel patterns based on the vehicle on-board diagnosticdata from 3 570 private cars in Beijing within two months. According to the parameters recorded from vehicles, a clustering algorithm called Clustering by Fast Search and Find of Density Peaks was used to classify different drivers into high-frequency travelers, commuting travelers, long-distance and occasional travelers and dangerous travelers, and analyzed from the aspects of average travel distance, travel frequency, travel time and dangerous driving behavior times of 100 km, to reflect the variability and regularity of driver's travel pattern. According to the clustering result, the multi-dimensional discrete Hidden Markov Model was used for modeling and measurement. Results indicate that the algorithm proposed shows good accuracy on the identification of drivers' travel patterns. For different kinds of drivers, the averagecorrect recognition rate exceed 91% while the highest recognition rete can reach 94.5%. -
图 5 不同类别百公里危险驾驶行为次数分布情况图
Figure 5. Distribution of dangerous driving behaviors in different categories
表 1 处理完成的数据格式
Table 1. Format of processed data
OBD ID 平均出行距离/km 出行天数 最频首次出行时段 最频末次出行时段 百公里危险驾驶行为次数 25BWFKA7 54 7 2 5 4 212A26G4 40 50 1 4 5 21E4JG67 33 27 1 4 12 
下载: 导出CSV
表 2 4个类别的具体数量
Table 2. Specific quantity of four categories
类别 类别1 类别2 类别3 类别4 共计 数量 1 071 1 071 1 248 180 3 570
下载: 导出CSV
表 3 训练数据格式
Table 3. Format of training data
出行模式 输人参数 状态 高频出行者 (32, 52, 1, 4, 5) 1 通勤出行者 (35, 33, 1, 4, 3) 2 长距偶发出行者 (53, 11, 2, 5, 3) 3 危险出行者 (57, 15, 2, 3, 15) 4
下载: 导出CSV
表 4 模型测试结果
Table 4. Results of model test
输人输出 高频出行者 通勤出行者 长距偶发出行者 危险出行者 高频出行者 298 27 15 0 通勤出行者 21 288 6 0 长距偶发出行者 2 4 335 3 危险出行者 0 2 18 52 总人数 321 321 374 55 准确率/% 92.8 89.7 89.6 94.5
下载: 导出CSV
-
[1] BHAT C. An analysis of evening commute stop-making behavior using repeated choice observations from a multi-day survey[J]. Transportation Research Part B: Methodological, 1999, 33(7): 495-510. doi: 10.1016/S0191-2615(99)00003-X [2] BOWMAN J L, BEN-AKIVA M E. Activity-based disaggregate travel demand model system with activity schedules[J]. Transportation Research Part A: Policy & Practice, 2001, 35(1): 1-28. http://www.sciencedirect.com/science/article/pii/S0965856499000439 [3] 柴彦威, 刘志林, 李峥嵘, 等. 中国城市的时空间结构[M]. 北京: 北京大学出版社, 2002.CHAI Yanwei, LIU Zhilin, LI Zhengrong, et al. The time-space structure of Chinese cities[M]. Beijing: Peking University Press, 2002. (in Chinese) [4] 陆化普, 孙智源, 屈闻聪. 大数据及其在城市智能交通系统中的应用综述[J]. 交通运输系统工程与信息, 2015, 15(5): 45-52. doi: 10.3969/j.issn.1009-6744.2015.05.007LU Huapu, SUN Zhiyuan, QU Wencong. Big data and its applications in urban intelligent transportation system[J]. Journal of Transportation Systems Engineering and Information Technology, 2015, 15(5): 45-52. (in Chinese) doi: 10.3969/j.issn.1009-6744.2015.05.007 [5] 聂霖杰. 基于移动数据的出行模式识别方法研究[D]. 长春: 吉林大学, 2020.NIE Linjie. Research on travel pattern recognition method based on mobile data[D]. Changchun: Jilin University, 2020. (in Chinese) [6] 刘晓波, 蒋阳升, 唐优华, 等. 综合交通大数据应用技术的发展展望[J]. 大数据, 2019, 5(3): 55-68. https://www.cnki.com.cn/Article/CJFDTOTAL-DSJU201903006.htmLIU Xiaobo, JIANG Yangsheng, TANG Youhua, et al. Development prospect of integrated transportation big data application technology[J]. Big Data Research, 2019, 5(3): 55-68. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-DSJU201903006.htm [7] 周素红, 杨利军. 广州城市居民通勤空间特征研究[J]. 城市交通, 2005, 3(1): 62-67. doi: 10.3969/j.issn.1672-5328.2005.01.015ZHOU Suhong, YANG Lijun. Study on spatial characteristic of commuting in Guangzhou[J]. Urban Transport of China, 2005, 3 (1): 62-67. (in Chinese) doi: 10.3969/j.issn.1672-5328.2005.01.015 [8] 何山, 朱健梅. 基于BP神经网络的个体出行模式识别研究[J]. 公路工程, 2014(5): 304-306. doi: 10.3969/j.issn.1674-0610.2014.05.067HE Shan, ZHU Jianmei. Research on individual travel mode detection by BP neural network[J]. Highway Engineering, 2014 (5): 304-306. (in Chinese) doi: 10.3969/j.issn.1674-0610.2014.05.067 [9] MA X, WANG Y, MCCORMACK E, et al. Understanding freight trip-chaining behavior using a spatial data-mining approach with GPS data[J]. Transportation Research Record Journal of the Transportation Research Board, 2016(2596): 44-54. http://www.researchgate.net/publication/306074834_Understanding_Freight_Trip-Chaining_Behavior_Using_a_Spatial_Data-Mining_Approach_with_GPS_Data [10] 马聪. 基于OBD技术的驾驶行为习惯评价方法研究[D]. 南京: 南京大学, 2016.MA Cong. Study on the evaluation method of driving behavior habits based on OBD technology[D]. Nanjing: Nanjing University, 2016. (in Chinese) [11] 何亚楠. 基于马尔科夫模型的出行目的地预测[D]. 长春: 吉林大学, 2017.HE Yanan. Predicting trip destination with Markov model[D]. Changchun: Jilin University, 2017. (in Chinese) [12] 卢笙. 基于车载诊断系统的轻型乘用车油耗影响因素量化分析[D]. 北京: 清华大学, 2018.LU Sheng. Quantitative analysis of factors affecting fuel consumption of Light-duty passenger vehicles based on On-board diagnostic approach[D]. Beijing: Tsinghua University, 2018. (in Chinese) [13] 张芷毓, 鹿应荣, 马晓磊, 等. 基于OBD数据的危险驾驶行为与建成环境空间关联研究[J]. 交通信息与安全, 2017, 35 (1): 35-43. http://www.jtxa.net/tiasn/paper/editpaper.do?flag=abstract&PAPERID=2016-00307ZHANG Ziyu, LU Yingrong, MA Xiaolei, et al. A study on the association between risky driving behavior and built environment using OBD data[J]. Journal of Transport Information and Safety, 2017, 35(1): 35-43. (in Chinese) http://www.jtxa.net/tiasn/paper/editpaper.do?flag=abstract&PAPERID=2016-00307 [14] KUMAR S, PAEFGEN J, WILHELM E, et al. Integrating on-board diagnostics speed data with sparse GPS measurements for vehicle trajectory estimation[C]. The SICE Annual Conference, Nagoya, Japan: IEEE, 2013. [15] 萧超武. 基于HMM的驾驶模式识别方法研究及应用[D]. 广州: 华南理工大学, 2015.XIAO Chaowu. Research and application of driving pattern recognition based on Hidden Markov model[D]. Guangzhou: South China University of Technology, 2015. (in Chinese) [16] 郭继孚, 孙建平, 温慧敏, 等. 基于车载OBD数据的小汽车出行特征分析: 以北京市为例[J]. 城市交通, 2017, 15(5): 70-77. https://www.cnki.com.cn/Article/CJFDTOTAL-CSJT201705014.htmGUO Jifu, SUN Jianping, WEN Huimin, et al. Characteristics of travel by car based on OBD data: A case study in Beijing[J]. Urban Transport of China, 2017, 15(5): 70-77. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-CSJT201705014.htm [17] 张梦歌, 蔡晓禹, 杜蕊. 驾驶行为数据驱动的城市道路交通安全风险辨识方法探讨[J]. 交通与运输, 2020, 36(2): 30-34. doi: 10.3969/j.issn.1671-3400.2020.02.008ZHANG Mengge, CAI Xiaoyu, DU Rui. Exploration of urban road safety risk identification method driven by driving behavior data[J]. Traffic & Transportation, 2020, 36(2): 30-34. (in Chinese) doi: 10.3969/j.issn.1671-3400.2020.02.008 [18] RODRIGUEZ A, LAIO A. Clustering by fast search and find of density peaks[J]. Science, 2014, 344(6191): 1492-1496. doi: 10.1126/science.1242072 [19] 张迎亚. 基于隐马尔可夫模型的车辆轨迹预测算法的研究[D]. 南京: 南京邮电大学, 2017.ZHANG Yingya. Research on the algorithm of vehicle trajectory prediction based on HMM[D]. Nanjing: Nanjing University of Posts and Telecommunications, 2017. (in Chinese) [20] 肖献强, 任春燕, 王其东. 基于隐马尔可夫模型的驾驶行为预测方法研究[J]. 中国机械工程, 2013, 24(21): 2972-2976. doi: 10.3969/j.issn.1004-132X.2013.21.025XIAO Xianqiang, REN Chunyan, WANG Qidong. Research on driving behavior prediction method based on HMM[J]. Chinese Journal of Mechanical Engineering, 2013, 24(21): 2972-2976. (in Chinese) doi: 10.3969/j.issn.1004-132X.2013.21.025 -
点击查看大图
图(6) / 表(4)
计量
- 文章访问数: 467
- HTML全文浏览量: 280
- PDF下载量: 14
- 被引次数: 0