由于出租车行业普遍存在轮班制，一条出租车的营运轨迹并不完全是一位出租司机的运营轨迹，因此，采用单一的出租车轨迹数据源无法深入分析出租车司机个体及群体的移动行为特征。卫星导航定位系统和地面移动通讯网络，均可对道路移动目标进行跟踪定位，形成不同质量的时空轨迹数据源，多源数据给出租司机移动行为分析提供了一种新的思路。提出一种面向出租车司机的多源时空轨迹的同轨分析建模方法，集成上述两类数据，增强轨迹语义，并利用提出的多源轨迹之间的时空相似度度量指标，对出租车GNSS (global navigation satellite system)轨迹数据和手机Cell-ID数据进行关联分析与同质性检验建模，重建“出租车-司机-手机”关联关系并探测出租车司机出车、收车的时空位置。采用2016年8月4日采集的北京市出租车GNSS轨迹和移动手机信令数据开展验证实验，结果表明：1)本文方法可有效识别“出租车-司机-手机”的关联关系，其中，基于GNSS轨迹与Cell-ID轨迹匹配的“车辆-手机”关联识别精度F1分数为0.91，基于Cell-ID轨迹聚类的“手机-出租车”关联识别精度F1分数为0.94；2)同一出租车的轮班司机的交接间隔时长呈伽马分布，平均1.5 h左右，交接位置的平均间隔距离约91 m，出租车司机交接班点沿交通枢纽呈现空间聚集现象。本文结果与人工解译结果具有高度的一致性，验证了本文方法的有效性。
王卫锋
,
胡靖昊
,
贺琰
,
宋现锋
,
芮小平
,
刘军利
,
朱克忞
. 出租车司机的多源轨迹同轨分析[J]. 中国科学院大学学报, 2023
, 40(3)
: 313
-321
.
DOI: 10.7523/j.ucas.2021.0078
Due to the shifts among partner taxi drivers, a taxi GNSS (global navigation satellite system) trajectory is usually not a driver's operational trajectory, and thus it is impossible to deeply analyze the mobile behavior characteristics of individuals or community with a single GNSS data source. Both a satellite navigation and positioning system and a ground mobile communication network can track and locate the moving objects on the road, forming the spatio-temporal trajectory data sources of different qualities. In this paper, we propose a novel synchronized trajectory analysis for multi-source temporal and spatial trajectories of taxi drivers, integrating the above two kinds of data to enhance trajectory semantics and extract taxi driver travel space. Based on the track of the points accumulated weighted similarity of similarity metrics, in which the spatial association analysis and homogeneity test analysis were carried out between a taxi GNSS trajectory and a mobile Cell-ID trajectory and correspondingly the association of "taxi-driver-cellphone" was reconstructed and the space-time position of the taxi driver's start-of-work and end-of-work was detected. The taxi GNSS data of Beijing Taxi and the mobile signaling data of Beijing Mobile collected on August 4, 2016 were used for experimental analysis. The statistical results show that the F1 score of identifying cellphone Cell-ID trajectories by matching a GNSS trajectory is 0.91, and the F1 score of recognizing cellphone user by clustering analysis is 0.94. The averaged time and space difference between drivers during their shifting a taxi are 1.5 h and 91 m respectively. Moreover, the handover points of taxi drivers are densely distributed nearby transportation hubs. The modeling results are highly consistent with the manually interpreted ones, well verifying the effectiveness of the proposed method.
[1] 许飒, 杨新征, 彭虓. 网约车与巡游出租车抽成比例研究:基于网约车司企分配模式视角的分析[J]. 价格理论与实践, 2019(10):137-140. DOI:10.19851/j.cnki.cn11-1010/f.2019.10.032.
[2] 万传荣, 孙英隽. 互联网背景下网约车与传统出租车行业的博弈分析[J]. 电子商务, 2018(5):8-9. DOI:10.14011/j.cnki.dzsw.2018.05.004.
[3] Su R, Fang Z. A review of studies in taxi mobility and e-hailing taxi service[J]. Journal of Smart Cities, 2019, 4(1):2-6. DOI:10.18063/JSC.2019.01.002.
[4] Zheng Z, Rasouli S, Timmermans H. Modeling taxi driver search behavior under uncertainty[J]. Travel Behaviour and Society, 2021, 22:207-218. DOI:10.1016/j.tbs.2020.09.008.
[5] Kottayil S S, Tsoleridis P, Rossa K, et al. Investigation of driver route choice behaviour using bluetooth data[J]. Transportation Research Procedia, 2020, 48:632-645. DOI:10.1016/j.trpro.2020.08.065.
[6] Shahverdy M, Fathy M, Berangi R, et al. Driver behavior detection and classification using deep convolutional neural networks[J]. Expert Systems With Applications, 2020, 149:113240. DOI:10.1016/j.eswa.2020.113240.
[7] 姚德中. 时空轨迹数据的关联挖掘技术研究[D]. 武汉:华中科技大学, 2016.
[8] 吴华意, 黄蕊, 游兰, 等. 出租车轨迹数据挖掘进展[J]. 测绘学报, 2019, 48(11):1341-1356. DOI:10.11947/j.AGCS.2019.20190210.
[9] Ghahramani M, Zhou M C, Hon C T. Mobile phone data analysis:a spatial exploration toward hotspot detection[J]. IEEE Transactions on Automation Science and Engineering, 2019, 16(1):351-362. DOI:10.1109/TASE.2018.2795241.
[10] 曾昭博, 王睿, 刘伟, 等. 基于模糊平均综合相似度的航迹关联算法[J]. 电讯技术, 2009, 49(8):9-12. DOI:10.3969/j.issn.1001-893x.2009.08.003.
[11] Alt H, Godau M. Computing the Fréchet distance between two polygonal curves[J]. International Journal of Computational Geometry & Applications, 1995, 5(1n02):75-91. DOI:10.1142/s0218195995000064.
[12] Zhang Z, Huang K Q, Tan T N. Comparison of similarity measures for trajectory clustering in outdoor surveillance scenes[C]//18th International Conference on Pattern Recognition (ICPR'06). August 20-24, 2006, Hong Kong, China. IEEE, 2006:1135-1138. DOI:10.1109/ICPR.2006.392.
[13] Vlachos M, Gunopulos D, Das G. Rotation invariant distance measures for trajectories[C]//KDD '04:Proceedings of the 10th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2004:707-712. DOI:10.1145/1014052.1014144.
[14] Mao Y C, Zhong H S, Xiao X J, et al. A segment-based trajectory similarity measure in the urban transportation systems[J]. Sensors (Basel, Switzerland), 2017, 17(3):524. DOI:10.3390/s17030524.
[15] Chen L, Ng R. On the marriage of lp-norms and edit distance[M]//Proceedings 2004 VLDB Conference. Amsterdam:Elsevier, 2004:792-803. DOI:10.1016/b978-012088469-8.50070-x.
[16] Nanni M, Pedreschi D. Time-focused clustering of trajectories of moving objects[J]. Journal of Intelligent Information Systems, 2006, 27(3):267-289. DOI:10.1007/s10844-006-9953-7.
[17] Leontiadis I, Lima A, Kwak H, et al. From cells to streets:estimating mobile paths with cellular-side data[C]//CoNEXT '14:Proceedings of the 10th ACM International on Conference on Emerging Networking Experiments and Technologies. 2014:121-132. DOI:10.1145/2674005.2674982.
[18] Gong X R, Huang Z, Wang Y L, et al. High-performance spatiotemporal trajectory matching across heterogeneous data sources[J]. Future Generation Computer Systems, 2020, 105:148-161. DOI:10.1016/j.future.2019.11.027.
[19] Hanley J A, McNeil B J. The meaning and use of the area under a receiver operating characteristic (ROC) curve[J]. Radiology, 1982, 143(1):29-36. DOI:10.1148/radiology.143.1.7063747.
[20] Zhang D Z, Lee K, Lee I. Hierarchical trajectory clustering for spatio-temporal periodic pattern mining[J]. Expert Systems With Applications, 2018, 92:1-11. DOI:10.1016/j.eswa.2017.09.040.
[21] 苏卫星, 朱云龙, 刘芳, 等. 时间序列异常点及突变点的检测算法[J]. 计算机研究与发展, 2014, 51(4):781-788. DOI:10.7544/issn1000-1239.2014.20120542.
[22] Rybski D, Neumann J. A review on the Pettitt test[M]//In Extremis. Berlin, Heidelberg:Springer Berlin Heidelberg, 2010:202-213. DOI:10.1007/978-3-642-14863-7_10.
[23] Pettitt A N. A non-parametric approach to the change-point problem[J]. Journal of the Royal Statistical Society:Series C (Applied Statistics), 1979, 28(2):126-135. DOI:10.2307/2346729.
[24] Nachar N. The Mann-Whitney U:a test for assessing whether two independent samples come from the same distribution[J]. Tutorials in Quantitative Methods for Psychology, 2008, 4(1):13-20. DOI:10.20982/tqmp.04.1.p013.
[25] 周文宏. 影响GPS定位精度的因素及改进方法[J]. 安徽科技, 2009(9):49-51. DOI:10.3969/j.issn.1007-7855.2009.09.025.
[26] Pan G, Qi G D, Zhang W S, et al. Trace analysis and mining for smart cities:Issues, methods, and applications[J]. IEEE Communications Magazine, 2013, 51(6):120-126. DOI:10.1109/MCOM.2013.6525604.
