Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2024, Vol. 58 Issue (11): 2393-2405    DOI: 10.3785/j.issn.1008-973X.2024.11.021
土木工程、交通工程     
基于手机信令数据的贝叶斯优化出行链识别
王殿海(),徐望,蔡正义*(),曾佳棋,黄宇浪
浙江大学 智能交通研究所，浙江 杭州 310058
Bayesian optimized trip chain identification based on mobile signaling data
Dianhai WANG(),Wang XU,Zhengyi CAI*(),Jiaqi ZENG,Yulang HUANG
Institute of Intelligent Transportation Systems, Zhejiang University, Hangzhou 310058, China
 全文: PDF(3769 KB)   HTML
摘要：

为了克服手机信令数据定位信息的时空不确定性对出行识别的影响，分析手机信令数据的时空特性，在时空阈值识别停留点方法的基础上，引入兴趣面（AOI）信息和基站位置数据，提出可变参数滑动窗口的出行停留点识别方法. 建立出行链模型，采用贝叶斯多目标优化获得模型的最佳参数，实现时空阈值的动态调整，提高识别的准确性. 为了验证方法的有效性，组织志愿者采集真实出行GPS数据和出行信息标签作为验证数据，与应用模型于对应手机信令数据后的结果进行对比. 研究结果表明，手机信令数据的采样在移动和静止状态下存在特性差异；相较于对比方法，提出的出行链识别方法在泛化性能和最优性能方面都表现出较小的误差和较高的识别率，尤其在识别率指标上，相比其他的最新算法改进了3%~26%.
关键词： 出行链识别;  贝叶斯优化;  手机信令数据;  兴趣面(AOI);  停留点;  OD识别    
Abstract:

The spatio-temporal characteristic of mobile signaling data was analyzed to mitigate the impact of spatio-temporal uncertainty in the location information of mobile signaling data on trip identification. Area of interest (AOI) and base station locations were incorporated based on the spatio-temporal threshold-based method for identifying stay points. A method for identifying stay points using a variable-parameter sliding window was proposed. A trip chain model was established, and Bayesian multi-objective optimization was employed to determine the best parameters. The dynamic adjustment of spatio-temporal thresholds was realized to enhance recognition accuracy. Volunteers were organized to collect real travel GPS data and travel information labels serving as validation data and compared with the results after applying the model to the corresponding mobile phone signaling data in order to validate the effectiveness of above-mentioned method. The research results indicate that there are characteristic differences in the sampling of mobile signaling data between mobile and stationary states. The proposed method show reduced errors and improved recognition rates in terms of both generalization and optimal performance compared with the benchmark methods. There is an improvement ranging from 3% to 26% especially in recognition rate compared to other state-of-the-art algorithms.
Key words: trip chain identification    Bayesian optimization    mobile signaling data    area of interest (AOI)    stay point    OD identification
收稿日期: 2023-09-04 出版日期: 2024-10-23
CLC:  U 491  
基金资助: 国家自然科学基金资助项目（52131202，71901193，52072340）.
通讯作者: 蔡正义     E-mail: wangdianhai@zju.edu.cn;caizhengyi@zju.edu.cn
作者简介: 王殿海（1962—），男，教授，博士，从事智能交通研究的研究. orcid.org/0000-0001-6066-2274.E-mail： wangdianhai@zju.edu.cn
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章　　
王殿海
徐望
蔡正义
曾佳棋
黄宇浪

引用本文:

王殿海,徐望,蔡正义,曾佳棋,黄宇浪. 基于手机信令数据的贝叶斯优化出行链识别[J]. 浙江大学学报(工学版), 2024, 58(11): 2393-2405.

Dianhai WANG,Wang XU,Zhengyi CAI,Jiaqi ZENG,Yulang HUANG. Bayesian optimized trip chain identification based on mobile signaling data. Journal of ZheJiang University (Engineering Science), 2024, 58(11): 2393-2405.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2024.11.021    或    https://www.zjujournals.com/eng/CN/Y2024/V58/I11/2393

图 1  出行链相关元素的示意图
图 2  出行链识别算法框架的示意图
图 3  停留点识别的流程图
图 4  基站分布均匀度计算的示意图
图 5  杭州市主城区部分AOI的分布示意图
图 6  出行段分类示意图
图 7  信令数据经过模型处理后获得的出行链信息示意图
时间状态坐标
00:00:00—08:33:00停留120.1506*, 30.3329*
08:33:00—09:12:19出行120.1506*, 30.3329*—
120.1181*, 30.2682*
09:12:19—22:14:36停留120.1181*, 30.2682*
22:14:36—22:52:51出行120.1181*, 30.2682*—
120.1506*, 30.3318*
22:52:51—23:59:58停留120.1506*, 30.3318*
表 1  某用户某天的出行链信息
用户编号时间戳经度纬度
000142****fb11/9 12:55:11120.15334630.339069
000142****fb11/9 12:55:36120.16072330.338450
000142****fb11/9 12:56:58120.16306730.337942
000142****fb11/9 12:59:22120.16072330.338450
000142****fb11/9 13:07:13120.16281230.336914
表 2  手机信令数据的样例记录
图 8  某次出行的手机信令和GPS数据对比
图 9  手机信令数据的定位误差分布
图 10  手机信令采样频率与实际移动速度的关系
图 11  出行轨迹APP的定位与标注页面
用户id开始时间结束时间起点经度起点纬度终点经度终点纬度
1985*******2022-11-23 8:30:122022-11-23 9:09:53120.1512**30.3325**120.1286**30.2687**
1985*******2022-11-23 21:43:002022-11-23 21:48:00120.1161**30.2674**120.1251**30.2791**
1985*******2022-11-23 23:44:372022-11-24 0:04:43120.1251**30.2791**120.1508**30.3325**
1985*******2022-11-24 8:30:082022-11-24 9:09:28120.1512**30.3326**120.1186**30.2688**
1985*******2022-11-24 22:03:342022-11-24 22:59:32120.1164**30.2669**120.1509**30.3325**
表 3  某用户的出行OD标注数据示例
优化结果系数系数参数值阈值参数值Y
i = 0i = 1i = 2i = 3i = 4geu
取值范围ai(400, 650)(?0.7, 0.7)(?39, 39)(?4.2, 4.2)(?4.2, 4.2)(8, 15)(150, 300)(100, 240)—
bi(25, 35)(?0.04, 0.04)(?2.3, 2.3)(?0.25, 0.25)(?0.25, 0.25)
优化结果1ai617?0.42?13?3.13.314.0252196?0.006
bi32.5?0.0171.4?0.19?0.16
优化结果2ai622?0.37?36?1.52.512.8155137?0.006
bi33.0?0.002?1.0?0.22?0.18
优化结果3ai632?0.45?38?0.7?2.112.7257187?0.006
bi32.7?0.032?1.1?0.16?0.21
优化结果4ai628?0.36?34?1.50.212.7270181?0.005
bi33.2?0.004?0.9?0.21?0.07
优化结果5ai620?0.63?24?2.02.412.6202199?0.007
bi32.1?0.031?0.2?0.15?0.24
表 4  优化获得的损失函数及各项参数值
统计项$ \Delta {{{D}}_{\text{o}}}{+}\Delta {{{D}}_{\text{d}}} $/m$ \Delta {{{T}}_{\text{o}}}{+}\Delta {{{T}}_{\text{d}}} $/minM ? A
25%分位数564.613.533?0.7205
75%分位数603.014.326?0.6716
75%分位数?25%分位数38.40.7930.0489
计算权重25%25%50%
归一化系数1/15701/321
表 5  各项性能度量指标组合的变化、权重及系数
性能算法误差指标识别率指标
$ \Delta {{{D}}_{\text{o}}} $/m$ \Delta {{{D}}_{\text{d}}} $/m$ \Delta {{{T}}_{\text{o}}} $/min$ \Delta {{{T}}_{\text{d}}} $/minPRAM
最优性能Marra’s Heuristic5175978.68.20.5950.6280.8220.237
Trackintel2752827.27.10.6810.7690.8290.304
本文框架固定阈值法2673177.16.90.7740.8030.8720.168
本文方法2672597.06.00.8300.8260.8750.125
泛化性能Marra’s Heuristic4925489.37.30.5980.6360.8150.255
Trackintel2442346.87.30.6190.7520.8180.397
本文框架固定阈值法2612926.86.90.7810.7920.8410.164
本文方法2512486.85.90.8180.8040.8560.136
表 6  出行链识别算法的性能比较
数据集部分参数误差指标识别率指标
a0a1b0b3ge$ \Delta {{{D}}_{\text{o}}} $/m$ \Delta {{{D}}_{\text{d}}} $/m$ \Delta {{{T}}_{\text{o}}} $/min$ \Delta {{{T}}_{\text{d}}} $/minPRAM
训练1619?0.4530.8?0.2413.32622722596.95.70.8190.8140.8720.125
测试12593297.87.10.8080.8030.8680.125
训练2611?0.3831.9?0.1414.32152652517.06.40.8300.8260.8660.131
测试22873158.15.80.8260.8090.8940.099
训练3622?0.632.9?0.1412.91982662577.46.10.8370.8300.8790.118
测试32582456.46.10.7600.8160.8820.191
训练4632?0.4333.2?0.0814.21582732557.06.50.8090.8280.8860.142
测试42702387.15.50.8270.7770.8250.120
训练5623?0.531.9?0.1612.62252662556.96.20.8150.8130.8770.126
测试52612547.57.20.8150.8380.8660.152
平均训练——————268.4255.47.046.180.8220.8220.8760.128
平均测试267276.27.386.340.8070.8090.8670.137
表 7  出行链识别算法稳定性实验（五折交叉验证）
1 SHIFTAN Y Practical approach to model trip chaining[J]. Transportation Research Record: Journal of the Transportation Research Board, 1998, 1645 (1): 17- 23
doi: 10.3141/1645-03
2 杨励雅, 李娟 居民出行链、出行方式与出发时间联合选择的交叉巢式Logit模型[J]. 北京大学学报: 自然科学版, 2017, 53 (4): 722- 730
YANG Liya, LI Juan Cross-nested logit model for the joint choice of residential location, travel mode, and departure time[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 2017, 53 (4): 722- 730
3 HUANG Y, WANG D, XU W, et al Accurate map matching method for mobile phone signaling data under spatio-temporal uncertainty[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 25 (2): 1418- 1429
4 CAI Z, WANG D, CHEN X (M) A novel trip coverage index for transit accessibility assessment using mobile phone data[J]. Journal of Advanced Transportation, 2017, 2017 (1): 1- 14
5 WAN L, YANG T, JIN Y, et al Estimating commuting matrix and error mitigation: a complementary use of aggregate travel survey, location-based big data and discrete choice models[J]. Travel Behaviour and Society, 2021, 25: 102- 111
doi: 10.1016/j.tbs.2021.04.012
6 YU Q, LI W, YANG D, et al Mobile phone data in urban commuting: a network community detection-based framework to unveil the spatial structure of commuting demand[J]. Journal of Advanced Transportation, 2020, 2020 (1): 1- 15
7 CHIN K, HUANG H, HORN C, et al Inferring fine-grained transport modes from mobile phone cellular signaling data[J]. Computers, Environment and Urban Systems, 2019, 77 (2): 101348
8 陈艳艳, 张野, 孙浩冬 基于手机信令数据的旅游客流特征分析[J]. 北京工业大学学报, 2022, 48 (8): 842- 850
CHEN Yanyan, ZHANG Ye, SUN Haodong Analysis of tourist flow characteristics based on mobile phone signaling data[J]. Journal of Beijing University of Technology, 2022, 48 (8): 842- 850
9 CHEN X, WAN X, LI Q, et al Trip-chain-based travel-mode-shares-driven framework using cellular signaling data and web-based mapping service data[J]. Transportation Research Record: Journal of the Transportation Research Board, 2019, 2673 (3): 51- 64
doi: 10.1177/0361198119834006
10 BONNETAIN L, FURNO A, EL FAOUZI N E, et al TRANSIT: fine-grained human mobility trajectory inference at scale with mobile network signaling data[J]. Transportation Research Part C: Emerging Technologies, 2021, 130: 103257
doi: 10.1016/j.trc.2021.103257
11 MARTIN H, HONG Y, WIEDEMANN N, et al Trackintel: an open-source Python library for human mobility analysis[J]. Computers, Environment and Urban Systems, 2023, 101 (11): 101938
12 JIANG H, YANG F, SU W, et al Activity location recognition from mobile phone data using improved HAC and Bi-LSTM[J]. IET Intelligent Transport Systems, 2022, 16 (10): 1364- 1379
doi: 10.1049/itr2.12211
13 杨飞, 姜海航, 姚振兴, 等 基于手机信令数据的出行端点识别效果评估[J]. 西南交通大学学报, 2021, 56 (5): 928- 936
YANG Fei, JIANG Haihang, YAO Zhenxing, et al Evaluation of activity location recognition using cellular signaling data[J]. Journal of Southwest Jiaotong University, 2021, 56 (5): 928- 936
14 NI L, WANG X, CHEN X A spatial econometric model for travel flow analysis and real-world applications with massive mobile phone data[J]. Transportation Research Part C: Emerging Technologies, 2018, 86: 510- 526
doi: 10.1016/j.trc.2017.12.002
15 林楠, 尹凌, 赵志远 基于滑动窗口的手机定位数据个体停留区域识别算法[J]. 地球信息科学学报, 2018, 20 (6): 762- 771
LIN Nan, YIN Ling, ZHAO Zhiyuan Detecting individual stay areas from mobile phone location data based on moving windows[J]. Journal of Geo-information Science, 2018, 20 (6): 762- 771
doi: 10.12082/dqxxkx.2018.180087
16 MARRA A D, BECKER H, AXHAUSEN K W, et al Developing a passive GPS tracking system to study long-term travel behavior[J]. Transportation Research Part C: Emerging Technologies, 2019, 104: 348- 368
doi: 10.1016/j.trc.2019.05.006
17 JIANG H, YANG F, ZHU X, et al Improved F-DBSCAN for trip end identification using mobile phone data in combination with base station density[J]. Journal of Advanced Transportation, 2022, 2022 (1): 1- 17
18 HUANG Z, LING X, WANG P, et al Modeling real-time human mobility based on mobile phone and transportation data fusion[J]. Transportation Research Part C: Emerging Technologies, 2018, 96: 251- 269
doi: 10.1016/j.trc.2018.09.016
19 FEKIH M, BELLEMANS T, SMOREDA Z, et al A data-driven approach for origin–destination matrix construction from cellular network signalling data: a case study of Lyon region (France)[J]. Transportation, 2021, 48 (4): 1671- 1702
doi: 10.1007/s11116-020-10108-w
20 王梅红, 侯笑宇, 司连法, 等 地理空间数据结合手机信令等多源数据刻画城市居民出行特征[J]. 测绘通报, 2022, (5): 162- 165
WANG Meihong, HOU Xiaoyu, SI Lianfa, et al Geospatial data combined with multi-source data such as mobile phone signaling data to depict the travel characteristics of city residents[J]. Bulletin of Surveying and Mapping, 2022, (5): 162- 165
21 HUAN L, ZHENBO L. Identification method of residents’ medical travel behavior characteristics driven by mobile signaling data: a case study of Kunshan [C/OL]// 5th International Conference on Information Science, Computer Technology and Transportation . Shenyang: IEEE, 2020: 198-207[2022-11-15]. https://ieeexplore.ieee.org/document/9363760/.
22 SIŁA-NOWICKA K, VANDROL J, OSHAN T, et al Analysis of human mobility patterns from GPS trajectories and contextual information[J]. International Journal of Geographical Information Science, 2016, 30 (5): 881- 906
doi: 10.1080/13658816.2015.1100731
23 HORN C, KLAMPFL S, CIK M, et al Detecting outliers in cell phone data: correcting trajectories to improve traffic modeling[J]. Transportation Research Record: Journal of the Transportation Research Board, 2014, 2405 (1): 49- 56
doi: 10.3141/2405-07
24 CAI M, ZHANG Z, XIONG C, et al An adaptive staying point recognition algorithm based on spatiotemporal characteristics using cellular signaling data[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23 (8): 10458- 10468
doi: 10.1109/TITS.2021.3094636
25 BACHIR D, KHODABANDELOU G, GAUTHIER V, et al Inferring dynamic origin-destination flows by transport mode using mobile phone data[J]. Transportation Research Part C: Emerging Technologies, 2019, 101: 254- 275
doi: 10.1016/j.trc.2019.02.013
26 周常勇. 基于移动信令数据的城市交通出行轨迹匹配技术[D]. 成都: 西南交通大学, 2016.
ZHOU Changyong. On the cellular signaling based transport trajectory matching technologies in urban area [D]. Chengdu: Southwest Jiaotong University, 2016.
27 ZAR J H. Biostatistical analysis [M]. 5th ed. Upper Saddle River: Prentice Hall, 2010: 614-617.
28 MARDIA K V, JUPP P E. Directional statistics [M]. Chichester: Wiley, 2000: 14-19.
29 BERGSTRA J, BARDENET R, BENGIO Y, et al. Algorithms for hyper-parameter optimization [C]// 25th Annual Conference on Neural Information Processing Systems . Granada: Neural Information Processing Systems Foundation, 2011: 2546-2554.
[1] 季廷炜,莫邵昌,谢芳芳,张鑫帅,蒋逸阳,郑耀. 基于高斯过程回归的机翼/短舱一体化气动优化[J]. 浙江大学学报(工学版), 2023, 57(3): 632-642.
[2] 倪玲霖, 张帅超, 陈喜群. 基于手机信号令数据的居民出行空间效应[J]. 浙江大学学报(工学版), 2017, 51(5): 887-895.