Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2023, Vol. 57 Issue (8): 1607-1617    DOI: 10.3785/j.issn.1008-973X.2023.08.013
土木工程、交通工程     
基于最优汇集时间间隔的城市间断交通流预测
王殿海(),谢瑞,蔡正义*()
浙江大学 智能交通研究所，浙江 杭州 310058
Prediction of urban interrupted traffic flow based on optimal convergence time interval
Dian-hai WANG(),Rui XIE,Zheng-yi CAI*()
Intelligent Transportation Research Institute, Zhejiang University, Hangzhou 310058, China
 全文: PDF(2168 KB)   HTML
摘要：

针对城市交通流受信号控制的影响而呈现出间断性、周期性和随机性的特点，提出基于最优汇集时间间隔的城市间断交流预测方法. 该方法首先基于傅里叶变换和自相关分析获取城市间断交通流的信号控制周期，再利用交叉验证均方差模型确定最优汇集时间间隔与信号周期的关系，在此基础上提出融合贝叶斯神经网络和深度学习模型的LSTM-BConv预测模型. 基于实测数据的实验结果表明：1）基于最优汇集时间间隔统计交通流数据能有效提升城市间断交通流预测模型的预测精度；2）城市间断交通流数据的最优汇集时间间隔为交通信号控制周期的倍数；3）对比试验结果表明，LSTM-BConv预测模型优于常见的预测模型，平均绝对百分比误差提升了4.57%. 预测结果可以为信号控制方案的优化提供参考依据.
关键词： 城市间断流;  最优汇集时间间隔;  信号控制周期;  短时交通流预测;  贝叶斯卷积神经网络    
Abstract:

An urban discontinuous traffic flow prediction method based on optimal convergence time interval was proposed, aiming at the discontinuity, periodicity and randomness of urban traffic flow affected by signal control. Firstly, the signal control period of urban discontinuous traffic flow was obtained based on Fourier transform and autocorrelation analysis, and then cross validation mean square error model was used to determine the relationship between optimal convergence time interval and signal period. A LSTM_BConv prediction model combining Bayesian neural network and deep learning model was proposed based on the previous analysis. Experimental results show that: 1) Traffic flow data statistics based on optimal convergence time interval can effectively improve the prediction accuracy of urban discontinuous traffic flow prediction model; 2) The optimal convergence time interval of urban discontinuous traffic flow data is a multiple of the traffic signal control cycle; 3) Comparison test results show that LSTM_BConv model is superior to common benchmark models, and the average absolute percentage error is increased by 4.57%. The prediction results can provide reference for the optimization of signal control scheme.
Key words: urban interrupted flow    optimal convergence time interval    signal control cycle    short term traffic flow prediction    Bayesian convolutional neural network
收稿日期: 2022-12-03 出版日期: 2023-08-31
CLC:  U 491  
基金资助: 国家自然科学基金资助项目(52131202, 71901193, 52072340); 浙江省"尖兵""领雁"研发攻关计划资助项目（2023C01240, 2023C03155)
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，博导，从事交通控制研究. 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
谢瑞
蔡正义

引用本文:

王殿海,谢瑞,蔡正义. 基于最优汇集时间间隔的城市间断交通流预测[J]. 浙江大学学报(工学版), 2023, 57(8): 1607-1617.

Dian-hai WANG,Rui XIE,Zheng-yi CAI. Prediction of urban interrupted traffic flow based on optimal convergence time interval. Journal of ZheJiang University (Engineering Science), 2023, 57(8): 1607-1617.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2023.08.013    或    https://www.zjujournals.com/eng/CN/Y2023/V57/I8/1607

图 1  直行交通流随时间的变化图
图 2  交叉验证均方差（CVMSE）计算过程
图 3  交叉口不同转向交通流的相关性
图 4  传统神经网络结构
图 5  贝叶斯神经网络结构
图 6  贝叶斯卷积神经网络卷积过程
图 7  LSTM_BConv预测模型框架
图 8  基于最优汇集时间间隔的城市间断交通流预测模型框架图
图 9  目标交叉口真实地理位置分布
图 10  目标交叉口傅里叶变换幅值图
图 11  目标交叉口与不同滞后阶数的相关性
图 12  目标交叉口不同时段实际信号控制周期
图 13  目标交叉口不同汇集时间间隔的交叉验证均方差
图 14  交叉口1 LSTM_BConv模型预测精度MAPE
图 15  交叉口2 LSTM_BConv模型预测精度MAPE
图 16  预测模型在不同汇集时间间隔下的预测MAPE对比
转向 统计学 机器学习 深度学习
HA ARIMA SVR KNN CapsNet GRU STAWnet LSTM_BConv
S_L 3.35 3.97 3.24 3.30 3.36 3.12 3.04 3.06
S_S 7.81 11.35 7.54 8.06 9.87 7.30 7.92 7.23
N_L 3.08 3.53 2.98 3.09 3.01 2.93 2.89 2.87
N_S 9.59 13.75 9.26 9.53 12.43 8.65 9.48 8.61
W_L 3.10 4.25 3.20 3.30 4.05 3.06 3.02 3.01
W_S 1.72 1.92 1.69 1.74 1.81 1.60 1.64 1.58
E_L 2.23 2.26 2.15 2.19 2.14 2.05 2.07 2.03
E_S 3.91 3.87 3.60 3.72 3.69 3.48 3.62 3.41
Mean 4.35 5.61 4.21 4.37 5.05 4.02 4.21 3.98
表 1  不同预测模型在交叉口1的预测MAE对比
转向 统计学 机器学习 深度学习
HA ARIMA SVR KNN CapsNet GRU STAWnet LSTM_BConv
S_L 4.24 4.91 4.09 4.15 4.21 3.91 3.84 3.82
S_S 10.25 14.24 9.85 10.37 12.20 9.47 10.09 9.36
N_L 3.85 4.37 3.71 3.83 3.71 3.64 3.62 3.57
N_S 12.27 16.74 11.84 12.09 15.04 11.21 12.23 11.19
W_L 3.98 5.26 4.07 4.20 5.16 3.94 3.89 3.88
W_S 2.09 2.39 2.04 2.10 2.17 1.99 2.00 1.94
E_L 2.76 2.83 2.63 2.70 2.62 2.51 2.53 2.48
E_S 4.75 4.80 4.39 4.55 4.49 4.23 4.41 4.17
Mean 5.52 6.94 5.33 5.50 6.20 5.11 5.33 5.05
表 2  不同预测模型在交叉口1的预测RMSE对比
转向 统计学 机器学习 深度学习
HA ARIMA SVR KNN CapsNet GRU STAWnet LSTM_BConv
S_L 26.17 32.36 25.63 26.17 24.09 24.50 24.17 21.61
S_S 15.83 26.23 13.65 14.63 18.63 13.14 14.23 12.67
N_L 32.72 38.89 35.06 36.10 31.72 33.14 34.10 28.45
N_S 19.57 30.61 17.47 17.87 24.16 16.61 17.59 15.54
W_L 29.64 43.02 36.63 38.32 43.90 32.73 34.00 29.18
W_S 41.02 44.29 60.60 64.07 45.96 54.29 39.59 36.88
E_L 36.77 36.24 42.46 44.31 33.03 39.78 42.65 30.56
E_S 31.27 33.64 30.12 31.45 26.99 28.77 30.44 25.34
Mean 29.12 35.66 32.70 34.12 31.06 30.37 29.60 25.03
表 3  不同预测模型在交叉口1的预测MAPE对比
图 17  消融实验对比结果
1 郭敏, 肖翔, 蓝金辉 道路交通流短时预测方法综述[J]. 自动化技术与应用, 2009, 28 (6): 8- 9
GUO Min, XIAO Xiang, LAN Jin-hui Summary of short-term prediction methods for road traffic flow[J]. Automation Technology and Application, 2009, 28 (6): 8- 9
doi: 10.3969/j.issn.1003-7241.2009.06.003
2 王均, 关伟 基于Kalman滤波的城市环路交通流短时预测研究[J]. 交通信息与安全, 2006, 24 (5): 16- 19
WANG Jun, GUAN Wei Research on short-term prediction on of urban ring road traffic flow based on Kalman filtering[J]. Traffic Information and Safety, 2006, 24 (5): 16- 19
doi: 10.3963/j.issn.1674-4861.2006.05.005
3 EMAMI A, SARVI M, BAGLOEE S A Short-term traffic flow prediction based on faded memory Kalman Filter fusing data from connected vehicles and Bluetooth sensors[J]. Simulation Modelling Practice and Theory, 2020, 102: 1- 17
4 KUMAR S V, VANAJAKSHI L Short-term traffic flow prediction using seasonal ARIMA model with limited input data[J]. European Transport Research Review, 2015, 7 (3): 1- 9
5 GIRAKA O, SELVARAJ V K Short-term prediction of intersection turning volume using seasonal ARIMA model[J]. Transportation Letters, 2020, 12 (7): 483- 490
doi: 10.1080/19427867.2019.1645476
6 傅贵, 韩国强, 逯峰, 等 基于支持向量回归的短时交通流预模型[J]. 华南理工大学学报: 自然科学版, 2013, 41 (9): 71- 76
FU Gui, HAN Guo-qiang, LU Feng, et al Short term traffic flow prediction model based on support vector machine regression[J]. Journal of South China University of Technology: Natural Science Edition, 2013, 41 (9): 71- 76
7 CHENG S F, LU F, PENG P, et al Short-term traffic forecasting: an adaptive ST-KNN model that considers spatial heterogeneity[J]. Computers Environment and Urban Systems, 2018, 71: 186- 198
doi: 10.1016/j.compenvurbsys.2018.05.009
8 ZHAO Z, CHEN W H, WU X M, et al LSTM network: a deep learning approach for short-term traffic forecast[J]. IET Intelligent Transport Systems, 2017, 11 (2): 68- 75
doi: 10.1049/iet-its.2016.0208
9 殷齐, 丁飞, 朱跃, 等 基于CNN与多尺度特征融合城市交通流预测模型[J]. 计算机技术与发展, 2022, 32 (10): 175- 181
YIN Qi, DING Fei, ZHU Yue, et al An urban traffic flow prediction model based on CNN and Multi-scale feature fusion[J]. Computer Technology and Development, 2022, 32 (10): 175- 181
doi: 10.3969/j.issn.1673-629X.2022.10.029
10 KIM Y, WANG P, ZHU Y F, et al. A capsule network for traffic speed prediction in complex road networks [C]// 2018 Sensor Data Fusion: Trends, Solutions, Applications (SDF). Bonn: IEEE, 2018: 1–6
11 TIAN C Y, CHAN W K Spatial-temporal attention wavenet: a deep learning framework for traffic prediction considering spatial-temporal dependencies[J]. IET Intelligent Transport Systems, 2021, 1 (4): 549- 561
12 TANG J J, ZENG J Spatiotemporal gated graph attention. network for urban traffic flow prediction based on license plate recognition data[J]. Computer-aided Civil and Infrastructure Engineering, 2022, 37: 3- 23
doi: 10.1111/mice.12688
13 路佳玲, 魏志成, 田多. 基于时空图卷积网络的城市交通流预测模型 [J]. 武汉理工大学学报: 交通科学与版, 2023, 47(2): 234–238.
LU Jia-ling, WEI Zhi-cheng, TIAN Duo. Urban traffic flow prediction model based on spatio-temporal graph convolutional network [J]. Journal of Wuhan University of Technology: Transportation Science and Engineering, 2023, 47(2): 234–238.
14 LI F X, FENG J, YAN H, et al. Dynamic graph convolutional recurrent network for traffic prediction: benchmark and solution [EB/OL]. [2021-05-03]. https://arxiv.org/abs/2104.14917v2.
15 ZHENG C P, FAN X L, WANG C, et al GMAN: a graph multi-attention network for traffic prediction[J]. Association for the Advancement of Artificial Intelligence, 2020, 34 (1): 1234- 1241
16 ZELLNER A, MONTMARQUETTE C A study of some aspects of temporal aggregation problems in econometric analyses[J]. The Review of Economics and Statistics, 1971, 53 (4): 335- 342
doi: 10.2307/1928734
17 ROSSANA R J, SEATER J J Temporal aggregation and economic time series[J]. Journal of Business and Economic Statistics, 1995, 13: 441- 451
18 VLAHOGIANNI E I, KARLAFTIS M G, GOLIAS J C Statistical methods for detecting nonlinearity and nonstationarity in univariate short-term time-series of traffic volume[J]. Transportation Research Part C, 2006, 14 (5): 351- 367
doi: 10.1016/j.trc.2006.09.002
19 BYRON J, SHAWN M T, WILLIAM L E, et al. Intelligent transportation system data archiving: statistical techniques for determining optimal aggregation widths for inductance loop detectors[J]. Transportation Research Record, 2000: 85-93.
20 PARK D, RILETT L R, GAJEWSKI B J, et al Identifying optimal data aggregation interval sizes for link and corridor travel time estimation and forecasting[J]. Transportation, 2009, 36 (1): 77- 95
doi: 10.1007/s11116-008-9180-x
21 SMITH B L, ULMER J M Freeway traffic flow rate measurement: investigation into impact of measurement time interval[J]. Journal of Transportation Engineering, 2003, 129 (3): 223- 229
doi: 10.1061/(ASCE)0733-947X(2003)129:3(223)
22 于雷, 陈旭梅, 耿彦斌, 等 基于小波分解的智能交通系统数据集成方法[J]. 清华大学学报: 自然科学版, 2004, 44 (6): 793- 796
YU Lei, CHEN Xu-mei, GENG Yan-bin, et al Data integration method of intelligent transportation system based on wavenet decomposition[J]. Journal of Tsinghua University: Natural Science Edition, 2004, 44 (6): 793- 796
23 陆振波, 夏井新, 焦恬恬, 等 城市道路交通流监测数据最优汇集时间间隔分析[J]. 东南大学学报: 自然科学版, 2012, 42 (5): 1000- 1005
LU Zhen-bo, XIA Jing-xin, JIAO Tian-tian, et al Analysis of optimal collection time interval of urban road traffic flow monitoring data[J]. Journal of Southeast University: Natural Scienc Edition, 2012, 42 (5): 1000- 1005
24 HOCHREITER S, SCHMIDHUBER J Long short-term memory[J]. Neural Computation, 1997, 9 (8): 1735- 1780
doi: 10.1162/neco.1997.9.8.1735
[1] 龚越, 罗小芹, 王殿海, 杨少辉. 基于梯度提升回归树的城市道路行程时间预测[J]. 浙江大学学报(工学版), 2018, 52(3): 453-460.
[2] 任沙浦, 沈国江. 短时交通流智能混合预测技术[J]. J4, 2010, 44(8): 1473-1478.