Journal of Jilin University(Engineering and Technology Edition) ›› 2021, Vol. 51 ›› Issue (5): 1708-1715.doi: 10.13229/j.cnki.jdxbgxb20200535

Previous Articles    

Data imputation approach for lane⁃scale traffic flow based on tensor decomposition theory

Wen-qi LU1(),Tian ZHOU2,Yuan-li GU2,Yi-kang RUI1(),Bin RAN1   

  1. 1.School of Transportation，Southeast University，Nanjing 211189，China
    2.Key Laboratory of Transport Industry of Big Data Application Technologies for Comprehensive Transport，Ministry of Transport，Beijing Jiaotong University，Beijing 100044，China
  • Received:2020-07-16 Online:2021-09-01 Published:2021-09-16
  • Contact: Yi-kang RUI E-mail:lplwq93@126.com;101012189@seu.edu.cn

RICH HTML

 2   

PDF (PC)

147

Abstract:

To reduce the impact of the missing traffic data on traffic flow prediction， advanced driving assistance， traffic state estimation， and other traffic applications， an adaptive rank Tucker decomposition-based imputation approach （ARTDI） is proposed to recover the missing data and improve the data quality. Multi-lane traffic flow data are represented as a tensor to make full use of the spatial-temporal characteristics of traffic flow. The objective function of the imputation approach is determined through Tucker decomposition and solved by the momentum gradient descent method. Three missing types of traffic datasets including the completely random missing， random missing， and mixed missing were constructed based on the lane-scale traffic speed data of the multiple road sections of the expressway in Beijing. The experimental results indicate that the mean absolute percentage errors （MAPE） of the ARTDI approach are 11.67%， 12.03%， and 11.89% respectively under the three missing types. Besides， the results demonstrate that with the increase of the missing rate， the ARTDI approach outperforms the benchmark approaches in terms of recovery accuracy under different missing types and the MAPEs of the ARTDI approach does not increase significantly. Hence， the ARTDI approach is stable and applicable.

Key words: traffic system engineering, data imputation, tensor decomposition theory, Tucker decomposition, lane-scale traffic flow

CLC Number: 

  • U491

Fig.1

Tucker decomposition of a third-order tensor"

Fig.2

Research scope"

Fig.3

Missing types of multi-lane traffic data"

Fig.4

Tensor pattern construction"

Table 1

Parameters summary of compared algorithms"

模型名称模型参数

KNN

SVR

TDI

CP-WOPT

欧氏距离，k=7

径向基函数，gamma=0.1，C=5

R］

R=30

Fig.5

Comparison of recovery results under different thresholds P"

Table 2

Comparison of MAPEs of different data imputation approaches under MCR mode %"

缺失率/%KNNSVRCP-WOPTTDIARTDI
1011.6912.2516.5511.8210.84
2014.0514.2218.0712.6411.22
3016.3116.6118.8113.6311.68
4020.6620.9421.3814.3512.07
5029.0728.6224.7216.5112.52
平均值18.3618.5319.9013.7911.67

Table 3

Comparison of MAPEs of different data imputation approaches under MR mode %"

缺失率/%KNNSVRCP-WOPTTDIARTDI
1012.3311.3116.4214.3711.52
2014.5414.7518.0315.6211.59
3016.1915.9320.3617.6412.02
4020.4021.8222.1019.0512.38
5028.0231.4727.6521.8212.66
平均值18.3019.0520.9117.7012.03

Table 4

Comparison of MAPEs of different data imputation approaches under MCR&MR mode"

缺失率/%KNNSVR

CP-

WOPT

TDIARTDI
1012.2512.4116.6812.8211.19
2014.2214.7717.7215.0411.70
3016.6116.9718.7916.0611.73
4020.9421.8521.0116.9912.26
5028.6232.7724.7119.6812.58
平均值18.5319.7519.7816.1211.89

Fig.6

Recovery results of different lanes"

Table 5

Comparison of MAPE of ARTDI model under different missing modes"

张量模式MRMCRMR&MCR平均值
30×24×319.0317.6718.5018.77
30×168×312.0311.6711.8911.96
30×168×916.9115.9516.7816.85
1 Raza A, Zhong M. Hybrid artificial neural network and locally weighted regression models for lane-based short-term urban traffic flow forecasting[J]. Transportation Planning and Technology, 2018, 41(8): 901-917.
2 Gu Y L, Lu W Q, Qin L Q, et al. Short-term prediction of lane-level traffic speeds: a fusion deep learning model[J]. Transportation Research Part C: Emerging Technologies, 2019, 106: 1-16.
3 Lu W Q, Rui Y K, Yi Z W, et al. A hybrid model for lane-level traffic flow forecasting based on complete ensemble empirical mode decomposition and extreme gradient boosting[J]. IEEE Access, 2020, 8: 42042-42054.
4 Zhong M, Sharma S, Lingras P. Genetically designed models for accurate imputation of missing traffic counts[J]. Transportation Research Record, 2004(1879): 71-79.
5 Williams B M, Hoel L A. Modeling and forecasting vehicular traffic flow as a seasonal ARIMA process: Theoretical basis and empirical results[J]. Journal of Transportation Engineering, 2003, 129(6): 664-672.
6 Bermúdez J D, Corberán-Vallet A, Vercher E. Forecasting time series with missing data using Holt's model[J]. Journal of Statistical Planning & Inference, 2009, 139(8): 2791-2799.
7 王新颖, 隽志才, 吴庆妍, 等. KNN算法的数据优化策略[J]. 吉林大学学报: 信息科学版, 2010, 28(3): 309-313.
Wang Xin-ying, Zhi-cai Juan, Wu Qing-yan, et al. Data optimization strategy of KNN algorithm[J]. Journal of Jilin University(Information Science Edition), 2010, 28(3): 309-313.
8 陈亮, 王金泓, 何涛, 等. 基于SVR的区域交通碳排放预测研究[J]. 交通运输系统工程与信息, 2018, 18(2): 13-19.
Chen Liang, Wang Jing-hong, He Tao, et al. Forecast study of regional transportation carbon emissions based on SVR[J]. Journal of Transportation Systems Engineering and Information Technology, 2018, 18(2): 13-19.
9 Qu L, Zhang Y, Hu J M, et al. A BPCA based missing value imputing method for traffic flow volume data[J]. IEEE Intelligent Vehicles Symposium, 2008: 985-990.
10 Qu L, Li L, Zhang Y, et al. PPCA-based missing data imputation for traffic flow volume: a systematical approach[J]. IEEE Transactions on Intelligent Transportation Systems, 2009, 10(3): 512-522.
11 Acar E, Dunlavy D M, Kolda T G, et al. Scalable tensor factorizations for incomplete data[J]. Chemometrics and Intelligent Laboratory Systems, 2011, 106(1): 41-56.
12 Tan H C, Feng G D, Feng J S, et al. A tensor-based method for missing traffic data completion[J]. Transportation Research Part C: Emerging Technologies, 2013, 28: 15-27.
13 Kolda T G, Bader B W. Tensor decompositions and applications[J]. SIAM review, 2009, 51(3): 455-500.
14 de Lathauwer L, de Moor B, Vandewalle J. A Multilinear singular value decomposition[J]. SIAM Journal on Matrix Analysis and Applications, 2000, 21(4):1253-1278.
15 Chen X Y, He Z C, Wang J W. Spatial-temporal traffic speed patterns discovery and incomplete data recovery via SVD-combined tensor decomposition[J]. Transportation Research Part C: Emerging Technologies, 2018, 86: 59-77.
16 Polyak B T. Some methods of speeding up the convergence of iteration methods[J]. USSR Computational Mathematics and Mathematical Physics, 1964, 4(5): 1-17.
17 Li L C, Du B W, Wang Y G, et al. Estimation of missing values in heterogeneous traffic data: application of multimodal deep learning model[J]. Knowledge-Based Systems, 2020, 194: 105592.
18 Chen X Y, He Z C, Sun L J. A bayesian tensor decomposition approach for spatiotemporal traffic data imputation[J]. Transportation Research Part C: Emerging Technologies, 2018, 98: 73-84.
19 谷远利, 张源, 芮小平, 等. 基于免疫算法优化LSSVM的短时交通流预测[J]. 吉林大学学报: 工学版, 2019, 49(6): 1852-1857.
Gu Yuan-li, Zhang Yuan, Rui Xiao-ping, et al. Short-term traffic flow prediction based on LSSVM optimized by immune algorithm[J]. Journal of Jilin University(Engineering and Technology Edition), 2019, 49(6): 1852-1857.
20 杨飞, 方滨兴, 王春露, 等. 基于小波和回声状态网络的交通流多步预测模型[J]. 吉林大学学报: 工学版, 2013, 43(3): 646-653.
Yang Fei, Fang Bin-xing, Wang Chun-lu, et al. Multi-step traffic flow prediction model based on wavelet and echo state network[J]. Journal of Jilin University(Engineering and Technology Edition), 2013, 43(3): 646-653.
21 龚勃文, 林赐云, 李静, 等. 基于核自组织映射-前馈神经网络的交通流短时预测[J]. 吉林大学学报: 工学版, 2011, 41(4): 938-943.
Gong Bo-wen, Lin Xi-yun, Li Jing, et al. Short-term traffic flow prediction based on KSOM BP neural network[J]. Journal of Jilin University(Engineering and Technology Edition), 2011, 41(4): 938-943.
[1] Hao LI,Hao CHEN. Mixed traffic network equilibrium with battery electric vehicles considering charging queuing time [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(5): 1684-1691.
[2] Ying-ying MA,Si-yuan LU,Xiao-ming ZHANG,Wen-shu WEI. Model of highway travel selection considering individual risk preference difference [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(5): 1673-1683.
[3] Rong-han YAO,Wen-yan QI,Liu-jie ZHENG,Da-yi QU. Settings of guiding markings for left⁃turning vehicles based on lane selection and vehicle trajectory [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(5): 1651-1663.
[4] Li-sheng JIN,Bai-cang GUO,Fang-rong WANG,Jian SHI. Dynamic multiple object detection algorithm for vehicle forward based on improved YOLOv3 [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(4): 1427-1436.
[5] Jin XU,Cun-shu PAN,Jing-hou FU,Jun LIU,Dan-qi WANG. Speed behavior characteristic on typical driving scenarios and along switched scenarios [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(4): 1331-1341.
[6] Kai LU,Wei WU,Guan-rong LIN,Xin TIAN,Jian-min XU. Combination forecasting model for number of assembling passengers at transportation terminal based on KNN regression algorithm [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(4): 1241-1250.
[7] Bo PENG,Yuan-yuan ZHANG,Yu-ting WANG,Ju TANG,Ji-ming XIE. Automatic traffic state recognition from videos based on auto⁃encoder and classifiers [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(3): 886-892.
[8] Guo-zhu CHENG,Rui CHENG,Liang XU,Wen-hui ZHANG. Risk assessment of roadside accidents based on occupant injuries analysis [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(3): 875-885.
[9] Ren HE,Xiao-cong ZHAO,Yi-bin YANG,Jian-qiang WANG. Man⁃machine shared driving model using risk⁃response mechanism of human driver [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(3): 799-809.
[10] Wei-xiong ZHA,Qi-yan CAI,Jian LI,Li-xin YAN. Optimization of offset of urban arterial signal coordination under condition of vehicle entry and exit on side road [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(2): 565-574.
[11] Jian ZHANG,Kun-run WU,Min YANG,Bin RAN. Double⁃ring adaptive control model of intersection during intelligent and connected environment [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(2): 541-548.
[12] Cai-hua ZHU,Xiao-li SUN,Yan LI. Forecast of urban public bicycle traffic demand by station classification [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(2): 531-540.
[13] Dian-hai WANG,Xin-yi SHEN,Xiao-qin LUO,Sheng JIN. Offset optimization with minimum average vehicle delay [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(2): 511-523.
[14] Wen-hui ZHANG,Jia-qi QIN,Hong-tao LI,Hong-xing DENG,Guo-zhu CHENG. Travel choice model between bus transit and subway considering crowding degree [J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(1): 200-205.
[15] Xian-min SONG,Ming-ye ZHANG,Zhen-jian LI,Xin WANG,Ya-nan ZHANG. Setting of dynamic bus lane and its simulation analysis and evaluation [J]. Journal of Jilin University(Engineering and Technology Edition), 2020, 50(5): 1677-1686.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
Copyright © Journal of Jilin University(Engineering and Technology Edition), All Rights Reserved.
Tel: +86-431-85095297　Fax: +86-431-85094074　E-mail: xbgxb@jlu.edu.cn
Powered by Beijing Magtech Co. Ltd