Research on short-term traffic flow forecast based on attention-T-GRU

Abstract

Abstract: Since some key road sections need accurate prediction results, the prediction modeling of a single road was carried out, and the adjacent roads in the same direction of the predicted road are selected for research considering the temporal and spatial correlation of traffic flow. First, a velocity matrix was constructed according to the correlation between the research road and its upstream and downstream roads. Secondly, the velocity matrix was input into the attention mechanism network to extract the spatial connection among the roads. Finally, the data output of the attention mechanism was decomposed into several sequences T, which was input into the GRU network to extract features, forming the short-term traffic flow prediction of the ATGRU (Attention-T-GRU) combination model. The proposed ATGRU combined model is verified by using the traffic flow data in Xi’an. The results show that the ATGRU model has a better prediction accuracy compared to models such as T-LSTM, CNN-LSTM and ACGRU.

Key words: short-term traffic flow forecast, spatio-temporal feature, attention mechanism, combination model

CLC Number:

U491.14

ZHANG Xi-Jun, SU Jin, CHEN Xuan, SHANG Ji-yang, CUI Yong. Research on short-term traffic flow forecast based on attention-T-GRU[J]. Journal of Lanzhou University of Technology, 2025, 51(4): 107-113.

References

[1] STEPHANEDES Y J,MICHALOPOULOS P G,PLUM R A.Improved estimation of traffic flow for real-time control[J].Transportation Research Record,1981(795):28-38.
[2] MIAO F,TAO L,XUE J,et al.A queue hybrid neural network with weather weighted factor for trafficflow prediction[C]//2021 IEEE 24th International Conference on Computer Supported Cooperative Work in Design (CSCWD).Dalian:IEEE,2021:788-793.
[3] 陈刚,郭晓梅.基于时间序列模型的非平衡数据的过采样算法[J].信息与控制,2021,50(5):522-530.
[4] BOX G E P,PIERCE D A.Distribution of residual autocorrelations in autoregressive-integrated moving average time series models[J].Journal of the American Statistical Association,1970,65(332):1509-1526.
[5] NIHAN N L,HOLMESLAND K O.Use of the box and Jenkins time series technique in traffic forecasting[J].Transportation,1980,9(2):125-143.
[6] WILLIAMS B M,DURVASULA P K,BROWN D E.Urban freeway traffic flow prediction:application of seasonal autoregressive integrated moving average and exponential smoothing models[J].Transportation Research Record,1998,1644(1):132-141.
[7] 董春娇,邵春福,诸葛承祥,等.基于时空特性的交通自由流短时预测状态空间模型[J].土木工程学报,2013,46(8):111-118.
[8] 程颖,陈艳艳,刘莹,等.城市交通排放高分辨率分析方法研究——北京实证[J].交通运输系统工程与信息,2018,18(2):236-244.
[9] 王晓原,吴磊,张开旺,等.非参数小波算法的交通流预测方法[J].系统工程,2005,23(10):44-47.
[10] 马林才,许玮珑,刘大学.基于小波变换的城市市区典型路段交通流量预测[J].控制与决策,2011,26(5):789-793.
[11] 王晨辉.基于组合模型的短时交通流预测方法研究[D].兰州:兰州理工大学,2021.
[12] MOU L,ZHAO P,XIE H,et al.T-LSTM:a long short-term memory neuralNetwork enhanced by temporal information for traffic flow prediction[J].IEEE Access,2019,7:98053-98060.
[13] 刘明宇,吴建平,王钰博,等.基于深度学习的交通流量预测[J].系统仿真学报,2018,30(11):4100-4105.
[14] 张玺君,王晨辉.基于SARIMA-GA-Elman组合模型的短时交通流预测方法[J].兰州理工大学学报,2022,48(5):107-113.
[15] 桂智明,李壮壮,郭黎敏.基于ACGRU模型的短时交通流预测[J].计算机工程与应用,2020,56(21):260-265.
[16] VASWANI A,SHAZEER N,PARMAR N,et al.Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems.Long Beach:[s.n.],2017:6000-6010.