Accurate short-term traffic flow prediction plays an important role in traffic guidance and traffic safety. In order to solve the problem that traffic flow is difficult to predict because of the sharp nonlinearity and randomness, a deep learning framework, i.e., the temporal convolutional network (TCN) was explored to capture the nonlinear spatiotemporal characteristics of traffic flow. By adjusting the hyper-parameters of TCN, a traffic flow predictor is proposed. Besides, this paper adopts a method to calculate the designated cross-section traffic volume of freeways from toll data, which makes up for the inability to obtain traffic volume at some cross-section locations due to damage or lack of traffic detection equipment. Through data merging, data cleaning, data reconstruction, data filtering, a traffic flow data set which was selected as the data input with a time interval of 15 min was constructed based on the toll data of Shaanxi Province freeway network from December 2018 to April 2019 as the data source. The TCN model was compared with the SVM, SAE, LSTM and GRU models in terms of mean absolute error (MAE) and root mean square error (RMSE). The prediction results of multiple cross-sections showed the TCN model performs best with superior prediction accuracy, which indicate the TCN model has good robustness and generalization ability. The traffic volume calculation algorithm may provide a practical method for deriving the traffic volume without installing any additional regularly maintained detectors and equipment on the freeway. And the prediction results of TCN model can provide strong support for traffic control and traffic induction.
Short-Term Traffic Flow Prediction on a Freeway with Multiple Spatial Toll Data Via Temporal Convolutional Network
Lect. Notes Electrical Eng.
International Conference on Green Intelligent Transportation System and Safety ; 2021 November 19, 2021 - November 21, 2021
28.10.2022
19 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Short-Term Traffic Prediction considering Spatial-Temporal Characteristics of Freeway Flow
DOAJ | 2021
|Short-Term Freeway Traffic Flow Prediction: Bayesian Combined Neural Network Approach
Online Contents | 2006
|Short-Term Freeway Traffic Flow Prediction: Bayesian Combined Neural Network Approach
British Library Online Contents | 2006
|