Abstract
The implementation of the toll free during holidays makes a large number of traffic jams on the expressway. Real-time and accurate holiday traffic flow forecasts can assist the traffic management department to guide the diversion and reduce the expressway’s congestion. However, most of the current prediction methods focus on predicting traffic flow on ordinary working days or weekends. There are fewer studies for festivals and holidays traffic flow prediction, it is challenging to predict holiday traffic flow accurately because of its sudden and irregular characteristics. Therefore, we put forward a data-driven expressway traffic flow prediction model based on holidays. Firstly, Electronic Toll Collection (ETC) gantry data and toll data are preprocessed to realize data integrity and accuracy. Secondly, after Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) processing, the preprocessed traffic flow is sorted into trend terms and random terms, and the spatial-temporal correlation and heterogeneity of each component are captured simultaneously using the Spatial-Temporal Synchronous Graph Convolutional Networks (STSGCN) model. Finally, the fluctuating traffic flow of holidays is predicted using Fluctuation Coefficient Method (FCM). Through experiments of real ETC gantry data and toll data in Fujian Province, this method is superior to all baseline methods and has achieved good results. It can provide reference for future public travel choices and further road network operation.
Funder
Fujian Expressway Science & Technology Innovation Research Institute Co., Ltd. Open project
The Local Science and Technology Development Project guided by the central government
Provincial Candidate Project of Fujian Province Million Talents Project
Patent Grant Project
Crosswise project
Research Platform Open Project
Publisher
Public Library of Science (PLoS)
Reference44 articles.
1. Chen H, Zou F, Guo F, Gu Q, Yu X, Luo Y, et al. A ETC gantry information calibration method based on trajectory data of special transportation vehicles. In: International Conference on Frontiers of Electronics, Information and Computation Technologies; 2021. p. 1–7.
2. A novel on-demand vehicular sensing framework for traffic condition monitoring;SA Rahman;Vehicular Communications,2018
3. An infrastructure-assisted crowdsensing approach for on-demand traffic condition estimation;SA Rahman;IEEE Access,2019
4. The implication of road toll discount for mode choice: Intercity travel during the chinese spring festival holiday;X Lin;Sustainability,2018
5. Huang L. Construction scheme of intelligent comprehensive application platform for expressway toll collection big data; 2022; 8: 17. Available from: https://mp.weixin.qq.com/s/rgA5bitqJfBqDacL48KJlw.