Collaborative prediction method and system for entry and exit traffic flow of highway toll station

By applying a traffic flow collaborative prediction method based on the spatiotemporal attention interactive network on highway toll stations, the problem of low prediction accuracy of traditional methods in abnormal scenarios such as holidays and other abnormal scenarios is solved, and more accurate long-term traffic flow prediction is achieved.

CN119723889APending Publication Date: 2025-03-28SHANDONG TRAFFIC PLANNING DESIGN INST +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202510030096.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-01-08
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

Traditional traffic flow prediction methods cannot effectively cope with the uncertainty and suddenness of highway toll stations in abnormal scenarios such as holidays, resulting in low prediction accuracy. Traditional deep learning models only consider short-term predictions and cannot accurately predict traffic flow information for a long period of time.

Method used

The traffic flow collaborative prediction method based on the spatiotemporal attention interaction network is adopted to construct a time series augmentation matrix by obtaining holiday historical data, and a pre-trained model is used to predict the traffic flow in and out of the station, and the characteristic information of the traffic flow in and out of the station is shared to improve the prediction accuracy.

Benefits of technology

The traffic flow prediction accuracy of expressway toll stations in abnormal scenarios such as holidays, can more accurately predict traffic flow information for a long period of the future, and adapt to the complex space-time characteristics of expressway toll stations.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides an in-out station traffic flow collaborative prediction method and system for an expressway toll station, and belongs to the technical field of traffic flow prediction, and the method comprises the steps: obtaining the historical in-out station traffic flow data and the in-out station traffic flow data before festivals and holidays of the expressway toll station in a to-be-detected region; on the basis of the obtained historical inbound and outbound traffic flow data of the holidays and festivals and the obtained pre-holiday inbound and outbound traffic flow data, time sequence augmentation matrixes of inbound and outbound traffic flows of the holidays and festivals are constructed respectively; based on the obtained time sequence augmentation matrix of the incoming and outgoing traffic flows, obtaining an incoming and outgoing traffic flow collaborative prediction result through a pre-trained traffic flow collaborative prediction model based on a space-time attention interaction network; the problems that a traditional prediction method is low in prediction precision due to the fact that a traditional method cannot adapt to uncertainty and burstiness of traffic flow in an abnormal scene, and a traditional prediction model based on deep learning only considers short-time prediction, and the total prediction time span is small are effectively solved.
Need to check novelty before this filing date? Find Prior Art