The invention provides a traffic flow prediction method based on an interactive space enhanced graph convolution model, and the method comprises the following steps: S1, collecting original traffic flow data X of different time granularities every day, every week and recent, and enabling the original traffic flow data X to pass through a parallel and same binary tree structure; each binary tree structure is composed of a plurality of IS blocks, and generating new traffic flow data and a dynamic fusion image; and S2, enabling the new traffic flow data and the dynamic fusion graph obtained in the step S1 to pass through respective diffusion graph convolution modules, fusing outputs of different time granularities generated by all the diffusion graph convolution modules by using a multilayer perceptron, and obtaining a traffic flow prediction result. According to the method, a dynamic time warping graph, a road connection graph and original traffic flow data are fused into a dynamic fusion graph, dynamic global spatial correlation is extracted through graph attention and a gating mechanism, multi-scale temporal correlation is extracted through a plurality of IS blocks, combination of the dynamic spatial correlation and the multi-scale temporal correlation is achieved, and the dynamic spatial correlation and the multi-scale temporal correlation are combined. And the traffic flow prediction precision is improved.
本发明提供的一种基于交互空间增强图卷积模型的交通流预测方法,包括如下步骤:S1,分别采集每天、每周和最近的不同时间粒度的原始交通流数据X,分别通过并行且相同的二叉树结构。每个二叉树结构都由多个IS块构成,生成新的交通流数据和动态融合图;S2,将步骤S1获得的新的交通流数据和动态融合图经过各自的扩散图卷积模块,使用多层感知机融合所有扩散图卷积模块生成的不同时间粒度的输出,获得交通流预测结果。该方法将动态时间规整图、道路连接图和原始交通流数据融合为动态融合图,通过图注意力和门控机制提取动态全局空间相关性,再通过多个IS块提取多尺度时间相关性,实现动态空间相关性和多尺度时间相关性的结合,提高了交通流预测的精度。
Traffic flow prediction method based on interactive space enhanced graph convolution model
一种基于交互空间增强图卷积模型的交通流预测方法
2024-07-05
Patent
Electronic Resource
Chinese
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06Q Datenverarbeitungssysteme oder -verfahren, besonders angepasst an verwaltungstechnische, geschäftliche, finanzielle oder betriebswirtschaftliche Zwecke, sowie an geschäftsbezogene Überwachungs- oder Voraussagezwecke , DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES |
European Patent Office | 2023
|Traffic flow prediction method based on double graph convolution
European Patent Office | 2024
|Traffic flow prediction method based on graph attention convolution network
European Patent Office | 2020
|Double-end graph convolution traffic flow prediction method with graph learning
European Patent Office | 2023
|Traffic prediction method based on dynamic graph convolution
European Patent Office | 2023
|