The invention discloses a traffic flow prediction method based on a self-adaptive dynamic fusion graph convolutional network, and the method comprises the steps: enabling a self-adaptive dynamic graph generator to capture a dynamic time graph and a dynamic space graph of a traffic network under the condition of given traffic sequence input, and respectively capturing the dynamic time and space trends; then, capturing a dynamic space-time fusion image of the traffic network through a space-time fusion image module, inputting the acquired traffic sequence and the dynamic space-time fusion image into a space-time block, enhancing a nonlinear dependency relationship and a dynamic trend, obtaining space-time characteristics of the traffic sequence, and inputting the acquired traffic sequence into an internal semantic enhancement module; and finally, integrating the spatial-temporal characteristics of the traffic sequence, and inputting the integrated spatial-temporal characteristics and the traffic sequence subjected to semantic enhancement into a multi-layer perceptron to obtain a prediction result. The model provided by the invention is better in effect and can predict the traffic flow more accurately.
本发明公开了一种基于自适应动态融合图卷积网络的交通流量预测方法,在给定交通序列输入的情况下,自适应动态图生成器捕获交通网络的动态时间图和动态空间图,分别捕获动态的时间和空间趋势。然后,通过时空融合图模块捕获交通网络的动态时空融合图,将采集的交通序列和动态时空融合图输入到时空块,增强非线性依赖关系和动态趋势,得到交通序列的时空特征,并将采集的交通序列输入内在语义增强模块,得到语义增强后的交通序列,最后对交通序列的时空特征进行整合,然后与语义增强后的交通序列输入多层感知机,得到预测结果。本发明所提出的模型效果更好且能够更准确地预测交通流量。
Traffic flow prediction method based on adaptive dynamic fusion graph convolutional network
一种基于自适应动态融合图卷积网络的交通流量预测方法
2024-06-28
Patent
Electronic Resource
Chinese
Multibranch Adaptive Fusion Graph Convolutional Network for Traffic Flow Prediction
DOAJ | 2023
|Space-time adaptive dynamic graph convolutional network traffic flow prediction method
European Patent Office | 2024
|European Patent Office | 2024
|European Patent Office | 2025
|European Patent Office | 2023
|