The invention discloses a space-time traffic speed prediction method and system based on attention and multivariate graph convolution fusion. The traffic space-time prediction method is technically characterized by comprising the following steps: acquiring auxiliary data and main data; respectively substituting into expansion space-time diagram convolution to obtain a diagram adjacency matrix; fusing the graph adjacency matrixes to obtain a graph adjacency matrix with fused features; extracting channel features of the main data, and adding the channel features and the graph adjacency matrix of the fused features point by point to obtain the output of the main data; extracting channel characteristics of the auxiliary data, and adding the channel characteristics of the auxiliary data and the graph adjacency matrix point by point to obtain output of the auxiliary data; and repeating the above steps, ending when a cycle reaches a predetermined number of times, and performing jump connection on the graph adjacency matrix with the fused features to obtain traffic speed prediction data. According to the invention, a deep learning space-time model for traffic speed prediction is constructed, and dynamic space features and time features are extracted adaptively.
本发明公开了一种注意力与多元图卷积融合的时空交通速度预测方法及系统。涉及交通时空预测领域,其技术要点在于:获取辅助数据和主要数据;分别代入扩张时空图卷积中得到图邻接矩阵;对图邻接矩阵进行融合得到融合特征的图邻接矩阵;提取主要数据的通道特征,将通道特征和融合特征的图邻接矩阵进行逐点相加,得到主要数据的输出;提取辅助数据的通道特征,将辅助数据的通道特征和图邻接矩阵进行逐点相加,得到辅助数据的输出;重复上述步骤,当循环达到预定次数时截至,将得到融合特征的的图邻接矩阵进行跳跃连接,得到交通速度预测数据。本发明构建了一种用于交通速度预测的深度学习时空模型,自适应地提取动态空间特征和时间特征。
Attention and multivariate graph convolution fusion space-time traffic speed prediction method and system
注意力与多元图卷积融合的时空交通速度预测方法及系统
2023-07-28
Patent
Electronic Resource
Chinese
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen |
European Patent Office | 2023
|Dynamic graph convolution traffic speed prediction method
European Patent Office | 2020
|Traffic flow prediction method based on graph attention convolution network
European Patent Office | 2020
|Traffic speed prediction method based on multi-graph cross attention fusion
European Patent Office | 2023
|European Patent Office | 2024
|