The invention relates to the technical field of traffic flow prediction, in particular to a multi-head attention space-time decoupling dual-channel MAE pre-training traffic flow calculation method, which combines mass data records of expressway traffic flow to construct an intelligent traffic system in the field of traffic prediction, and performs masking self-coding pre-training in a time dimension and a space dimension respectively to obtain a pre-training result. Trend and event dual channels are integrated into a pre-training architecture, traffic space-time prediction is completed, and the problem that the prediction accuracy is limited due to the fact that a complex space-time dynamic dependency relationship of traffic space-time data is difficult to capture is solved.
本发明涉及交通流量预测技术领域,具体为一种多头注意力时空解耦双通道MAE预训练交通流量计算方法,结合高速公路交通流量的大量数据记录,构建交通预测领域智能交通系统,分别在时间维度和空间维度上进行掩蔽自编码预训练,把趋势和事件双通道融入预训练的架构,完成交通时空预测,解决了难以捕获交通时空数据复杂的时空动态依赖关系而导致预测准确率受限的问题。
Multi-head attention space-time decoupling dual-channel MAE pre-training traffic flow calculation method
多头注意力时空解耦双通道MAE预训练交通流量计算方法
2024-10-08
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 |
Space-time heterogeneous decoupling method based on multi-mode traffic flow prediction
European Patent Office | 2024
|European Patent Office | 2024
|European Patent Office | 2025
|Traffic flow prediction method based on multi-head attention mechanism
European Patent Office | 2023
|Distributed traffic flow prediction method and system based on space-time decoupling
European Patent Office | 2023
|