The invention relates to an urban traffic network traffic flow prediction method and system based on Transform and GCN. Relates to the technical field of traffic prediction, in particular to the technical field of urban traffic network traffic flow prediction. In order to overcome the defects of existing research and invention, the invention provides the following scheme. The method comprises the following steps of: performing Z-score normalization processing on data collected by a traffic system, collecting and coding time-space attributes of a traffic network, pre-defining a topological structure according to a road network and traffic data, establishing a neural network model based on Transform and GCN, training the proposed model by utilizing historical data, and performing data analysis by adopting an autoregressive reasoning mode. And predicting future traffic load data according to the historical traffic data. The method is well applied to the traffic field, the accuracy and reliability of traffic flow prediction can be improved, and powerful support is provided for traffic management and planning.
一种基于Transformer与GCN的城市交通网络交通流预测方法与系统。涉及交通预测技术领域,具体涉及城市交通网络交通流预测技术领域。为解决现有研究发明的不足,本发明提供如下方案,将交通系统收集的数据进行Z‑score归一化处理、对交通网络时空属性进行收集与编码、根据道路网络与交通数据进行预定义拓扑结构结构、建立一个基于Transformer与GCN的神经网络模型、利用历史数据对提出的模型进行训练、采用自回归的推理模式,根据历史交通数据预测未来交通荷载数据。本发明在交通领域中有较好的应用,可以提高交通流量预测的准确性和可靠性,为交通管理和规划提供有力支持。
Transform and GCN-based urban traffic network traffic flow prediction method and system
一种基于Transformer与GCN的城市交通网络交通流预测方法与系统
2024-08-30
Patent
Electronic Resource
Chinese
European Patent Office | 2023
|Urban traffic road network traffic flow prediction method considering carbon emission model
European Patent Office | 2024
|Traffic flow prediction method and system for urban road network
European Patent Office | 2021
|