The invention provides a road network traffic state prediction method under a limited observation condition. The method comprises the following steps: determining a data space-time coding method and a data mask matrix; establishing a dynamic graph convolutional neural network module and a multi-head attention module; establishing a generative network framework to complement missing data; and constructing a multi-task learning network model to predict the traffic state. According to the method, the data distribution of the road network traffic state is captured through the generative neural network, and the traffic data is complemented, so that the prediction precision of the road network traffic state under a limited observation condition is improved to a certain extent, and a more scientific and accurate reference can be provided for a traffic system manager; the traffic jam phenomenon can be relieved, the road network traffic operation efficiency is improved, the economic loss is reduced, and the carbon emission is reduced.
本发明提供了一种有限观测条件下的路网交通状态预测方法,包括:确定数据时空编码方法和数据掩码矩阵;建立动态图卷积神经网络模块和多头注意力模块;建立生成式网络框架对缺失数据进行补全;构建多任务学习网络模型预测交通状态。本发明通过生成式神经网络捕捉路网交通状态的数据分布并补全交通数据,在一定程度上提高了有限观测条件下路网交通状态的预测精度,进而可为交通系统管理者提供更为科学准确的参考,有利于缓解交通拥堵现象、提高路网交通运行效率、减少经济损失和降低碳排放量。
Road network traffic state prediction method under limited observation condition
一种有限观测条件下的路网交通状态预测方法
2024-02-20
Patent
Electronic Resource
Chinese
All-road traffic flow estimation method under limited observation condition
European Patent Office | 2024
|European Patent Office | 2021
|Traffic road condition prediction method, device and equipment
European Patent Office | 2023
|Whole road network traffic state prediction method and device
European Patent Office | 2021
|