The invention discloses a road network traffic space-time state uncertainty calculation method based on a graph convolution model. The method comprises the following steps: collecting historical traffic space-time data; constructing a road network traffic state point prediction model based on a DCRNN model; establishing a loss function method for describing accidental uncertainty, and performing traffic state uncertainty training on the road network traffic state point prediction model based on MC-Dropout of a Bayesian framework by using historical traffic spatio-temporal data to obtain a trained road network traffic state interval range prediction model; and inputting to-be-predicted traffic spatio-temporal data into the trained road network traffic state interval range prediction model to obtain a traffic state in a predicted time range. According to the method, the uncertainty of the traffic state is quantified more deeply and effectively by combining the deep learning technology and the Bayesian framework, and more flexible and comprehensive decision support is provided for traffic management and scheduling.
本发明公开了一种基于图卷积模型的路网交通时空状态不确定性计算方法,包括:采集历史交通时空数据;基于DCRNN模型构建路网交通状态点预测模型;建立描述偶然不确定性的损失函数方法,并利用历史交通时空数据,基于贝叶斯框架的MC‑Dropout对路网交通状态点预测模型进行交通状态不确定性训练,获得训练好的路网交通状态区间范围预测模型;将待预测交通时空数据输入到训练好的路网交通状态区间范围预测模型中得到预测的时间范围内的交通状态。本发明通过结合深度学习技术和贝叶斯框架,更深入有效地量化交通状态的不确定性,为交通管理和调度提供更灵活和全面的决策支持。
Road network traffic space-time state uncertainty calculation method based on graph convolution model
基于图卷积模型的路网交通时空状态不确定性计算方法
2024-10-29
Patent
Electronic Resource
Chinese
Road network traffic flow prediction method based on gated loop graph convolution attention network
European Patent Office | 2023
|Traffic flow prediction method based on interactive space enhanced graph convolution model
European Patent Office | 2024
|Urban traffic prediction method and device based on space-time mixed graph convolution
European Patent Office | 2023
|Road section traffic flow prediction method based on time convolution neural network
European Patent Office | 2020
|Road traffic safety prediction method based on dynamic graph attention space-time network
European Patent Office | 2024
|