The invention relates to an expressway toll station traffic flow prediction device considering adjacent nodes, which comprises the following steps: acquiring traffic flows of a current expressway toll station and toll station nodes strongly associated with the current expressway toll station through a traffic flow monitoring system, generating a data set by using the parameters, and extracting spatial association features between toll stations through a graph convolutional neural network; taking the extraction result of the graph convolutional neural network as the input of a long-short term memory network, and extracting the time correlation characteristics of the toll station traffic flow data; a combined prediction model is established, the traffic flow of a toll station in the future is predicted through traffic flow data collected in real time, and traffic guidance and issuing of scheduling measures are guided.
本发明涉及一种考虑相邻节点的高速公路收费站交通流量预测装置,通过交通流量监控系统获取当前高速收费站以及与之强关联收费站节点的交通流量,将以上参数生成数据集,通过图卷积神经网络提取收费站之间的空间关联特征;将图卷积神经网络的提取结果作为长短时记忆网络的输入,提取收费站交通流数据的时间相关性特征;建立组合预测模型,通过实时采集的交通流数据对未来收费站的交通流量进行预测,并指导交通引导和调度措施的发布。
Expressway toll station traffic flow prediction device considering adjacent nodes
一种考虑相邻节点的高速公路收费站交通流量预测装置
2024-03-29
Patent
Electronic Resource
Chinese
Expressway toll station lane configuration method considering short-time traffic flow
European Patent Office | 2023
|Expressway toll station flow prediction method and system
European Patent Office | 2024
|Expressway section traffic flow backstepping method based on toll station ETC data
European Patent Office | 2023
|