The invention discloses an expressway traffic flow prediction method. The method comprises the steps of collecting traffic information data of an expressway section within a preset time length; converting the acquired traffic information data into corresponding road section traffic volume data; distributing the road section traffic volume data according to time and space to obtain a road section traffic volume matrix; extracting the space-time characteristics of the traffic flow through a road section traffic volume matrix and an LSTM method, and establishing a model; capturing the key features of the space-time traffic volume obtained in the previous step by utilizing an attention model, and predicting the future traffic volume by adopting a single-layer full-connection neural network. The method constructs the depth of the network based on an attention mechanism, achieves the adaptive attention of the features of the key part in the traffic data, extracts the features, achieves the prediction of the traffic flow of a future expressway, and can achieve the more precise prediction of the traffic flow of the future expressway compared with a conventional LSTM and other prediction methods.
本发明公开了一种高速公路交通流量预测方法,该方法包括采集预设时长内高速公路路段的交通信息数据;将采集的交通信息数据转换为相对应的路段交通量数据;将路段交通量数据按时间和空间分布得到路段交通量矩阵;通过路段交通量矩阵以及LSTM方法对交通流时空特征进行提取并建立模型;利用注意力模型捕获前一步得到的时空交通量关键特征,采用单层全连接神经网络对未来交通量进行预测。本发明基于注意力机制构造网络的深度,实现交通数据中关键部分的特征的自适应关注,并将其特征提取出来,从而对未来高速公路交通流量预测,该模型与现有的LSTM和其他预测方法相比,可以更精确的预测未来高速公路交通流量。
Highway traffic flow prediction method
一种高速公路交通流量预测方法
2020-06-12
Patent
Elektronische Ressource
Chinesisch
Highway traffic flow prediction method based on weather conditions
Europäisches Patentamt | 2023
|Highway traffic flow speed prediction method based on traffic factor state network
Europäisches Patentamt | 2020
|Highway short-time traffic flow prediction method based on deep learning
Europäisches Patentamt | 2023
|