The invention provides a traffic flow prediction method based on generative spatio-temporal evolution modeling. According to the method, a generative adversarial network is adopted to capture an evolution relationship between traffic nodes along with time, a multi-node traffic sequence is converted into two-dimensional plane data by using a Grubrum angle field, a node embedding combination arrangement generation matrix is constructed to simulate association arrangement of a global traffic network, and a global traffic condition characteristic graph representing a certain specific moment is generated. In the prediction problem solving process, the evolution relation of the traffic characteristics is learned through the generative adversarial network, the two-dimensional image representation of each traffic detection node is generated by using the Grubrum angle field, and the characteristic graphs of all time sequences are obtained. And generating input feature arrangement through node embedding and arrangement generation matrixes, and training a generative adversarial network to generate traffic sequence data of each node in a prediction period. The method fully considers the independence and interaction evolution relationship of the time sequence on each node, and has high prediction accuracy.
本发明提出一种基于生成式时空演化建模的交通流量预测方法。该方法采用生成对抗网络来捕获交通节点之间随时间的演化关系,利用格拉姆角场将多节点交通序列转换为二维平面数据,并构建节点嵌入结合排布生成矩阵模拟全域交通网络的关联排布,生成表征某一特定时刻全域交通状况特征图。在预测问题的求解过程中,通过生成对抗网络对交通特征的演化关系进行学习,使用格拉姆角场生成每个交通侦测节点的二维图像表示,得到所有时间序列的特征图。通过节点嵌入和排布生成矩阵生成输入的特征排布,训练生成对抗网络生成预测时段的每个节点的交通序列数据。该方法充分考虑了每个节点上时间序列的独立以及交互演化关系,具有较高的预测准确性。
Traffic flow prediction method based on generative spatio-temporal evolution modeling
一种基于生成式时空演化建模的交通流量预测方法
2023-06-09
Patent
Electronic Resource
Chinese
Spatio-Temporal AutoEncoder for Traffic Flow Prediction
IEEE | 2023
|Traffic flow prediction method for traffic flow spatio-temporal data information
European Patent Office | 2024
|Short-term traffic flow prediction method based on spatio-temporal correlation
European Patent Office | 2020
|Global spatio‐temporal dynamic capturing network‐based traffic flow prediction
DOAJ | 2023
|Global spatio‐temporal dynamic capturing network‐based traffic flow prediction
Wiley | 2023
|