The invention discloses a road network traffic data restoration method based on DTW-RGCN. The method comprises the following steps: 1) reconstructing a road traffic network based on road historical data similarity; 2) preprocessing the road traffic data flow and constructing a traffic flow state matrix data set; 3) based on the road traffic network and the road traffic flow state matrix, using a graph convolution network based on Gaussian distribution to extract features: taking the road traffic network and the road traffic state matrix as inputs of a graph convolution layer based on Gaussiandistribution, and extracting node features of the road traffic network; and 4) realizing traffic flow data restoration based on a Gaussian distribution graph convolution network: taking the traffic road network data containing data missing after DTW processing as the input of RGCN, defining a model loss function, and continuously optimizing the parameters of the model through back propagation to realize restoration of the traffic road network missing data. According to the invention, the robustness and accuracy of traffic flow data restoration are effectively improved.
一种基于DTW‑RGCN的路网交通数据修复的方法,包括以下步骤:1)基于道路历史数据相似性重构道路交通网络;2)对道路交通数据流进行预处理并构建交通流状态矩阵数据集;3)基于道路交通网络和道路交通流状态矩阵,使用基于高斯分布的图卷积网络提取特征:将道路交通网络和道路交通状态矩阵作为基于高斯分布的图卷积层的输入,提取道路交通网络的节点特征;4)基于高斯分布的图卷积网络实现交通流数据修复:将DTW处理后的含有数据缺失的交通路网数据作为RGCN的输入,定义模型损失函数,经过反向传播不断优化模型的参数,实现对交通路网缺失数据的修复。本发明有效提高了交通流数据修复的鲁棒性与准确性。
Road network traffic data restoration method based on DTW-RGCN
一种基于DTW-RGCN的路网交通数据修复的方法
2020-10-16
Patent
Electronic Resource
Chinese
Trajectory restoration system based on urban traffic road network
European Patent Office | 2023
|Road network traffic time sequence characteristic data quality diagnosis and restoration method
European Patent Office | 2020
|GraphSAGE-GAN-based traffic network data restoration method
European Patent Office | 2021
|Intelligent network connection traffic data restoration method
European Patent Office | 2023
|