A Traffic Prediction Method Based on Attention Temporal Graph Convolutional Network

Abstract

The invention belongs to the field of intelligent transportation, and discloses a traffic prediction method based on attention temporal graph convolutional network. First, the urban road network is modeled as a graph structure, and the nodes of the graph represent road sections, and the edges are connections between road sections relationship, describing the time series of each road segment as the attribute characteristics of nodes; secondly, using the attention temporal graph convolutional network model to capture the temporal and spatial characteristics of traffic flow, and using gated recurrent units to learn the characteristics of traffic flow on urban roads. Time change trend to capture time dependence, use the attention mechanism to learn the global time change trend of traffic flow; then, use the fully connected layer to obtain the traffic flow state at different times on each road segment; finally, use different evaluation indicators to estimate the traffic flow on the urban road. The difference between the actual value of the traffic flow and the predicted value is used to evaluate the predictive ability of the model. Experiments prove that the method proposed by the present invention can effectively realize the task of traffic flow prediction on urban roads.

Description

technical field [0001] The invention relates to a traffic prediction method based on attention temporal graph convolution network, belonging to the technical field of intelligent traffic. Background technique [0002] With the deployment and development of intelligent transportation systems, the task of traffic flow prediction has been paid more and more attention. It is a key part of advanced traffic management systems and an important part of traffic planning, traffic management and traffic control. Traffic flow forecasting is the process of analyzing the traffic status on urban roads, including flow, speed and density, etc., mining traffic operation rules and predicting the trend of traffic status changes on the road. It can not only perceive traffic congestion in advance for traffic managers, limit vehicle It can also provide a scientific basis for urban travelers to choose appropriate travel routes and improve travel efficiency. But accurate real-time traffic flow pred...

AI Technical Summary

Problems solved by technology

However, since CNN is essentially applicable to Euclidean space, such as images, grids, etc., it has limitations on traffic networks with complex topological structures, so it cannot essentially describe the spatial dependence of traffic flow. Therefore, this type of method also exists certain defects

Images