Prediction method for road network traffic flow based on hierarchical sequence diagram convolutional network

Abstract

The invention discloses a prediction method for road network traffic flow based on a hierarchical sequence diagram convolutional network. The prediction method comprises the following steps (1) establishing a road network graph, and constructing an adjacency matrix according to the road network graph; (2) taking a flow on each node as a signal of the node; (3) calculating the Laplace matrix by using the adjacency matrix of the graph, for the flow signal, establishing a spectrogram convolution module, and extracting the influence of the spatial structure of the road network on the flow; (4) constructing clustering pooling and anti-pooling modules, extracting hierarchical information in the road network, and looking for flow characteristics of similar structures; (5) constructing a GRU unitby the spectrogram convolution module and the clustering pooling module together, and realizing the extraction of time characteristics of the flow; and (6) completing the prediction of the flow from one sequence to another sequence by using a self-encoder structure. The hierarchical sequence diagram convolutional network provided by the invention has the advantages that the prediction of huge andcomplex traffic flow is more accurate; travel routes for urban residents are convenient to plan; travelling time is reduced, travelling experience is improved; computing efficiency is increased; and the problem of memory overflow is not liable to appear.

Description

technical field [0001] The invention belongs to the technical field of deep learning, and in particular relates to a road network traffic flow prediction method based on a hierarchical time sequence graph convolution network. Background technique [0002] With the continuous development of cities, urban transportation and convenient travel have become more and more topics of concern, and smart city transportation has also become a research hotspot, that is, the intelligent transportation system IIS (Intelligent Transport System). The core issue of this system is how to use historical traffic data, including traffic flow and flow speed, etc., to comprehensively extract urban traffic characteristics, accurately predict short-term or long-term traffic conditions in the future, and then provide timely guidance for individual travel and government decision-making. Reasonable advice. [0003] The prediction of traffic flow can mainly consider the core factors that affect traffic ...

AI Technical Summary

Problems solved by technology

Although this model has been able to better capture the impact of several upstream and downstream roads on traffic flow, the important factor of hierarchical structure is not considered in the factors affecting space.

Images