A spatial-temporal prediction scheme of urban traffic flow based on graph convolutional neural network

Abstract

The present invention relates to a spatial-temporal prediction scheme of urban traffic flow based on graph convolutional neural network, comprising the following steps: obtaining urban road network topological structure diagram and historical flow data; constructing and training ST-GCN network model; predicting and obtaining all Class A intersections use the Adjacent algorithm to obtain the predicted flow of all B-type intersections; use the Adjacent algorithm to obtain the predicted flow of all C-type intersections; then use the Similar algorithm to obtain the predicted flow of D-type intersections; finally output the predicted flow of all intersections. Compared with the prior art, the present invention first obtains the predicted traffic of intersections with historical traffic data by constructing and training the ST-GCN network model, and then through the Adjacent-Similar algorithm, for intersections without historical traffic data, according to its adjacent intersections For the "island" intersection, find the similar intersection and use the predicted flow of the similar intersection as the predicted flow of the intersection, which provides a new solution for the flow prediction of the intersection without historical flow data.

Description

technical field [0001] The invention relates to the field of intelligent transportation, in particular to a spatial-temporal prediction scheme of urban traffic flow based on a graph convolutional neural network. Background technique [0002] Traffic forecasting is a key part of intelligent transportation system and an important means to realize traffic planning, traffic management and traffic control. Accurate traffic flow forecast information can provide traffic managers with timely traffic decision-making basis, better manage urban traffic, improve road operation efficiency, and also allow drivers to understand traffic conditions in advance, change routes in time, and choose reasonable traffic roads , save travel time, thereby reducing traffic congestion. In addition, it can also reduce environmental pollution to a certain extent and improve traffic safety. Therefore, urban flow forecasting is an important part of urban traffic system. [0003] However, there are still ...

AI Technical Summary

Problems solved by technology

However, in actual use, due to the limited geographical location and equipment of some intersections, it is often impossible to obtain historical traffic data, so it is impossible to predict traffic flow data based on historical traffic.

Images