Short-term traffic flow prediction method for complex urban traffic network

Abstract

The invention discloses a short-term traffic flow prediction method for a complex urban traffic network, belonging to the field of intelligent traffic analysis. The method comprises the steps of: obtaining a data source comprising bayonet traffic statistics and vehicle trajectory data extraction; in order to extract the local spatial characteristics of the traffic network traffic, providing a network2image traffic network matrixing algorithm to mine the intersection relationship from the vehicle trajectory to reconstruct the traffic network to obtain the traffic network traffic picture; employing a convolutional neural network to perform local convolution calculation; for reduction of the computing complexity problem, employing a traffic network image algorithm network2image algorithm to mine a relation of intersections in the traffic network from the trajectory data to reconstruct the traffic network to images, and employing the convolutional local calculation of the convolutional neural network to achieve the purposes of extraction of the traffic network local features, reduction of the computing complexity and improvement of the generalization capability; and introducing a small-scale recurrent neural network in a convolutional layer to determine the local traffic network long-time association. The short-term traffic flow prediction method effectively improves the flow short-time prediction accuracy in the actual complex urban traffic network.

Description

technical field [0001] The invention belongs to the field of data mining and intelligent transportation, relates to a local calculation model of a road network, in particular to a short-term traffic flow data prediction method for complex urban traffic networks. Background technique [0002] With the rapid development of traffic informatization, traffic data has grown explosively. Among them, traffic flow per unit time is an important index to describe traffic flow, and its changing law reflects the basic nature of traffic flow. Short-term traffic flow prediction is a research hotspot in the field of intelligent transportation. Accurate short-term traffic flow prediction is one of the key technologies for traffic control and traffic guidance. [0003] So far, there is a large research literature on short-term traffic flow forecasting. From the use of traffic flow data characteristics, it can be summarized into two types: the use of time correlation and the use of time-spac...

AI Technical Summary

Problems solved by technology

But its limited scope is a highway, not a complex urban road network traffic

Images