Highway traffic flow prediction method based on convolutional neural network

Abstract

The invention relates to a highway traffic flow prediction method based on a convolutional neural network. The highway traffic flow prediction method is characterized by comprising the following steps: 1) pre-processing traffic flow data and external factor data, constructing a time adjacency matrix, a daily periodicity matrix, a weekly periodicity matrix and an external factor vector as input ofa model; 2) using a convolutional neural network to model the time adjacency, daily periodicity and weekly periodicity of traffic flow, and extracting different temporal and spatial characteristics; 3) fusing external factors based on the temporal and spatial characteristics extracted in step 2), to form a fusion vector as the input of a fully connected layer; and 4) using the fully connected layer to fuse all the characteristics to complete the final traffic flow prediction. According to the highway traffic flow prediction method, a deep learning model is utilized to model the traffic flow ofa high-speed road network, the spatio-temporal characteristics of traffic flow and external influence factors are combined, so that the model can cope with complex traffic flow characteristics, prediction results are obtained, and simultaneous prediction of all road sections in an area is achieved.

Description

technical field [0001] The invention relates to a method for predicting expressway traffic flow based on a convolutional neural network, which belongs to the field of artificial intelligence-intelligent transportation. Background technique [0002] Traffic flow prediction is an integral part of Intelligent Transportation System (ITS). As the core subsystem of ITS, traffic control and guidance systems need to rely on accurate traffic flow prediction. Traffic flow prediction is also the basis for determining the scale of road projects, road surface maintenance, technical standards and economic evaluation. [0003] For effective traffic flow prediction, some challenging capabilities are required, such as being able to form more intelligent and efficient predictions to simulate the spatial and temporal characteristics of traffic flow, being able to deal with complex external feature privacy, and being able to process large-scale traffic data Wait. The problem is how to constru...

AI Technical Summary

Problems solved by technology

However, most of the existing methods only predict one road section, and it is difficult to effectively use the spatial characteristics of the traffic network, which makes the expressway traffic management department lack the overall grasp of the traffic status change trend of the road network, which is not conducive to the macro regulation of the road network

Images