Urban road network short-term traffic flow prediction method based on space-time residual hybrid model

Abstract

The invention discloses an urban road network short-term traffic flow prediction method based on a space-time residual hybrid model, and the method comprises the steps: collecting the current latitude and longitude data of a vehicle, storing the data in a big data cluster database, and carrying out the data preprocessing of the original data; dividing the urban road network into a traffic grid network according to longitude and latitude, mapping the longitude and latitude data of the vehicle into the traffic grid network, and generating traffic grid data; performing standardization processing on the traffic raster data, and constructing a training set and a test set; constructing an urban road network short-term traffic flow prediction model based on the space-time residual hybrid model; and training the constructed urban road network short-time traffic flow prediction model based on the space-time residual hybrid model to predict traffic raster data at the next moment. According to the method, the residual hybrid model is introduced under space-time analysis of the road network, and the capability of capturing the traffic flow time and space of the road network is improved, so that the relative error when the space-time residual hybrid model predicts the traffic flow is reduced, and the prediction precision is improved.

Description

technical field [0001] The invention belongs to the field of deep learning and intelligent traffic flow forecasting, in particular to a short-term traffic flow forecasting method for urban road networks based on a spatiotemporal residual error mixed model. Background technique [0002] With the popularity of traffic sensors and the deployment of new sensors, traffic data has shown an explosive growth, and traffic flow prediction largely relies on historical and real-time traffic flow data collected from various sensors, including GPS, cameras, coils, and more. The purpose of traffic flow prediction is to provide traffic flow information. Accurate traffic flow prediction is essential to improve the reliability and safety of intelligent transportation systems, and can provide accurate and reliable traffic information for traffic travelers and traffic management departments. [0003] Traffic flow forecasting methods can be roughly divided into two categories: model-driven and d...

AI Technical Summary

Problems solved by technology

Model-driven methods are generally pre-determined based on theoretical assumptions, which are highly theoretical but low in practicability, and the actual situation is generally more complicated than the theoretical assumptions; data-driven methods mainly benefit from the growing traffic data, and traditional data-driven methods Lack of ability to process high-dimensional data. With the rise of deep learning theory, data-driven methods combined with deep learning models can effectively model high-dimensional data and capture feature information in the data.

In order to capture the spatio-temporal information of the urban road network, a spatio-temporal residual model has been proposed. This method uses 2D convolution in space for spatial feature mining, and uses residual network for capture in time. However, it lacks the screening of time series and is easy to capture. Time series with low correlation, and when the residual method captures small changes, it is easy to ignore long-term time features, thereby reducing the prediction accuracy of the model

Images