Road traffic flow prediction method for scats system based on spatial graph convolutional neural network

Abstract

A method for predicting road traffic flow in SCATS system based on spatial graph convolutional neural network. First, the lanes are used as nodes, and the nodes with interconnection are obtained according to the traffic adjacency matrix, and the correlation of the connected nodes is sorted from high to low, and then Find the neighborhood nodes of each target node, construct a high-order neighborhood traffic state matrix from the traffic flow state data of the target node and its neighborhood nodes as the input of CNN, and finally get the traffic flow state prediction result of the predicted lane. The invention fully exploits the correlation between time and space of the traffic flow state, which not only improves the accuracy but also enhances the anti-interference of random results.

Description

technical field [0001] The invention belongs to the field of intelligent traffic forecasting, and relates to a road traffic flow forecasting method of a SCATS system. Background technique [0002] Nowadays, the improvement of living standards and the progress of automobile manufacturing technology have led to a significant increase in the number of cars owned. When cars bring convenience to our lives, they have also brought many problems. Among them, the problem of urban road congestion is closely related to life. To alleviate the congestion problem, the first thing to do is to allocate urban road resources rationally, and road traffic flow state prediction can help the rational allocation of urban space to a large extent. [0003] Currently widely used forecasting models include: historical average model, time series model, autoregressive moving average model, non-parametric regression model, etc. However, these models are not sufficiently correlated in time and space to f...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the deficiency that the existing technology cannot accurately predict the traffic flow state, the present invention provides a SCATS system road traffic flow prediction method based on the spatial map convolutional neural network, which can combine the predicted road and its neighborhood roads to predict the future time The traffic flow state, the prediction result has a strong ability to resist random interference

Images