A spatio-temporal prediction method of road network traffic flow for intelligent transportation and intelligent driving

Abstract

A road network traffic flow spatio-temporal prediction method for intelligent traffic and intelligent driving, which mainly includes: an acquisition step, an encoding step, a decoding step and a prediction step, so as to obtain the traffic flow of the road network within the prediction time period according to the prediction map signal . On the one hand, due to the construction of the inverted tooth-shaped diffusion convolution cycle module in the encoding step and the decoding step, a state variable is retained in the convolution layer, which helps to reduce the number of state variables to be maintained. Through the inverted tooth-shaped cycle Towards the convolution diffusion process that can optimize the state variables, and the restructured diffusion convolution cycle processing process can realize the operation of updating the state twice at the same time, which can enhance the short-term dependence of time; on the other hand, due to the compilation step In the paper, the daily periodogram signal and the weekly periodogram signal are taken into consideration in the decoding process, so that the basis information is no longer single, and the noise that may be introduced in the progressive learning process can be effectively suppressed, thereby improving the prediction accuracy of the prediction map signal.

Description

technical field [0001] The invention relates to the fields of intelligent transportation, intelligent driving and artificial intelligence, in particular to a spatio-temporal prediction method of road network traffic flow oriented to intelligent transportation and intelligent driving. Background technique [0002] With the sharp increase in car ownership, the limited road network capacity can no longer meet the blowout traffic demand, and traffic congestion has become a common phenomenon in modern transportation. Traffic congestion not only wastes people's precious daily time, but also causes a series of chain problems, such as frequent traffic accidents, deterioration of social environment, waste of energy resources and so on. To this end, on the one hand, it is necessary to develop an intelligent transportation system (Intelligent Transportation System, ITS), to scientifically dispatch traffic from a global perspective, to reasonably disperse traffic pressure, and to maximi...

AI Technical Summary

Problems solved by technology

Among them, the expressway network is relatively small in scale compared with the urban dense road network, and it is easy to obtain public data resources. Therefore, the research on the short-term traffic flow and multi-time prediction of the expressway network has been relatively mature. Some scholars have proposed the road network The spatio-temporal prediction model can already capture the dynamic characteristics of time and space, but there are still some problems. This kind of spatio-temporal prediction model only considers the spatial local correlation and the short-term dependence of time, and does not consider the long-term dependence of time and periodic characteristics, making the highway network There are some inaccurate situations in the traffic flow prediction results, and the prediction performance needs to be further improved

Images