Research On Short-term Prediction Of Urban Road Traffic Flow

With economic growth and improvement of people's livelihood,car ownership has been keeping a great growth momentum,which brings about the prominent contradiction between traffic demand and supply of road resources.Therefore,relieving traffic congestion appears to be particularly urgent.Thus accurately grasping the inherent laws of traffic flow and making scientific forecasting of traffic flow at certain times in the future will be of great significance to traffic guidance,traffic management,and traffic planning.In this paper,the short-term traffic flow forecasting of the road section and the intersection is taken as the study object.Secondly,the BP Neural Network model and the grey BP Neural Network model for short-term traffic flow forecasting are constructed,including the selection of network structure,related parameter settings,and network structure training.Finally,based on the fault identification and restoration of the collected data,the short-term traffic flow of the section and the intersection is forecast.The results show that: Whether it's a section or intersection,the prediction accuracy of the grey BP Neural Network is better than that of the BP Neural Network in the study of short-term traffic flow forecasting.Meanwhile,as for section,the grey BP Neural Network established by the traffic data of the same period in the previous few days is more effective for forecasting.And for subsection,the gray BP Neural Network established on traffic data of the previous time intervals of the same day possesses better forecasting.This is mainly due to the fact that the traffic at the intersection is more complex and periodicity of the daily traffic flow of the intersection is less noticeable than that of the section whose stability of the traffic flow at the same time on different days is more stable.As an essential component of the intelligent transportation system,traffic information flow is also an influencing factor of traffic guidance and control.