Urban Traffic Congestion Identification And Early Warning Research Based On Hybrid Deep Learning Model

Since the reform and opening up,the urbanization process has accelerated,the living standards of residents have improved,and the rapid growth of urban motor vehicle possession has brought convenience to people,but it has also brought about increasingly severe traffic congestion problems,which restricts the city to a certain extent.Development,good urban traffic operation and social benefits are inextricably linked,so it is of great significance to study urban traffic congestion.With the rapid development of big data and deep learning,intelligent transportation system(ITS)has become a new direction of future traffic development.This paper aims to identify and warn urban traffic congestion by using deep learning model.Among them,the traffic congestion in the frequent urban areas has obvious time and place distribution law,which has strong predictability;the occasional congestion is the random congestion phenomenon caused by traffic accidents,emergencies or weather,etc.The randomness is harder to predict.This paper comprehensively considers the blocking factors of traffic congestion in frequent and sporadic cities,and realizes the identification and early warning of urban traffic congestion,so that the traffic management department can control the urban traffic congestion.The hybrid deep learning model constructed in this paper mainly consists of two parts.The bottom layer is a typical deep learning model self-encoding network(AE).By using a large amount of data to carry out feature learning training,the potential law of traffic congestion in urban roads can be obtained.After learning the data characteristics,the upper-layer support vector machine(SVM)model is used to classify the urban road traffic congestion and obtain five different levels of traffic operation status.Finally,based on the depth learning model,the congestion time-space distribution map is drawn,and the congestion propagation model based on the rectangular method is established.By focusing on the manifestation of the congestion state,the visualization method is used to present the traffic characteristics of road traffic congestion.At the same time,combined with qualitative analysis and quantitative analysis method based on traffic wave theory,the paper analyzes the propagation and evolution of urban traffic congestion,realizes the identification and early warning of urban traffic congestion,and takes timely measures for traffic management departments to solve urban traffic congestion.Accurate,targeted and targeted solutions and measures provide theoretical basis and decision support.