Research On Traffic State Prediction Based On Deep Learning Algorithm Of SNN Model

With the increasing demand for traffic,the problem of road traffic congestion is more and more serious.Traffic state prediction is one of the important bases for traffic management and control accurately.Based on this,a traffic state prediction method based on deep learning algorithm of SNN model is researched and a urban road network traffic state prediction model considering large vehicles factors is established.The main research contents and innovations of this paper are as follows:(1)Traffic state analysis and traffic data preprocessing.The velocity is used as the standard of traffic congestion state identification,and the K-Means clustering algorithm is used as the traffic state identificaiton method.According to the traffic data characteristics,the finite impulse response filter is designed.(2)Algorithm design of traffic flow velocity prediction.The deep gated recurrent unit neural network is established.The information transmission mechanism based on integrate-fire neuron in spiking neural network is considered into the hidden layer of the gated recurrent unit neural network,and the information transmission mechanism is re-established between the neurons.The prediction effect is evaluated by using the open data which is from the United States Federal Highway Administration.The results show that the Spiking-GRU neural network has good prediction accuracy and convergence speed.(3)Urban road network traffic state prediction considering large vehicles factors.The driving flow is taken as the research object,and the relationship between the proportion of large vehicles and the traffic state score of road network is analyzed by grey relational analysis.Based on the temporal and spatial characteristics of road network,the prediction time is analyzed.Through the vehicle data of road network simulation,the traffic state of the congested road is predicted.The experimental results show that the Spiking-GRU neural network has better prediction accuracy and convergence speed when it is consider large vehicles factors.According to the characteristics of traffic data,road traffic state prediction is analysed from three aspects:data preprocessing,deep learning neural network and urban road large vehicles factors.The results show that the deep learning algorithm of the SNN model can predict the road traffic state accurately and provide a good reference for traffic management and control.