Forecasting Method And Application Of Traffic Flow Based On 5G Vehicle Network

With the continuous improvement of people's living standards,urban travel has become more and more diversified.The increasing volume of travel has caused tremendous pressure on the current transportation system.Under this background,intelligent traffic control systems have emerged and developed rapidly.Where,urban short-term traffic flow forecasting is particularly important as an important link and branch in intelligent traffic control systems.Then the car network realizes the interconnection between the car and the car,the car and the road,the car and the person and the car and the Internet through the V2 X transmission protocol and the wireless communication,so as to ensure the dynamic interaction of the information and the intelligent control of the vehicle.5G mobile network has many advantages such as high speed and low latency.This paper designs a 5G vehicle network traffic flow prediction model by combining the Internet of Vehicles with the 5G network,and is committed to promoting the breakthrough development of ITS.Firstly,aiming at the inefficiency of KNN algorithm in short-term traffic flow forecasting,this paper proposes an improved KNN short-term traffic flow forecasting method based on non-parametric regression.This method is based on the idea of non-parametric regression.Through the KNN algorithm,candidates input data similar to the current state are searched for the system in the non-predictive period of time,and then pre-processed.The best decision input data for prediction is identified from candidate input data.Finally,the best decision input data is used to generate prediction through prediction algorithm.Secondly,in order to solve the problem that the gradient of RNN is easy to disappear or explode in the short-term traffic flow prediction based on neural network,a short-term traffic flow prediction method based on the combination of KNN and LSTM is proposed.In this method,KNN algorithm is used to select the similarity of the road sections in the predicted target area.By setting different thresholds,K detection points which are most similar to the target monitoring points are selected.Then,the K detection points selected by the commander are input into the LSTM model for traffic flow prediction,and the prediction result of K value corresponding to the minimum error is taken as the final short-term traffic flow prediction.Finally,aiming at the two short-term traffic flow forecasting methods proposed in this paper,based on the combination of 5G mobile network technology and vehicle networking technology,the forecasting model is designed for practical application,which includes three modules: information collection module,traffic flow forecasting module and information release module.