Neural Network Model of Traffic Flow Forecasting

The article presents the results of building a neural network model of traffic flow forecasting at the crossroads of the city of Krasnoyarsk. The obtained model is used in the automated traffic control system of the city. The forecast is based on data from traffic detectors located at crossroads that evaluate the characteristics of traffic flow in real time. The developed models show a good result in forecasting the number of passing cars through the considered crossroads. The predicted values of the number of passing cars over a large interval are quite close in value to the true values.