Traffic flow prediction using neural network

Abstract

Traffic flow management and analysis have become essential for both individuals to better manage and route their daily commutes, and for transportation planners to optimally schedule road infrastructure maintenance tasks. Therefore the ability to predict the nature of the traffic stream accurately is one of the most important requirements of traffic management systems. In this research, we will propose an intelligent method to predict traffic flow based on real data for the years 2016 and 2017 provided by the Moroccan center for road studies and research. The proposed solution focuses on training a neural network model to estimate future traffic flow on an hourly basis. Results determined by the simulation gave a good prediction to the traffic data.