Online Traffic Prediction Using Time Series: A Case study

Developments of technology and rise in vehicles on the streets have made the traffic management a challenging issue in societies. To apply Intelligent Transportation Systems (ITS) effectively, it is necessary to get access to the online data of traffic flow in specific locations. The application of these systems let the engineers to respond the traffic congestions easier and more effectively. In this paper, a time series model is proposed to predict the traffic flow for a certain intersection, which can be used to control the signaling of that intersection. One of the advantages of a model based on time series compared to other models is that having previous traffic data of an intersection, it will make it possible for engineers to predict the upcoming traffic condition of that intersection. Using this method for Moallem Blvd. in Mashhad demonstrated that the model is able to predict the traffic flow with 88.74% and 81.96% accuracy for 15 minutes ahead and 1 hour ahead, respectively.