Using neural networks for route and destination prediction in intelligent transport systems

Abstract

Route prediction and destination prediction based on the past routes are a missing piece in intelligent transport systems (ITS). These predictions can be useful in many areas: congestion prediction, traffic control, upcoming traffic hazards and targeting advertisements next to the roads are some of the obvious ones. Simply said, if we can estimate the future location of cars which are already on the road network, we will be able to estimate future congestions and upcoming traffic hazards. The GPS units in the new generation of smartphones provide a good data source for prediction algorithms. Google maps application already collects this data. This paper discusses several algorithms and methods which have been used in similar areas and a route prediction method based on artificial neural networks using the past routes of a vehicle. © 2012 Springer-Verlag.