An integrated method for short-term prediction of road traffic conditions for intelligent transportation systems applications

The paper deals with the short-term prediction of road traffic conditions within Intelligent Transportation Systems applications. First, the problem of traffic modeling and the potential of different traffic monitoring technologies are discussed. Then, an integrated method for short-term traffic prediction is presented, which integrates an Artificial Neural Network predictor that forecasts future states in standard conditions, an anomaly detection module that exploits floating car data to individuate possible occurrences of anomalous traffic conditions, and a macroscopic traffic model that predicts speeds and queue progressions in case of anomalies. Results of offline applications on a primary Italian motorway are presented.