A general approach to network status estimation and traffic prediction is a must for traffic control and user information ITS applications in the contexts where a variety of real time, quasi-real time and offline traffic data sources are available. In this paper, the models adopted for the implementation of the Network Observer - a key function of the Town Supervisor operational in Torino, are described. These models are employed to estimate homogeneous traffic status parameters on network links where traffic is measured through different techniques and on different time bases. Extensions are also available for the links where traffic is modelled by traffic demand and assignment complementary models. Some result is reported in graphic way, which relates to the application of the models to the Stockholm site during the Town Supervisor transferability analysis.
Abstract