A procedure for urban route travel time forecast based on advanced traffic data: Case study of Rome

In this paper, authors report a procedure to forecast the route travel time, based on different advanced traffic data, both historical and current. The procedure is articulated in two main steps: the first consists in apportioning current route travel time among the links based on historical Floating Car Data; in second step, the obtained link travel times are combined with the current loop detectors data through the measurement data fusion technique; moreover, the Extended Kalman Filter (EKF) correction is applied to a second order traffic model, in order to forecast the link speeds and consequently link and route travel times. Real applications of the proposed procedure concerning both a signalized route of Lungotevere and 6.5 kilometers urban freeway with 14 on/off-ramps in Rome are reported. The results of the forecasted route travel time show a good accuracy until the detected data used to correct the estimation are coherent and the input data, such as flows on ramps, are reliable. The results of the link speeds in terms of RMSE and RME with respect to the data detected by the loop sensors seem to be accurate, however further development will deal with the comparison to the other existing methods. Finally, the results underline, as it is well known, the need of real and reliable data in order to provide a good forecast.