The traffic data collection from vehicles offers new possibilities to traffic engineers as far as the improvement of knowledge on traffic patterns evolution is concerned. One might intuitively expect that the use of these facilities by a large number of subscribers would allow for a complete knowledge of dynamic O/D (origin destination) on the network so that every further data collection might be avoided. In this paper, both the limit of validity of this assumption and a method for the integrated use of floating cars data and conventional sensors for the real time estimation of traffic patterns are presented. The technical approach is to consider the availability of all the models and strategies developed within the CAR-GOES project for real time turning flows estimation at junctions, within-day and day-to-day filtering of journey times and suitable flow-time models which have been improved by combining the two independent sources of information (e.g. cars and sensors). A technique is then presented for identifying the evolution of major traffic patterns on the network when the output of the previously mentioned models is available. (A) For abstracts of part A and B see IRRD 848020 and IRRD 848021
Samenvatting