The most important cities are nowadays equipped with Area Traffic Control Systems. Those systems provide their operators with amounts of data measured over the controlled network, including a great quantity of information, but which are always difficult to process due to the multidimensional nature of the traffic phenomena. The target of this study is therefore to develop a set of methods for processing those data and getting some practical indications for traffic systems implementation or updating. The first part of the study here presented deals with the first two steps of the processing: - a correction procedure making possible the processing of data sets including missing or meaningless data, - a global analysis procedure, resulting in qualitative informations on the traffic phenomena occuring in the studied network. A second report will present further steps of the processing: - automatic traffic situations classification, dividing traffic variations into homogeneous classes, in order to develop specific strategies (signal timings), - automatic selection of representative points for on-line choice of signal timings. (Author/publisher)
Abstract