Within traffic management systems online traffic flow data from various measurement sites is prepared in order to model a consistent traffic state for the entire road network. Typical sources for online data are detectors of vehicle actuated signal control systems and dedicated autonomous sensors communicating via GSM. Based on the current state of the network, a short term forecast can be computed to estimate the development within the next 15 to 30 minutes. To estimate the current state of the network, an assignment method called path flow estimator is used, and a mesoscopic traffic flow simulation tool is used to compute the forecast. The algorithms for state estimation and forecast must be supplied with information about the infrastructure, i.e. essentially the road network, and information about traffic demand. Traffic demand is provided in the form of origin-destination-matrices constructed from a set of preclassified matrices and matrices representing traffic demand induced by special events. A typical traffic management system is designed in a modular way. Static information about road network and traffic demand is managed by a graphical modelling tool, interfaces to systems providing online data are implemented as independent subsystem converters that transform the data in standardised formats and store them in a data base. The state estimation and forecasting procedures interface to the data base and can be adapted to different technical environments in a flexible way.
展开▼
