Real time traffic models, decision support for traffic management

Abstract

Reliable and accurate short-term traffic state prediction can improve the performance of real-time traffic management systems significantly. Using this short-time prediction based on current measurements delivered by advanced surveillance systems will support decision-making processes on various control strategies and enhance the performance of the overall network. By taking
proactive action deploying traffic management measures, congestion may be prevented or its effects limited. An approach of short-term traffic state prediction is presented and implemented in a real life case for the city of Assen in the Netherlands. This prediction is based on connecting online traffic measurements with a real time traffic model using the macroscopic dynamic
traffic assignment model StreamLine in a rolling horizon implementation. Different monitoring data sources consisting of both fixed-point and floating car data are used. The advantage of the rolling horizon approach is that no warming-up period is needed for the dynamic traffic assignment taking less computation time while keeping results consistent. Further, the current traffic state
estimation is done by combining model estimates of previous predictions and current measurements. The results of predictions made in the real life case are presented as well as several tested methods for improving the current state estimations showing promising results.