On-line Simulation and Traffic Forecast: Applications and Results

We study an approach to forecast vehicular traffic using a micro-simulator. In order to provide network-wide information about the current traffic state a cellular automaton traffic flow model is combined with measured data. The framework is applied to the freeway network of North Rhine-Westphalia, where data from about 3,500 loop detectors are available and provided on-line minute by minute. Further, heuristics based on the statistical analysis of historical data are developed. Combining these methods and the results obtained we present an Internet application. This can provide the traffic information given by the on-line simulation such as travel times and the current traffic state for a broad public.