Online Traffic Information and Forecast for Routing Services

This paper presents a practical implementation of two different algorithms for estimating and forecasting the traffic situation in a network composed of various types of roads. The two different methods cover different parts of the network: one is specifically aimed at freeway sections, the other one is based on micro-simulation and can explicitly handle traffic control like signal control. Both methods are implemented into an online system. This online system receives the raw detector data of the complete road network, controls the forecasting and sends back the forecast results. The whole system started practical operation in December 2004 and was running online in a 2-month test phase for each of the two test sites, Munich and Magdeburg. The work presented here was part of the German research Initiative INVENT, subproject Network Equalizer.