Calibrating Traffic Simulation Models in SUMO Based upon Diverse Historical Real-Time Traffic Data – Lessons Learned in ITS Upper Austria

Traffic information services and traffic simulations represent a crucial element for today’s mobility. Traffic data may be gained using different types of sensor technologies and measurement approaches. However, there is no “one fits for all solution” related to the application of sensor technology and providers of traffic information services need to carefully decide when to apply which kind of sensor technology and measurement approach to provide traffic information. In Upper Austria, ITS Upper Austria represents such a traffic information provider. For the calculation of travel times and delays, real-time traffic sensors and a traffic simulation are currently in use. The latter is required when the amount of current real–time traffic information related to a link is too low for providing reasonable traffic information. ITS Upper Austria implemented its traffic simulation using the SUMO software. The demand model used for the simulation was built years ago, mainly using data from a household survey in Upper Austria in 2012. Based on this demand model, a route file was composed, which serves as input for the mesoscopic simulation. However, to increase the quality of the simulation, the route file needs to be continuously updated with respect to changing traffic behaviors (e.g. route traces, amount of cars). Different types of sensor data might trigger the calibration of traffic simulation models. For example Floating-Car-data, Bluetooth-data, data gained by permanent counting stations or even traffic times gained within test rides. Triggering updates of the traffic simulation model requires a careful analysis of the data basis and an appropriate update algorithm. This paper presents a traffic simulation update algorithm based upon diverse traffic data sources. Furthermore, findings related to the applicability of different sensor technologies for triggering simulation model updates are discussed. The findings stem from developments and empirical tests of ITS Upper Austria. The results could inform traffic E. Wießner, L. Lücken, R. Hilbrich, Y.-P. Flötteröd, J. Erdmann, L. Bieker-Walz and M. Behrisch (eds.), SUMO2018 (EPiC Series in Engineering, vol. 2), pp. 25–42 Calibrating Traffic Sim. Models in SUMO –– Lessons Learned in ITS UA Flitsch et al. information service provides when selecting sensor technology or when designing update mechanisms related to traffic simulation models.