Online traffic state estimation with multi-agent simulations
2008 (English)In: Proceedings of the 4th International Symposium Networks for Mobility, 2008Conference paper (Refereed)
This article describes a novel approach to online traffic state estimation with multi-agent simulations. The presented method identifies individual-level motorist behavior from aggregate measurements of flows, densities or velocities that are obtained at a limited set of network locations. Behavioral dimensions that can be estimated range from single route choice to the selection of full-day plans. The estimation logic only requires a simulation-based representation of the behavioral dimensions under consideration in order to simulate the most plausible driver behavior given the additional information contained in the sensor data. No mathematical description of the behavioral driver model is needed such that the broad modeling capabilities of a multi-agent simulation can be fully exploited.
Place, publisher, year, edition, pages
Transport Systems and Logistics
IdentifiersURN: urn:nbn:se:kth:diva-76843OAI: oai:DiVA.org:kth-76843DiVA: diva2:491216
4th International Symposium Networks for Mobility. Stuttgart, Germany. September 25 - 26, 2008
TSC import 1002 2012-02-06. QC 201205142012-02-062012-02-062012-05-14Bibliographically approved