Lehrstuhl für Computergestützte Modellierung und Simulation
Ingenieurfakultät Bau Geo Umwelt
Technische Universität München

MultikOSi – Support systems for urban events: Multicriterial integration for openness and safety



Research into scale transitions in context of computer-aided pedestrian simulations


Joint research project Homepage:



Team: P. Kielar, O. Handel, D. Biedermann, A. Borrmann


Cooperation partner: 

Hochschulreferat 6 - Sicherheit und Strahlenschutz (RefSi), TU München

Fakultät für Informatik und Mathematik, Hochschule München

Lehrgebiet Stadtsoziologie, TU Kaiserslautern

Fachbereich Mathematik, TU Kaiserslautern

Fachbereich Mathematik/Naturwissenschaften, Universität Koblenz-Landau

IMS Gesellschaft für Informations- und Managementsystem bmH

VDS GmbH - Veranstaltung | Dienstleistung | Sicherheit


Associated partner:

Strategisches Innovationszentrum der Bayerischen Polizei (SIZ) Bayerisches Landeskriminalamt



Das Bundesministerium für Bildung und Forschung (BMBF)

Running period:

01.09.2013 till 31.08.2016


Sub-project description

Events like public viewing, city festivals or concerts are an integral component of urban life. In order to create accident-prevcorridorenting safety concepts, the organizers and security forces of such events are still in need of scientifically proven planning aids. The purpose of the project is to support those responsible with a better foundation for decisions by evolving new methods combining and integrating known ones. The research objective of the CMS Department is to develop novel multiscale techniques for computational simulations of pedestrian flows. The simulation results help experts to plan and execute urban events.

A multiscale coupling of simulation models featuring different spatial discretization levels is to be developed for the purpose of reducing the disadvantages of the different models. Macroscopic simulation models, which rest upon a network abstraction of the scenario, may be highly efficient but ignore interactions between individuals. In contrast, microscopic models exhibit an excellent spatial resolution but tend to involve a great amount of computational input. In order to overcome the conflicting goals of high spatial accuracy and fast computation times, we will realize a scale-sensitive integration of existing methods. This approach will lower the impact of the respective downsides. Based on the given precision requirement, an area can be selected within a simulation scenario in order to implement a different simulation engine in each chosen region. The main scientific challenge of this approach is to research spatial resolution transition methods at the bounds of the simulations areas. The multiscale transition approach has to ensure that the simulation methods employed remain stable. A dynamic switching between different simulation systems can extend the promised multiscale techniques. Switching the simulation models leads to a zoom-in approach. It makes it possible to swap the simulation models of arbitrary areas of the scenario and thus change the spatial resolution required.

webseiteMultiskalen03 eng 01

The second objective is to utilize existing information on the structural character of the venue as a direct input parameter for the simulation. For this purpose, we will research new methods for extracting geometrical information from digital building and city models. The scientific challenge is to provide all the information acquired at an appropriate level of abstraction for simulation purposes.

A demonstrator implementation will display the feasibility of the acquired methods and concepts. In this context, the new simulation framework MomenTUM v2 pedestrian simulator was developed.



▪  Kielar, P. M.; A., Borrmann:
Spice: A Cognitive Agent Framework for Computational Crowd Simulations in Complex Environment
Autonomous Agents and Multi-Agent Systems, 2018
DOI: 10.1007/s10458-018-9383-2

doi pdf bibtex mediatum
▪  Kielar, Peter M.; Biedermann, Daniel H.; Kneidl, Angelika; Borrmann, André:
A unified pedestrian routing model for graph-based wayfinding built on cognitive principles
Transportmetrica A: Transport Science 14 (5), pp. 406-432, 2017
DOI: 10.1080/23249935.2017.1309472

doi pdf bibtex mediatum
▪  Bareth, Thomas:
Baudimensionierung hinsichtlich Fluchtwegesicherheit und Komfort – eine Betrachtung von ingenieurtechnischen Berechnungsmethoden vor dem Hintergrund der baurechtlichen Vorschriften
Betreuer: Kielar, Peter
Masterarbeit, 2017

pdf bibtex mediatum
▪  Handel, O.; Borrmann, A.:
Service bottlenecks in pedestrian dynamics
Transportmetrica A: Transport Science 14 (5), pp. 392-405, 2017
DOI: 10.1080/23249935.2017.1280712

doi bibtex mediatum
▪  Lahr, S.:
Durchführung einer mikroskopischen Personenstromanalyse zur Optimierung und Evaluation der Abläufe, von Umbaumaßnahmen des Münchner Hauptbahnhofes
Betreuer: Kielar, Peter
Masterarbeit, 2017

bibtex mediatum
▪  Kielar, Peter M.; Hrabák, Pavel; Bukáček, Marek; Borrmann, André:
Using Raspberry Pi for Measuring Pedestrian Visiting Patterns via WiFi-Signals in Uncontrolled Field Studies
In: Proc. of the 12. Conference on Traffic and Granular Flow, Washington DC, USA, 2017

pdf bibtex mediatum
▪  Lauterbach, Sven:
Performanceoptimierung von Fußgängersimulationen durch Einsatz von Parallelisierungstechniken
Betreuer: Peter, Kielar
Masterarbeit, 2017

pdf bibtex mediatum
▪  Biedermann, D. H.; Kielar, P. M.; Riedl, A. M.; Borrmann, A.:
Oppilatio+ - A data and cognitive science based approach to analyze pedestrian flows in networks
Collective Dynamics 1, pp. 1-30, 2016
DOI: 10.17815/CD.2016.9

doi pdf bibtex mediatum
▪  Cheng, Zhibin:
Modelling Pedestrian Group Behaviors on a Music Festival Event Using an Agent-based Method
Betreuer: Borrmann, A.; Handel, O.
Masterarbeit, 2016

bibtex mediatum
▪  Thieme, Christian:
Implementierung einer Sensorik für virtuelle Fußgänger im Kontext der agentenbasierten Modellierung und Simulation
Betreuer: Bungartz, H. J.; Handel, O.
Bachelorarbeit, 2016

pdf bibtex mediatum
Alle Einträge anzeigen


Peter Kielar

© 2021 CMS TUM | Technische Universität München