We study how we can automatically create a data structure that represents the walkable surfaces in virtual environments, and how it can be updated dynamically and efficiently when it changes. We refer to this structure as a navigation mesh. This mesh enables efficient crowd simulation, which is our next topic of research. We study and develop a crowd simulation framework and its components, which ranges from global (AI) planning to local animation. We create models for realistic crowd behaviors, which includes studying how (groups of) people move and avoid collisions in such environments, based on agent profiles and semantics (such as terrain annotations).
Our crowd simulation research and development has appeared at many places. On this page, we present many examples of case studies that we and companies have done by using our software.
The Grand Départ of the Tour de France was expected to draw anywhere from 600,000 to 800,000 spectators to Utrecht in July 2015. Simulations in a virtual Utrecht support the municipal government in planning the surroundings of the cycling course and testing the prognoses for the flows of the spectators. These innovative simulations in an unprecedentedly detailed virtual world were the result of collaboration between computer scientists at Utrecht University and the engineering and consulting bureau Movares. Read more...
This project was carried out with Movares (Jan Beumer), city of Utrecht (Ron Looy) and MSc student Marijn van der Zwan.
In 2016, we studied how long it would take to clear each of three metro station that are part of the new Noord/Zuid-lijn in Amsterdam. We compared these evacuation times with scenarios in which a small percentage of the passengers brought a bicycle along with them.
This project was carried out with Movares (Gea Kolk), city of Amsterdam and MSc student Martijn Koenis.
We have created an interactive demo displaying a part of the city of Utrecht where thousands of pedestrians are walking around. Users can interact with the simulation by inserting or removing illuminated blocks. This allows them to play with different scenario's in an interactive and intuitive way.
This augmented-reality crowd simulation demo was especially created for all EU ambassadors and policy makers who paid a visit to Utrecht on May 20, 2016. The demo was also also shown at the Campus Party (2016), the Annual Congress Technology Foundation (STW) (2016), Weekend van de Wetenschap (2016), Betweterfestival (2016), the 4th National eScience Symposium (2016), and at the Utrecht Science Museum (2017).
The demo has been made possible by Utrecht University, Z25.org, the COMMIT/ project, and STW. Many people helped to realize and demonstrate this project, including my PhD students, Arne Hillebrand and Wouter van Toll, scientific programmers Angelos Kremyzas, Patrick Eijk and Rens de Heer, and Machiel Veltkamp and Wijnand Veneberg from Z25.org. Special thanks goes to MSc students Jordi Vermeulen and Mihai Polak who assisted me regularly.
In June 2016, a 2D path planning demo was given at the International Computational Geometry Multimedia Exposition (SoCG 2016) in Boston. It loads a number of 2D test environments and computes their ECMs. In the environment that is shown, an agent repeatedly computes and traverses a path to a random goal position. Users can switch between environments and change the agent's goal, radius, and preferred clearance to obstacles. The demo was created by my PhD student Wouter van Toll. You can download the Windows demo here.
Our team, consisting of Wouter van Toll, Arne Hillebrand, Angelos Kremyzas, Mihai Polak and me, has won NWO's ICT.OPEN 2016 Meet the Demo award for 'Towards making a city smarter'. The demo showed a city center in which crowds were simulated. This Unity3D application featured a crowd that was composed of groups of people. See our publication on social groups and corresponding movie for more information.
In this project, we extended our UU crowd simulation software to calibrate a crowd simulation in real-time (using video camera's) while an event is going on. In addition, we programmed a prediction model that can predict what the (near) future state of a crowd will be. The project was showcased during Festival deBeschaving in Utrecht (see below), executed by Arne Hillebrand. Read more...
This project is the result of a software project carried out by our bachelor computer science students: Patrick van Eijk, Rens de Heer, Raphael Claasen, Sophie Huiberts, Martin Korevaar, Jurriaan Pijpers, Adolfo Ochahavia, Tim de Bie, Patrick Habekotte and Guido Oskam. Special thanks goes to Atlas Cook (voice), and my PhD students Arne Hillebrand and Wouter van Toll for their endless effort.
Welcome on my home page on crowd simulation and motion planning
Dr. Roland Geraerts | Associate Professor | Program Director Bachelor Computer Science | Information and Computing Sciences | Utrecht University | Princetonplein 5, room 4.08 | 3508 TB Utrecht | r.j.geraerts@uu.nl | www.uu.nl/staff/RJGeraerts
Our department page describes how you can visit us.
