Menu Close

Recap CityFlows Webinar #3: Crowd Management at Multimodal Hubs and Train Stations

From:  Cornelia Dinca (AMS Institute)
Last edited: 11 November 2020

On Tuesday, November 3, a group of 23 crowd-management practitioners and researchers came together for the third edition of the CityFlows webinar series. The program consisted of two presentations and a discussion on topics related to crowd management at multimodal hubs and train stations.

If you missed the session, you can watch the full recording below or via

Innovations in Crowd Simulation and Forecasting

Roland Geraerts, Assistant Professor at Utrecht University and CEO of UCrowds, shared several innovations in crowd simulations and forecasting. He started with a few case studies illustrating the role of simulations in improving the safety and comfort of crowds. In one project, the evacuation of commuters was simulated for the North-South Metro line in Amsterdam.  Several “what-if” scenarios shed light on the impact on evacuation time when some people bring their bicycle on the metro. Another example is Utrecht Train Station where the station area serves not only as a transfer area, but also as public space. Simulations played a role in informing the design of street furniture to optimize the flow of people and placemaking functions.  Or at Schiphol Train Station, where an origin-destination model was created to evaluate different “what-if” scenarios, giving insight into possible bottlenecks and informing mitigation strategies.

Roland also shared two current projects where real-time crowd monitoring and prediction models are being developed at St Pancreas Station in London, and at Utrecht Train Station. At the St Pancreas station, thirty-two 3D cameras have been placed at entrance and exit points to measure crowd flows. This data automatically generates an origin-destination matrix which is updated in real time. A similar innovation project is underway at the Utrecht Train Station, measuring flows of people, bicycles and cars at the station area and generating a real-time model. The model is currently being validated, and in the future it will be possible to use it to make predictions and inform interventions.

Roland also shared several examples of projects showing how social distancing effects are being integrated into crowd simulations. He concluded his presentation with a live demo of UCrowds software, illustrating the process of building simulation models by “drag-and-dropping” different buildings blocks.  Roland emphasized the need for real data about the environment and the crowd being simulated, which are the foundations of any model.  

Improving Security at Transport Hubs: The NATO INSTEAD Project

The second presentation was by Luigi de Dominicis, Researcher at ENEA who provided an introduction to the NATO-funded INtegrated System for Threats EArly Detection (INSTEAD) project focused on improving security at transport hubs. The INSTEAD project is a collaboration between Italian, Finish and Dutch experts and part of DEXTER, a larger program that is developing innovative solutions for the detection of firearms and explosives in crowded areas.

The impetus of the INSTEAD project is the long and tragic record of suicide bombing attacks at transport hubs. In most cases, post-attack forensic analysis shows the attackers were spotted by the station surveillance systems. This brings up the question of whether it’s possible to develop a distributed active surveillance technology to detect potential terrorists in time to allow for a successful response. The INSTEAD project sets out to develop an integrated system consisting of hardware, software and human resources that cooperate to enhance the prevention capabilities of security operators in control rooms. This should allow for a fast and reliable identification of suspects and immediate activation of intervention protocols.

A key element in the project is the development of algorithms used by 2D cameras for tracking and reidentifying possible suspects. Such algorithms must be sufficiently robust to track and reidentify commuters even in massive flows, while also ensuring privacy. Another innovation is the integration of wearable tools using augmented reality to respond quickly in case an alarm is raised in the system.  The integrated solution will be tested at the Rome Underground Anagnina Station. The INSTEAD project is open to constructive feedback and evaluation from operators of transportation hubs.


Following these presentations, Dorine Duives, Principle Investigator at TU Delft, kicked off a discussion with a reflection and several questions.  A key take-away was that the rapid development of crowd simulation technologies improves the ability to make predictions about crowd behavior. The discussion also touched on the ethical and security considerations when using invasive technologies as in the case of INSTEAD project, and the need for governmental authorities to learn from each other about how to deploy these technologies in a responsible manner.

Join us for the last CityFlows webinar of 2020 :

Are you a practitioner or researcher working on a relevant crowd management project and would like to share your work and findings with the CityFlows network? Send a short email explaining your project to CityFlows Communications Officer, Cornelia Dinca via