Managing crowds safely

Disaster hit at the 2010 Love Parade in Germany. What can research scientists do to prevent any repeat of this tragedy? How can we improve the safety of our citizens? Research into crowd safety under the German government's High-Tech Strategy is looking at these issues.

Enlarge image Techno-Fans enjoy Love Parade at the Brandenburger Gate in Berlin (© picture-alliance/ dpa) The white cap with a black dot in the middle does look a bit strange. And it is not only one young lady who is sporting this eye-catching head gear. More than one hundred men and women appear to share her unusual dress sense. As though they are remote controlled, the test subjects follow a designated route. Swiftly but without jostling. A camera films them from above. This is a research project into crowd control.

Professor Armin Seyfried from the Forschungszentrum Jülich is in charge of the Hermes Project. Numerous research institutes are involved as well as the police, fire brigade and security firms. Germany's Ministry of Education and Research is sponsoring Hermes under the Rescue and Protection of People Component of the Research Programme for Civil Security.
 
Judging crowds correctly

The tragic balance of the Love Parade in Duisburg in July 2010, read 21 dead and more than 500 injured. The organisers were way off in their estimates. Far more people came than expected. The access routes were too narrow. Pedestrians bunched, in a tunnel. Panic broke out, and the tragedy ran its course. Could this disaster have been prevented?

Professor Seyfried believes so, provided more information had been available on how crowds move and react. This knowledge and information about how crowds of people move and in what direction allow us to predict where and when routes can become too crowded.

We have not yet reached that point. The scientific research has not yet been performed. The tests conducted by the Hermes project began with lots of people wearing funny hats. This was vital so that the computer could initially evaluate the images of cameras. With the help of the hats the computers could trace the paths taken by the individuals, how they got in line and tried to make it to the exit.
 
The scientists took this data to devise mathematical models. In the form of symbols – circles and ellipses to indicate rapid movement, the movements of the crowds could be presented graphically. Can these findings be useful in real-life situations? Professor Seyfried is convinced that they can. He and his team are now working on an evacuation assistant.

Dealing safely with major events

Enlarge image Hard hats... (© dpa/pa) A computer-assisted system of this sort could make large-scale events much safer. It would offer support to organisers, security services, the police and the fire brigade by predicting at an early stage where crowds could bunch. Security staff could then steer the flows of people and possibly open additional exits if necessary. Conversely, individual exits or specific parts of the stands in a football stadium could be closed off for a limited time. This would make it possible to control different flows of people, so that they clear the area one after another.

Scientists are now able to recognise individuals using video images and computers even without the funny hats – another step towards practical application. On the basis of the automatically counted numbers of people in the different parts of a building, the software can forecast how crowds will move at the end of an event, or in the event of an emergency. Scientists are aiming to be able to predict flows within 15 minutes.

Along with the forecast, all services involved would be given the information they need to make decisions. This is built on models of dynamics in groups of pedestrians collated within the scope of extensive field studies and laboratory experiments. These models also examined construction law requirements, some of which are blatantly contradictory.
 
Averting risks

The Rescue and Protection of People Component focuses above all on situations which represent a risk for a large number of people. This might be in the wake of a terrorist attack, a major accident or a natural disaster. The aim is to identify problems at an early stage and take preventive action to manage crises. In this way research scientists can help initiate rescue and security measures swiftly and efficiently.

The Research Programme for Civil Security was launched in 2007. To date the Germany's Ministry of Education and Research has provided 235 million Euro  to ensure the safety of citizens. Activities are not exclusively technological in nature. Several projects are also looking at new organisational concepts and strategies for action. Research scientists and practitioners in a wide range of fields are cooperating very closely. The programme is aligned with European research policy.