Technology and quality

Simulating and predicting the behaviour of crowds

Calcolare la congestione di una folla

Today, more than ever, at large or small events, it is easy to find yourself suddenly surrounded by a large number of people. Just think of the crowds outside a concert, museum or sports event.

Being transported, pushed and suffocated by a chaotically moving crowd can be a terrible experience even for the most ardent of fans.

To mitigate the crowding and congestion of people in a limited physical space, Mitsubishi Electric Corporation has developed the world's first system to analyze real-time surveillance camera data in order to provide estimates and predictions for crowd movement towards event sites.

All of this is the fruit of a collaboration with the Research Center for Advanced Science and Technology (RCAST) of the University of Tokyo. The result is a real live simulator of crowd movements with a precision rate of 80%, approximately 30% more than conventional methods.

The high-speed RCAST human behaviour analysis model dramatically reduces the complexity of calculations, simulating crowd movements and focusing on the proximity of individuals.

Mitsubishi Electric Corporation has thus developed a quick and practical solution for event organizers, enabling them in real time to estimate areas at risk of excessive congestion and thus act accordingly with adequate countermeasures.

The result is increased safety for all participants, a more efficient event management and well-organized emergency routes away from situations of panic and danger.