Crowd cruncher

University of Melbourne researchers have developed a mathematical model which can predict and prevent dangerous crowd situations.

Engineering PhD graduate Ris Lee and her supervisor Professor Roger Hughes have developed a model for predicting when people are at greatest risk of being trampled or crushed in large crowds.

More than 2000 people are killed annually in crushing incidents such as the overnight stampede which has killed more than 800 people in a religious festival in Iraq.

Using data from some of the world's most notorious crowd disasters, Dr. Lee has tested a mathematical theory known as forward-backward autoregressive modelling to develop a system that can warn of impending crushes within five minutes.

These predictions could be used by organisers of events to suggest when crowd barriers should be collapsed or other crowd control methods such as directing the speed and flow of pedestrians could be implemented.

The research also identifies critical densities, points at which crowd masses become dangerous. It finds that people are in danger of being trampled when crowds reach a density of five people per square metres, and of being crushed at a density of about 10 people.