Dr. Majid Sarvi and his team from the Department of Civil Engineering and Institute of Transport Studies at Monash University have been conducting some rather unusual experiments which they say could help design buildings that are easier to evacuate in an emergency.
The team has been looking at the effects of emergency exit area layout in a panic escape scenario using stressed ants.
Doors and corridors are necessary elements in public infrastructure such as railway stations, airports and stadiums. Previous documented crowd disasters have showed that collective movement patterns are affected by the layout or the geometrical structure of the escape area. However, little research has been carried out to examine these interactions under panic situations due to a scarcity of data on human panic.
Using a bio-inspired approach “at low cost and with no need for ethical approval,” Sarvi has tested whether making appropriate architectural adjustments within a given escape area would change the collective movement patterns in a way that enhances the outflow of the crowd.
First, the team performed a series of experiments with ants under panic conditions to test the effects of different structural features to the panic escape in a chamber with fixed dimensions but with varied exits and obstruction positions.
The ants were forced into their fleeing pattern by use of a harmless citronella insect repellent. Those that tried to exit through holes in the middle of walls found their path blocked, leading to confusion, whereas those that headed to the corners were able to escape far more quickly.
By tweaking the exit points, results showed that evacuation times could be decreased by more than 90 per cent.
The experiment was then scaled up with a computer simulation of a human scenario. Using the optimum layout from the ant experiment, the team discovered that evacuation times were reduced by 62 per cent when people in the model left via a corner exit, even if it had a large structural column in front of it, compared to a lone exit in the middle of a wall.
Sarvi says the experiment clearly shows that careful design of small structural features could have significant effects on pedestrian traffic outflow. Insight into such microscopic variations would assist in advancing understanding of which properties of panic are inherent to the physical nature of the crowds and which properties depend on the idiosyncratic details. This gives additional perspective to aid in devising solutions that are more efficient and improve the safety of crowds.
Speaking to the New Scientist in the UK, Paul Townsend, simulation and modelling manager at Crowd Dynamics, has advised caution, however.
“Where real evacuations have been filmed, very rational behaviours are observed – with people helping people they have never met before,” he said. “So to relate ant behaviour to that of humans, a precise definition of ‘panic behaviour’ is needed if this is to properly inform space design,” he says.
To study non-panicking behaviour, Sarvi and his team will now turn to the woodlouse. As a larger species, Sarvi hopes the woodlouse will move en masse in normal modes, which are almost impossible to observe with ants.