Sign In

Communications of the ACM

ACM TechNews

Avoiding the Crush

View as: Print Mobile App Share:
People avoid walking into each other at the very busy Shibuya Crossing in Tokyo.

Researchers at the Department of Computer Science and Engineering at the University of Minnesota have developed a mathematical law that helps explain why people are able to move through crowds without running into each other.

Credit: Mlenny Photography/Getty Images

New crowd-modeling research by Minnesota University computer scientists describes a mathematical law called time-to-collision that helps characterize a person's ability to move through crowds without running into another.

Professor Stephen Guy says crowd modeling has up to now been limited by a failure to concentrate on how people actually move, instead of on how they think they move.

"One of the most important things is that we can anticipate the future," Guy notes. "I have a brain, you have a brain. When we see each other coming, we know what's going to happen in the near future and we can use that to anticipate and avoid each other."

The time-to-collision law suggests all individuals make subconscious calculations when moving, determining ahead of time who or what is likely to collide with them; these calculations are concurrent and continuous, but Guy says the only evaluations spurring action are those where a person or object is determined to be within two to three seconds of collision.

He says the visual data for proving the theory's veracity is widely available. "Thanks to surveillance cameras, thanks to advances in computer vision, we can get hundreds and hundreds of trajectories of people walking in different kinds of environments," Guy says.

From Australian Broadcasting Corp. News
View Full Article


Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA


No entries found