Editor’s Note: Don Lincoln is a senior scientist at Fermi National Accelerator Laboratory. He is the author of “The Large Hadron Collider: The Extraordinary Story of the Higgs Boson and Other Stuff That Will Blow Your Mind” and produces a series of science education videos. Follow him on Facebook. The opinions expressed in this commentary are his. View more opinion articles on CNN.
It’s the final days of the holiday season – and that means crowds. People are everywhere. They are crowding into malls, bustling from shop to shop along busy city streets and traveling in huge numbers to visit loved ones. Bus, train or plane, it’s off to grandmother’s house we go.
Navigating that mass of humanity is a timeless tradition, which requires that we adroitly dart around people to avoid them and try to maintain our positive holiday spirit when we inevitably collide with strangers. It’s all very stressful. While the experience has long been with us, science is just beginning to better understand how people in crowds interact – how we try to move past one another and navigate our way through the chaos. The insights could help us build better corridors and thoroughfares that make it easier for us to get from place to place, and just maybe make holiday travelers a little jollier.
To begin to understand how walkers behave in a natural environment, researchers in the Netherlands mounted cameras in the Eindhoven train station and let them run for six months. These cameras were located in one of the main hallways commuters passed as they went about their daily lives. Everything from people moseying about during off hours to peak rush hour behaviors were captured on film by the cameras. Millions of commuters were recorded.
From this behavioral data, the researchers modeled the motion of the people they observed running, walking, and turning around. They then watched how pedestrians modified their behaviors as they passed near one another – how much they changed their path when people were far away and the choices they made to avoid running into those who came very close to them. They even recorded some chance collisions.
Their model focused on two distinct types of interactions. The first are what the researchers called “social forces.” Those are ones where two pedestrians could see each other in advance and could adjust their trajectories. If two people came too close to one another, they did what you’d expect – gradually angle away from each other, so they would pass at a comfortable distance. But, more interestingly from a scientific point of view, the data also gave measurements of such things as how far apart is considered a comfortable distance and just how much people changed their trajectory to avoid getting too close. The minimum comfortable distance seems to be about 75 or 80 centimeters, or 29.5 to 31.5 inches.
The second type of interaction was “collision avoidance,” which occurred when two pedestrians got very close before becoming aware of one another. In this situation, the walkers came almost close enough to collide before making a marked and large change in their direction. The study didn’t explicitly say, but I can’t help but suspect that these close collisions were often the result of people staring at their phones, rather than focusing on where they were going.
The data that was collected was used to build a mathematical model, which was then implemented as a computer simulation that was tested against the data to assess its accuracy. Researchers found that, using the model, they could simulate how people interacted in a real world, crowded multi-person environment. While the data against which they tested their simulations was the same from which they devised their model, the simulation reproduced a broad array of observed behaviors.
The study is still in its infancy, but the researchers are now trying to use it to predict the motion of people in crowds and to try to figure out ways to optimize conditions to improve the movement of large groups of people. They still have a way to go, but it’s a good beginning. The next step is to use their computer model to simulate the motion of people in denser and denser crowds.
This work, while still preliminary, can be applied to a huge array of pedestrian environments. For instance, in a terminal bus or train station, people tend to be walking more or less in the same direction, as everyone is going to work or home at the same time. In transfer terminals, pedestrian behavior is more likely to include people walking in different directions, as they change trains or buses. And in places like intersections, pedestrians are often crossing one another’s paths, which presents an entirely different series of problems.
By using this data to be able to build computer simulations of common and complex pedestrian environments, researchers will be able to simulate many different pedestrian configurations and then alter the parameters of the corridors or thoroughfares to see how efficiently people can get around. Thus, these measurements and the simulation that the researchers devised could be used to help formulate better designs for people movement in crowded environments.
High density pedestrian congestion isn’t a new problem, but as the world’s population increases, it’s a growing problem. This recently released research will allow researchers to quickly test a range of designs without needing to do things the old-fashioned way of simply building a new public place and seeing how things go. Hopefully, the result will help make the daily lives of city dwellers better.
And those chance occasional collisions recorded by Dutch researchers? No word yet on whether they accurately predicted the beginning of a typical romantic comedy.