Too many cars equals a traffic jam, even without an external cause.

Mathematical Society of Traffic Flow

Traffic Jams Happen, Get Used to It

Ever wondered what causes those inexplicable traffic jams on open stretches of highway, the ones without any accidents, construction, or other obvious bottlenecks? A Japanese group has an answer: It's pure physics. A simple experiment shows that when the density of vehicles on a road passes a certain threshold, traffic jams emerge because of fundamental instabilities inherent in multiparticle interactions. In other words, just a few mildly inconsistent drivers on the road will eventually cause a wave of backups.

The cause of so-called phantom traffic jams has been quite controversial, says Dirk Helbing, who studies the physics of social interactions at the Swiss Federal Institute of Technology in Zürich, Switzerland. One camp of traffic researchers believes that even phantom jams have external causes, be they merging traffic, curves, hills, or even a few bozos abruptly changing lanes. But other researchers contend that jams will spontaneously appear simply if the vehicle density exceeds a certain critical value. Yuki Sugiyama, a physicist at Nagoya University in Japan, says the predictions of these models have matched observations of highway traffic.

But no one had performed a controlled experiment, so Sugiyama and his colleagues got volunteers to drive varying numbers of cars single file around a circular track 230 meters in circumference. Vehicles started out uniformly spaced, and drivers were told to cruise at 30 km/h and maintain a safe distance from the car ahead. As clearly seen in a video of the experiment, traffic initially flows smoothly. But because the drivers don't all maintain exactly the same speed, soon there are minor variations in vehicle spacing, and within less than a minute cars bunch up at one point around the track. Drivers have to slow down and even stop within the cluster of cars then speed up as they escape it. The threshold of critical density was 22 vehicles: Bunch-ups occurred with 22 or more cars, but they resolved themselves with fewer than 22 cars. "To tell the truth, I was impressed the experiment really performed just as theoretically predicted," Sugiyama says. The study is published in this month's issue of the New Journal of Physics.

Helbing, who wasn't involved in the experiment, thinks all researchers will now have to include the effect of inherent instability in traffic flows in their models. He adds that this work contributes to a growing understanding of situations in which numerous humans interact. But making this pay off in less traffic congestion will require much more work.

Related site