A Simple Math Equation Could Solve Highway Traffic Jams Forever

Who needs self-driving cars? Unnecessary traffic jams might be caused by human’s innate inability to drive so well, but one MIT computer scientist says there’s an algorithm for that.

[Image: Flickr user Joe Thomissen]

Traffic, one of the most annoying conditions of modern life (if you own a car), often happens for no real reason. Roads have carrying capacities, sure, but even drivers on closed tracks have shown that traffic jams appear to be hardwired in human nature.

This process burns time, gas, and creates pollution. But an MIT professor may have solved highway traffic congestion, or at least the unnecessary kind. Computer scientist Berthold Horn has developed a counter-intuitive approach to smooth out naturally occurring clumps of freeway traffic that he says could be included in a car’s existing computer system. The trick, as it happens, may just be that drivers needed to look behind them.

Horn explains that drivers unconsciously follow an equation in their heads: Look at the car ahead, try to maintain a safe distance. If it’s larger, accelerate, and if it’s shorter, brake.

“This modified system, which I call bilateral control, uses information from vehicle behind you,” Horn says. “You try to maintain the same distance ahead as behind you. Think of the car connected in front by string, and distance in car behind is the same.”

If you think about traffic flow as a fluid, Horn says, the first equation–which he calls the Car Following algorithm–will always end in disaster, creating vibrations, or points that oscillate instead of moving forward in a steady stream. The mathematical proof for bilateral control shows that if all cars maintained equal distances in front and behind, oscillations would be eliminated.

But how might all cars make the shift? Many cars already have rearview cameras, and high-end vehicles have something called adaptive cruise control. Together, Horn says that same system could easily accommodate his bilateral algorithm.

The flow only really works if all cars do it–but nothing is lost if one car abides by the rules and others don’t. “For a while I was collecting GPS data in my own commuting to support this research,” Horn says. “I found at some point I had kind of unconsciously adapted my driving modes myself, and had smoothed out some of these oscillations.”

By eliminating traffic jams, Horn’s simulation shows that bilateral control could make commutes faster and more efficient. It’s also promising for better air quality.

“It’s well-known that smoother traffic flow would lead to lower emission rates in general. More stop and go traffic results in a higher emission rate,” says Greg Rowangould, an assistant professor of civil engineering at the University of New Mexico who researches the impact of traffic pollution. Rowangould also points out that this system could one day be less expensive than adding a lane to a highway to relieve congestion.

Still, Rowangould says that bettering congestion can be tricky, and sometimes can have the opposite effect. “When you relieve congestion, you get to where you’re going more quickly, and that reduction in traffic time tends to attract more cars to the roadway,” he says.