MIT scientists and Maine students team up to tackle traffic problems

Traffic jams are a growing problem in Maine and elsewhere. Maine students and their robots are part of the solution.

This past August I found myself parked in the left lane of I95 on a usually lonely stretch of pavement between Bangor and Waterville. Stuck in a traffic jam that would extend my driving time by more than an hour, I expected to see some road construction, an accident or something that would explain my frustration. Instead, I saw nothing; after an hour of crawling the jam was gone and I was free to resume barreling down the highway.

These “phantom” traffic jams are becoming more common across the US, even in Maine. Increases in tourist visits to places like Acadia National Park have traffic on the rise in Maine, leading to numerous accidents and road closures this past summer.

Most traffic jams (besides those caused by construction or freak accidents) are caused by simple rules of driver behavior. Imagine this scenario: you are enjoying a drive up Coastal 1 when the tourist operating the car in front of you spills ice cream on his lap, causing him to slow down momentarily. You slam on your brakes to avoid rear-ending the tourist, and the driver behind you slams on their brakes to avoid hitting you, and so on. After the tourist in front of you removes the ice cream from his pants, he accelerates again. You wait for a moment before you accelerate in order to maintain a safe following distance, and the driver behind you does the same. These reasonable behaviors can cause traffic jams to develop; while you only slowed down momentarily, cars farther behind you in line will end up stopping. In this way, a traffic jam will grow behind you even though there was no restriction on traffic flow besides one car slowing down momentarily. Hence, the phantom traffic jam.

Scientists have been working to generate a solution to these phantom traffic jams for years. Besides being frustrating, traffic jams waste time and energy and can increase wear on roads and cars.

Professor Berthold Horn of the Massachusetts Institute of Technology’s (MIT’s) Computer Science and Artificial Intelligence Laboratory (CSAIL) has developed a simple solution to phantom traffic jams. Called “bilateral control,” the solution involves adjusting your speed in response to the car in front of you and the car behind you, with the goal of staying close to halfway between the two cars. Using calculus and compelling visual simulations, Professor Horn has shown that this bilateral control can quickly erase traffic jams.

To rigorously show that bilateral control can ease traffic jams, Professor Horn needed to demonstrate that his method worked with actual moving vehicles, not just simulations. Unable to recruit hundreds of drivers to be stuck in traffic for his studies, Professor Horn decided to build small robots that behaved like Massachusetts drivers. To help him develop these robots, Professor Horn turned to budding Maine scientists.

This past November, Professor Horn teamed up with middle and high school students at Gizmo Garden in Damariscotta. Founded by engineers and educators Bill and Judy Silver, Gizmo Garden offers programs to foster robotics and creativity in young Maine women and men with an emphasis on gender equality and a creative, rather than competitive, engineering environment.

Gizmo Garden students Rachel and Heaven work with MIT Professor Berthold Horn to build self-driving robots.

Gizmo Garden students Rachel and Heaven work with MIT Professor Berthold Horn to build self-driving robots.

Professor Horn and the Gizmo Garden gang (no, this is not the title of a Disney movie) built a fleet of small robots that act like real drivers, following each other and attempting not to crash. Professor Horn and his team at MIT are using these robots to test their bilateral control idea. Generating traffic patterns with these robots is easier, cheaper and less dangerous that creating traffic with real cars and drivers. Bill Silver, Gizmo Garden founder said, “This is a win-win-win-win. Professor Horn gets the robots he can’t find off-the-shelf. Gizmo Garden students get to put their electronic assembly skills to real-world use. And I get to play with robots.”

Robots

Professor Berthold Horn of MIT and Bill Silver of Gizmo Garden test gizmos for Horn’s traffic flow research.

How will Professor Horn’s bilateral control idea help real drivers? Incorporating the algorithm into adaptive cruise control, which is now available in about 30 different makes of automobiles, could put a dent in traffic jams. In a futuristic world of self-driving cars, bilateral control could even make traffic jams a thing of the past.

If you have a question about science in Maine or see some cool science you want to share, please get in touch via email, Twitter, or Facebook!

For more info on Gizmo Garden, check out their website or Facebook page.

Zachary Wood

About Zachary Wood

I am a hiker and kayaker turned research scientist. I work as a PhD student at the University of Maine in Orono, where I research fish evolution, teach biology and take every excuse I can to escape to the University Forest. I spend the rest of my time roaming the state of Maine, searching for interesting ecology. There is an amazing amount of research that happens in Maine’s outdoors. My goal as a writer is to seek out and find the most exciting natural science that hides just beyond our sight – underwater in our lakes, up high in our trees and swimming through our rivers. I have studied ecology from Costa Rica to Greenland, but I feel most at home living and working in Maine. My work has appeared in news publications, scientific articles, and several science symposia. I need your help! If you see something interesting or have a question about science in Maine, please get in touch via email zachary.t.wood@maine.edu or Twitter @wacharyzood