Car Sensors Will Help Solve Traffic Jams

Stop tailgating already, say MIT researchers.

|
Dec 14 2017, 1:00pm

Image: Shutterstock

We’ve all been there before: That jerk in the beat-up blue sedan is riding your tail at 60 miles an hour on the freeway, trying to make you go just a little faster. Maybe you nervously inch closer to the car you’re following, lest this bozo rear-end you. The driver in front of you then brakes, and we’ve got ourselves a good old-fashioned traffic jam—if not an outright car crash.

Well, researchers at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) have got some news: If we all just stopped tailgating, we might get to our destination twice as fast.

Integrating rear sensors into existing adaptive cruise-control systems may be just the ticket. In a new paper published in IEEE Transactions on Intelligent Transportation Systems, MIT professor Berthold K.P. Horn and co-author Liang Wang write that having front and rear sensors would help keep equal distance between vehicles—and that could help solve our traffic woes.

The authors kick off their paper by painting a portrait of the problem: The average American urban commuter spends 38 hours a year stuck in traffic. Combined, stop-and-go congestion wastes 2.9 billion gallons of fuel and emits 56 billion pounds of carbon dioxide across the United States.

Horn thought of the notoriously gridlocked I-93 near Boston, which he regularly endures on his drive from rural New Hampshire to the Cambridge university. “I think it’s just a common frustration that we all have on the highways,” Horn told me on the phone. Our existing road network isn’t been used to its maximum potential because of our inefficient driving style. “We fix so many problems—so why can’t we fix this one?”

Read More: New AI Algorithm Beats Even the World's Worst Traffic

We drive so closely to each other that every minor traffic disturbance creates a domino effect, which leads to traffic jams. However, the authors posit that maintaining an equal distance in front of and behind us would dramatically reduce those disturbances, as well as travel time and fuel consumption.

Of course, there’s always a difference between what we know we should do and what we actually do. Most drivers instinctively like to follow another car and we hate it when people cut us off, said Horn. “They want to be as close to the car in front as possible, but that really creates a problem.” So how do we make drivers actually stop tailgating?

Sensors might be the answer. Automakers are already producing cars with adaptive cruise-control systems—that is, a sensor embedded in the front grille that works to adjust a vehicle’s speed and distance from cars up ahead.

The MIT CSAIL researchers suggest adding similar sensors in the back of the car would further optimize traffic flow.

Executing this concept—dubbed “bilateral control” by Horn—could be done using any kind of sensor, be it radar, lasers, or cameras; as a professor of machine vision, Horn’s got a preference for cameras, which are also the cheapest option comparatively speaking and are already in many new cars.

Additionally, while the system would require the vehicle itself to have some intelligent capabilities, it wouldn’t require the kind of interconnectivity that autonomous cars require.

The researchers will be continuing to test their ideas with Toyota in the coming year—as well as working on ways to inform the public about the advantages of hanging back.

Get six of our favorite Motherboard stories every day by signing up for our newsletter .