News Release

Software uses in-road detectors to alleviate traffic jams

Peer-Reviewed Publication

Ohio State University



Cars pass over traffic detectors on a roadway in Columbus, Ohio. Each square outline cut into the pavement marks the spot where road crews have inserted a loop of wire. The wires act as metal detectors, allowing city engineers to monitor traffic flow. An Ohio State University engineer has developed software that detects traffic jams faster than previously possible, using data from the in-road detectors. Photo by Benjamin Coifman, courtesy of Ohio State University.

Full size image available through contact

COLUMBUS, Ohio – The same in-road detectors that control traffic lights and monitor traffic could soon respond quicker to traffic jams, thanks to software developed by an Ohio State University engineer.

In tests, the software helped California road crews discover traffic jams three times faster than before, allowing them to clear accidents and restore traffic flow before many other drivers would be delayed.

Benjamin Coifman

This technology could also provide drivers with the information they need to plan efficient routes, and even improve future road design, said Benjamin Coifman, assistant professor of electrical engineering and civil and environmental engineering at Ohio State.

Many drivers have probably noticed the buried detectors, called loop detectors, at intersections. A square outline cut into the pavement marks the spot where road crews have inserted a loop of wire. When a car stops over the loop, a signal travels to a control box at the side of the road, which tells the traffic light to change.

Though the loop detectors are barely more than metal detectors, they collect enough information to indicate the general speed of traffic, Coifman said. So he set out to use the detectors in a new way.

In the March issue of the journal Transportation Research, he describes how he was able pinpoint traffic congestion and accurately measure vehicles' travel time using standard loop detectors.

"Traffic is a fluid like no other fluid," Coifman said. "You can think of cars as particles that act independently, and waves propagate through this fluid, moving with the flow or against it."

With the software, a small amount of roadside hardware, and a single PC, a city could significantly improve traffic monitoring without compromising drivers' experience of the road, Coifman concluded. That's important, he said, because good traffic management can't be obtrusive.

"If transportation engineers are doing their job well, you don't even realize they've improved travel conditions," he said.

Coifman began this work while he was a postdoctoral researcher at the University of California, Berkeley. In 1999, he installed computer network hardware in control boxes along a three-mile-long stretch of road near the Berkeley campus, and took traffic data from loop detectors every third of a mile.

He then wrote computer algorithms that can capture a vehicle's length as it passes over a detector. Once a vehicle of similar length passed over the next loop, the computer could match the two signals and calculate the vehicle's travel time. Based on each car's travel time, the software was able to determine within three and a half minutes after traffic began to slow that a traffic jam had formed.

Because drivers' behavior isn't predictable, the new algorithms had to take many human factors into account. Among other factors, Coifman had to consider people changing lanes, entering and exiting from ramps, and "rubbernecking" -- the delay to drive time caused by people who slow down to look at accidents or other events.

"Traffic is a fluid like no other fluid," Coifman said. "You can think of cars as particles that act independently, and waves propagate through this fluid, moving with the flow or against it."

After an accident, it may take a long time for the telltale wave of slow moving traffic to propagate through the detectors. With the new algorithm, Coifman can detect delays without waiting for slowed traffic to back up all the way to a detector. This improved response time is important, because the personal and financial costs grow exponentially the longer people are stuck in traffic.

The detectors can't obtain any specific information about the make or model of car, he said, and a margin of error prevents the software from identifying more than a handful of cars in any one area at one time.

But that's enough information to gauge traffic flow, and the benefits to motorists can be enormous.

The average American city dweller wastes 62 hours per year stuck in traffic, according to the 2002 Urban Mobility Study by the Texas Transportation Institute. The institute measured traffic delays in 75 major cities, including Columbus, Ohio, where the average delay is 36 hours per year; Cleveland, where the average is 21 hours per year; and Cincinnati, where it's 43 hours per year.

According to the same study, traffic jams cost the average city $900 million in lost work time and wasted fuel every year.

The Ohio Department of Transportation (ODOT) has already begun using loop detectors to help motorists spend less time in traffic. When drivers head south into Columbus on Interstate 71 during business hours, an electronic sign just north of the city displays the average drive time into downtown.

As such information becomes more common, drivers can plan their routes more efficiently, Coifman said. He's working with ODOT to further improve travel time estimates.

The software would work with other vehicle detection systems too, such as video cameras. But installing these new systems can cost as much as $100,000 per location, and retrofitting existing equipment to use Coifman's software would only cost a fraction as much.

This work was supported by the Partners for Advanced Highways and Transit Program of the University of California, the California Department of Transportation, and the United States Department of Transportation, Federal Highway Administration.

###

Contact: Benjamin Coifman, 614-292-4282; Coifman.1@osu.edu
Written by Pam Frost Gorder, 614-292-9475; Gorder.1@osu.edu.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.