All in good time

Traffic-sensitive signals developed at the Robotics Institute are saving fuel, time and drivers' nerves

Gregory Barlow knows a lot about traffic, and not just how long it takes to commute from his home in Squirrel Hill to his office in Newell-Simon Hall, where he works with CMU research professor Stephen Smith on tackling traffic congestion in urban areas.

Barlow (CS'11), a post-doctoral researcher in the Robotics Institute, knows traffic problems are nothing new: The early Romans wrestled with them two millennia ago by banning wagons on their roads during certain times of the day. He also knows they're costly: In 2009, drivers in 439 urban areas in the United States traveled 4.8 billion hours longer and purchased 3.9 billion more gallons of fuel because of traffic snarls, for a total congestion cost of $115 billion.

And all this comes from Barlow, a man who admits, "I don't really like to drive."

Maybe that's why he's working to improve the driving experience in Pittsburgh, along with CMU colleagues Xiao-Feng Xie, a research associate, and Zachary B. Rubinstein, a senior systems scientist. They are part of a team led by Smith, director of the Intelligent Coordination and Logistics Laboratory in the Robotics Institute, where they developed a "smart" traffic signal that works in real-time. Smith has dubbed the project Scalable Urban Traffic Control (SURTRAC).  

Today, most signals operate by pre-setting the timing on the green light for several different periods throughout the day, depending on the number of vehicles expected to travel through an intersection. As part of the Traffic21 Initiative in Carnegie Mellon's H. John Heinz III College, the researchers developed software that allows a signal to respond to traffic as it is happening. Last spring, they installed adaptive signaling systems at nine intersections in Pittsburgh's East Liberty neighborhood. 

"Our system watches actual traffic flow through the cameras at each intersection and dynamically adjusts the green timing periods on a second-by-second basis," Smith says.  

According to Smith, similar software is already being used to manage traffic flow onto along arterial roadways--surface roads with a strong traffic flow in one direction and some side streets. With these systems, though, information is typically collected and sent to a central location, where data are analyzed and adjustments to the signaling are determined.  

"It can take anywhere from five to 15 minutes (to reprogram the signals), depending on the system being used," Smith says. "The beauty of our approach is that it is decentralized, which makes it inherently scalable, in principle." 

There also are decentralized systems used on arterial roads, says Smith, but they operate on the assumption that there is a dominant flow of traffic that does not change.  

SURTRAC, in contrast, is designed to discover the dominant flow of vehicles through an intersection and automatically adjust the signaling. It is designed to operate within urban grids, where the volume and direction of traffic can change throughout the day.  

At the East Liberty intersections, cameras take continuous shots of the traffic, which allows SURTRAC to create a schedule for moving vehicles through the intersections in the most efficient way possible. Each intersection also communicates via a fiber optic cable (or, in the case of one East Liberty intersection, a wireless signal) to its downstream neighbors as to what the projected outflow from its signal will be. The neighboring signals do the same thing, and together they create a communications network that's akin to having the watchful eyes of traffic police at every intersection. 

More impressive, though, are the net results. At the East Liberty intersections, the research team found that the wait time for people driving through the grid was reduced by 40 percent. Travel time was reduced by 26 percent, and projected vehicle emission by 21 percent. The researchers obtained the results by completing "before" and "after" drives along 12 pre-determined routes through the nine intersections. They did the drives during four set times throughout the day; GPS tracers on their cell phones allowed them to gather the data necessary to calculate the differences.  

Smith says the methodology is based on a model used frequently by the Southwestern Pennsylvania Commission, a regional planning agency that administers state and federal transportation and economic development funds.  

"When I saw the results, I was really shocked that they were as good as they were," says Allen Biehler, executive director of CMU's University Transportation Center and former secretary of the Pennsylvania Department of Transportation.  Although Biehler says that more "robust testing needs to be done" by expanding the pilot grid in an effort to obtain more data, he expects other cities throughout the United States will be interested in SURTRAC. In November, during the annual meeting of the American Association of State Highway Transportation Officials, Florida Transportation Secretary Ananth Prasad invited Biehler and Smith to discuss SURTRAC with his state's Department of Transportation. "This (technology) will have far-reaching implications," Biehler says. 

Nate Cunningham, director of real estate for East Liberty Development Inc., sees the ease in congestion at the East Liberty intersections as a boost to the local economy. "There is no more important thing to make a community competitive than being able to move people easily from place to place," Cunningham says. "I can see how (the ease in traffic flow) will have an impact on real estate in East Liberty." 

Another advantage, says Cunningham, is that municipalities can implement the technology incrementally. Once an intersection has been upgraded to include detection equipment, a SURTRAC system can be installed to monitor traffic in real time. "What's exciting is that this is so powerful and so cheap. Cities can bite off little chunks at a time, as funds become available." 

Amanda Purcell, municipal traffic engineer for the City of Pittsburgh, says that the pilot project in East Liberty "is working fine" and will help "establish a baseline for how the signals operate." 

She says talks are under way to expand the project, which has already received approval to install SURTRAC-supported signals eastward from East Liberty along Penn Avenue to Fifth Avenue. Nine more intersections will be added to the grid during spring of 2013. Adaptive signaling is also being considered along Baum and Center avenues to Craig Street in Oakland, creating what Smith likes to call "a virtual corridor to Downtown." 

A computer-enhanced corridor into Pittsburgh may not match the Roman accomplishment of "all roads leading to Rome," but a faster way into town should please many drivers. 

Traffic21 was launched in 2009 with funding from the Henry L. Hillman Foundation. Grants to Traffic21 from The Heinz Endowments' Breathe Project and from the Richard King Mellon Foundation provided the funding for the pilot.
For More Information: 
Jason Togyer | 412-268-8721 | jt3y@cs.cmu.edu