Study of traffic controls could slash drive times

Michael Chow/The Arizona Republic Metered stoplights on freeway ramps (top) trigger a variety of responses from drivers: Some people ignore them, others come to a screeching halt. Carl Jones (above) keeps an eye on the flow at the Traffic Operations Center in Phoenix.

By Kerry Fehr-Snyder
The Arizona Republic
Aug. 24, 1999
Some drivers gun right through them. Others screech to a dead stop - forcing everybody behind them to do the same.

Still other drivers perform that commonyet logically impossible maneuver called the "rolling stop."

No one seems sure how to respond to those metered stoplights on Valley freeway ramps, in part because engineers find them tricky to time properly even with the aid of high-powered computers.

Now, a University of Arizona researcher is conducting experiments at several Valley streets and one interchange that could - if successful - eventually reduce drive time by as much as one-third.

One key to the experiments is examining how the most unpredictable piece of the traffic puzzle - you - affects how quickly or slowly you get to work.

"The goal," engineering Professor Pitu Mirchandani said, "is to decrease the waiting times" - a bland statement that's music to every commuter's ears.

In fact, Mirchandani has developed three computer programs to move cars more efficiently on city streets as well as freeways.

The programs, known as computer algorithms, are being tested at stoplights on streets and freeway interchanges in Phoenix, Tempe, Tucson, Seattle and maybe Santa Clara, Calif.

The first program deals with traffic signals that line long corridors, such as Rural/Scottsdale Road. Called "real-time adaptive control," the system uses a series of copper loops buried underneath the road to measure the volume of cars approaching traffic lights.

"It's like an intelligent cop looking down the road and seeing what's coming," Mirchandani said. "It's more dynamic than a fixed system."

By seeing the traffic flow before it arrives, the system can adjust the lights in real time to benefit the biggest volume of cars.

 

Michael Chow/The Arizona Republic Metered stoplights on freeway ramps (top) trigger a variety of responses from drivers: Some people ignore them, others come to a screeching halt. Carl Jones (above) keeps an eye on the flow at the Traffic Operations Center in Phoenix.

Mirchandani also had worked out a computer program that moves traffic through freeway interchanges more efficiently.

Such interchanges are among the most complicated in traffic management because they involve lights that control cars at a previous intersection as well as the one at a freeway overpass or underpass.

"These two signals have to be coordinated, and it's difficult because the distance is so short," Mirchandani said.

Taken together, these two computerized systems are designed to cut trip times by 15 percent to 30 percent, according to UA computer models.

"If those kind of numbers really happen in the field, that's huge," said Tim Wolfe, an engineer with the Arizona Department of Transportation.

But less travel time isn't the only benefit.

"That equates to huge (pollution reduction) because people aren't idling at lights as long," Wolfe said. "And there are safety benefits because they're not having to stop at as many lights."

The final algorithm developed by Mirchandani involves those ramp meters that run along big stretches of freeway, specifically along Interstate 10.

By making ramp meters run more slowly on the entrances farthest away from the city's core, the system limits - though can't eliminate - bottlenecks that form on the freeway.

Unlike the other two systems, the ramp meters won't necessarily allow drivers to pile on the freeways faster than they otherwise could.

"On the (ramp) stuff, all we're really trying to do is manage the capacity there," Wolfe said. "We may not see huge trip reductions but we'll be able to manage them better."

But a later, more ambitious version of the ramp metering software will also tackle that greatest of variables - individual driving habits.

"The traffic light will learn the local differences of drivers," Mirchandani said - meaning that the ramp lights will eventually be metered at different rates, depending upon the behavior of drivers using that particular ramp.

The version being tested now can only function on the basis of the behavior of an average Arizona driver - who may be more mellow, relatively speaking, than is generally thought.

"In general, Arizona drivers are less aggressive than California drivers," Mirchandani said.

"And in general, California drivers are less aggressive than Boston drivers, so you can adaptively control the signals to control the behavior."

Kerry Fehr-Snyder can be reached at kerry.fehr-snyder@pni.com via e-mail or at 1-602-444-8975.

Home | General Info | Current Research | Researchers and Staff | Publications and Video | Media Info | Contact Atlas | Other Links