April 25, 2007 By Jessica Weidling
Just as a music conductor guides an orchestra, making interpretative decisions as to the tempo of a music passage, real-time intelligent traffic systems now help cities conduct light-rail and side-street traffic, resulting in a harmonious flow of transportation through bustling city streets.
Such technology can be used to regulate car commute times, increase the viability of light-rail systems and avoid the ever-growing problem of congested roadways.
A group of Arizona cities is working toward this end and plans to implement a "predictive priority" system for the Central Phoenix/East Valley Light Rail Transit Project - now under construction and set to open in December 2008.
Valley Metro is the local agency responsible for public transportation in the area. Phoenix, Tempe, Mesa and Glendale created the nonprofit METRO light rail, under Arizona statute, to construct, design and maintain Valley Metro's light rail.
Predictive priority is meant to synchronize traffic lights to increase the smooth circulation of car and light-rail traffic. Phoenix's system balances the need to give priority to approaching light-rail trains - ensuring the fewest red lights for public transportation - without disrupting traffic flow, said Pat Fuller, deputy project manager of design and construction at METRO light rail.
"Without the system, the train wouldn't be competitive in regard to travel times with vehicles," Fuller said. "And we had to prove that we were able to compete with vehicles, otherwise people just wouldn't ride it."
The localized traffic intelligence system is based on complex communication networks - sensor networks to interpret characteristics of oncoming traffic, and mathematical and predictive algorithms that compute optimum settings for traffic light cycles.
Houston and Salt Lake City recently built light-rail predictive priority systems, but Phoenix's system will be more extensive because its 20-mile track snakes through major streets in Phoenix, Tempe and Mesa, requiring more intersections to be equipped with the technology.
Construction for the Central Phoenix/East Valley Light Rail Transit Project began in 2003 and is one of the largest infrastructure projects in Arizona's history. Fuller said engineers at the Traffic Signal Test Center in Phoenix have worked to perfect the predictive priority system since 2005. The system is 50 percent complete and was tested on live trains in March 2006.
The predictive technology gives light-rail traffic an edge, but not an automatic green light, because this would halt normal traffic flow, he said. Also, added side-street gridlock from more red lights would be counterproductive by encumbering the drive to the light-rail train station.
The system is built with a gigabyte Ethernet network running along the 20-mile corridor, and will allow quick, clear communications to the on-site traffic controllers.
Intersections will be equipped with check-in and check-out detectors that trip when the train speeds over them. The detectors will broadcast the rail's real-time position to upcoming intersections, giving them several minutes to prepare, Fuller said. Then, sophisticated algorithms analyze the time it will take the train to reach each station and decide where the traffic light will be in its cycle upon the train's arrival.
"The theory is we can get far enough ahead of the train in arrival time, that that's enough to facilitate assuring the train a green on arrival," Fuller said.
At its core, the predictive priority system allows for coordination between adjacent traffic light signals - whereas most traffic lights operate independently, said Larry Head, interim department head of Systems and Industrial Engineering in the Engineering College at the University of Arizona.
System communication is organized in groups of five or six intersections that talk to four midlevel switches, located five miles apart, Fuller said. Then, a network backbone collects and disperses the information to three traffic management centers, which provide central control, observation and dissemination of