Creating the Smart Chevy

Casey Dexter Arizona Summer Wildcat At a lab at the University of Arizona (top), a radar system attached to the front of a car keeps a computerized driver on a safe course. The car of the future is a 1989 Chevrolet Celebrity that the engineers bought for $1.

At a University of Arizona laboratory, professors and graduate students are developing technology that they hope will clear up urban congestion.

UA engineers at the Advanced Traffic and Logistics Algorithms and Systems lab are working on projects ranging from more efficient traffic lights to self-driving automobiles.

Pitu Mirchandani, Systems and Industrial Engineering department head, Feiyue Wang, associate professor of systems and industrial engineering, and several graduate students have built a prototype of a self-driving vehicle using a system they call "adaptive cruise control."

Working with the Arizona Department of Transportation, they have equipped a 1989 Chevrolet Celebrity with a computerized radar and camera system that follows the lane markers along a programmed route.

"That's a problem, we have to have a continuous bright lane marker," Mirchandani said.

The radar system measures the distance from other vehicles and tells the computer, which modifies the speed as needed. The driver punches in a destination and a computerized "map" tells the car where to go. Reflectors by the side of the road will tell the car exactly where it is.

Pitu Mirchandani, UA Systems and Industrial Engineering department head, explains the programmable traffic sensors which will soon be tested in the university area. The sensors, which will be set up at intersections, will measure the volume of traffic and time the lights accordingly.

Richard Nassi, city traffic engineer for the city of Tucson, said that the system will be first used by the trucking industry on major highways. Trucks would travel in a convoy while their drivers relax. When one driver needs to leave the highway, he would switch to manual control.

"It looks and acts like a mass transit system, but when they want to get off the train, they can," he said.

The lab will be working with the city of Tucson to test another project which will use a computer program to study traffic flow at intersections and change the lights as needed.

Researchers hope it will end the frustration of waiting with 10 other cars at a red light while the green light shines on an empty road.

Called RHODES -- Real-time Hierarchical Optimized Distributed Effective System -- the program will analyze data from video cameras and radar detectors to determine the number of cars approaching an intersection.

The program will then calculate when to change the lights so that the most number of cars will be able to move at one time.

"This is going to be phenomenal," Nassi said. "The improvement (in speed) could be as good as an extra lane."

Engineers will test the system in the UA area sometime soon, but did not want to reveal exactly where or when the test will take place.

"A lot of people would shy away...if they know a test is underway," said David Lucas, a systems and industrial engineering graduate student.

"We want an unbiased sample," Mirchandani added.

Mirchandani estimated that after the system has been perfected, it will cost about $3,000 to outfit each intersection with sensors and computer equipment.

Researchers at the lab are developing a similar system to regulate traffic flow onto highways. On current highways, a driver entering at the wrong time can disrupt the flow of traffic, forcing other vehicles to slow down, Mirchandani said.

Lights at on-ramps will allow cars on the highway with as little disruption to traffic as possible, making for a smoother and faster ride to the drivers' destinations.

"The time from on-ramp to off-ramp will be significantly less," Mirchandani said. "Instead of 10 minutes it will be six or seven minutes."

Engineers will test the system in about six months on a 10-mile section of highway near Pheonix.