PITTSBURGH--Researchers at Carnegie Mellon University's Robotics Institute have automated two Houston Metro Transit Authority buses, a pair of Pontiac Bonnevilles and an Oldsmobile minivan for "AHS O97," an automated highway system (AHS) prototype demonstration taking place Aug. 7-10 on Interstate 15 near San Diego.
Carnegie Mellon is one of nine core members of the National Automated Highway System Consortium (NAHSC), established in 1994 with more than $200 million from the U.S. Department of Transportation and other sponsors to help solve congestion, safety and productivity problems on the nation's roads and highways.
The buses, cars and van--known as Navlabs six through 10--comprise one of seven teams from academic, industrial and government organizations that will demonstrate various aspects of highway and vehicle automation at the AHS demonstration.
Carnegie Mellon's autonomous vehicle technology, which has been under development for more than a decade, drives on existing roadways with no special modifications. During the demonstration, the vehicles will exhibit lanekeeping, speed and headway maintenance, lane departure, obstacle detection and avoidance, and lane changing and passing. In some instances they will tell other vehicles what lies ahead on the road.
To perform these tasks, the vehicles have been fitted with PCs hidden in their trunks. A camera on the rear view mirror "sees" the road ahead and turns the steering wheel with software known as "Ralph," which stands for Rapidly Adapting Lateral Position Handler. Researchers also have automated the vehicles' brakes and throttles, as well as the steering capabilities.
For headway maintenance and obstacle detection, the vehicles are fitted with radar systems that look forward, sideways and backward. "We use a radar system to sense vehicles in front of us and adjust our speed using the throttle and brake to keep us a set distance behind," explained Navlab team member Todd Jochem. "There is a rear looking laser to detect approaching cars, as well as four side looking radars to check our blind spots before we change lanes."
In addition, the vehicles are equipped with a laser range finder installed in back bumpers to warn of overtaking vehicles. The vehicles also are equipped with five antennae for communicating with other vehicles and a traffic management center.
"Everything is put together from things that are already built," said project manager Chuck Thorpe. "Everything on the demo is running on a PC that's less powerful than what people are buying for their desktops."
Researcher Dean Pomerleau noted that the Navlab researchers have logged more than 30,000 miles in automated vehicles--13,000 since February of this year. He said 16 versions of the Navlab vision system are being used in Europe, Japan and North America. Two copies are mounted on heavy trucks as a research tool.
"A lot of this technology has become commercially available," said Jochem. "Many of these systems can be adapted easily to a standard car. In the near term, before completely automated driving is possible, parts of this technology will be used for driver assistance. For example, the Ralph vision system can be used to watch the road while a person drives. If the driver becomes sleepy and starts to drift off the road, Ralph notices and can alert the driver."
Two Carnegie Mellon spin-off companies are helping to commercialize the technology used in the demonstration vehicles. Assistware Technologies, Inc., (Pittsburgh) produces Ralph vision software and hardware. K2T, Inc., provided the vehicle automation systems built into the two buses used for the demonstration.
At Carnegie Mellon, research into autonomous vehicles began in 1985 and has received continuous support from the Defense Advanced Research Projects Agency. More recently, additional funding has come from the U.S. Department of Transportation. The technology has advanced from an oversized blue Chevrolet van stuffed with a million dollars worth of computers to standard vehicles fitted with off-the-shelf sensors and computers that are less powerful than the standard equipment on today's desktop.
For more Information on Carnegie Mellon's autonomous vehicles check the Navlab Website at: http://www.cs.cmu.edu/afs/cs/project/alv/member/www/navlab_home_page.html
97-52