Carnegie Mellon's robocar navigates big win

A robot racing team from Carnegie Mellon University beat out a rival from Stanford University over the weekend to win DARPA's Urban Challenge, a 60-mile race involving self-guided vehicles that were judged on both time and how well they performed.

The Carnegie Mellon team, known as Tartan Racing, took home the US$2 million first prize, while the Stanford Racing Team grabbed a US$1 million check for finishing second. Team Victor Tango, which was made up of faculty and students from Virginia Tech, received US$500,000 for taking third place.

In all, 11 so-called autonomous vehicles raced at the abandoned George Air Force Base in the finals of the Urban Challenge, which pit the driverless cars against one another on a course that challenged the vehicles and their self-guidance systems to find their way through 60 miles of urban streets with multiple lanes, traffic circles and four-way stops.

It was the third race of its kind held by the Defense Advanced Research Projects Agency, but the first to involve an urban-style course.

"The urban setting added considerable complexity to the conditions faced by the vehicles, and was significantly more difficult than the fixed desert courses featured in the first two Grand Challenges," Norman Whitaker, DARPA's Urban Challenge program manager, said in a statement sent via e-mail.

"Tartan Racing, Stanford Racing and Victor Tango all did a great job getting their vehicles to navigate the course quickly and safely despite the challenging conditions."

The Urban Challenge called on the teams of researchers -- academics from schools that also included MIT and Cornell University, along with hobbyists and engineering professionals -- to match up their creativity and technical savvy.

DARPA sponsored the event and offered the hefty cash prizes in an effort to encourage researchers to develop technology for autonomous ground vehicles that could help save lives on the battlefield, according to Jan Walker, a spokeswoman for the agency. She said that self-driven vehicles could be used in military convoys and for bomb-detonation tasks or military reconnaissance missions.

The Tartan Racing team reported on its Web site that this weekend's finish was nearly a case of deja vu from the 2005 Grand Challenge, in which Stanford Racing finished first and the Carnegie Mellon team second.

During the Urban Challenge, Stanford's car, called Junior, again crossed the finish line ahead of Tartan's vehicle, known as Boss. But this time, the final outcome was different. As part of a staggered start, Boss began more than 20 minutes after Junior did but was only a couple of minutes behind at the finish, according to Tartan Racing.

Team Victor Tango's namesake car, which was the first vehicle to start the race, was third across the finish line, about six minutes behind Boss.

DARPA officials factored in how well the vehicles did in navigating the course, along with their times and any delays they encountered, to rank the teams.

The robot-powered cars had to accurately navigate the complicated course in less than six hours without human assistance while sharing the road with about 50 human-driven vehicles. No one was in the cars to turn the wheel, apply the brakes or figure out which way to go. A group of 35 semifinalists was winnowed down to the 11 teams that competed in the finals.