CMU wins $2 million in urban robot race

Carnegie Mellon University won the $2 million first place prize in DARPA's urban robot race this weekend, stealing the thunder from 2005's Grand Challenge leader, Stanford University.

The Defense Advanced Research Projects Agency (DARPA) Urban Challenge awarded a total of $3.5 million in prizes on Sunday, a day after the race. Stanford University took second place, with a $1 million cash prize, and Virginia Tech won $500,000 for third place.

CMU's robot, Boss, finished the Urban Grand Challenge about 20 minutes faster than Stanford's Junior to win the race.

(Credit: Stefanie Olsen/CNET News.com)

The Urban Challenge was a six-hour test of driverless vehicles on the suburban roads of the former George Air Force Base in Oro Grande, Calif., where the robotic cars were required to complete three missions while obeying traffic laws and avoiding obstacles and collisions with other driverless vehicles. The challenge was the first ever to test robots driving among other robots, and it was significantly harder than DARPA's 2005 desert Grand Challenge because of that interplay and the urban setting, according to race officials.

Despite skepticism about how the robots would perform, 6 of the 11 finalists-- including teams from MIT, Cornell and the University of Pennsylvania--crossed the finish line. Yet CMU, Stanford and Virginia Tech displayed the most competent driving skills and finished in the allotted six hours, so the race came down to speed, according to DARPA.

CMU's Boss, a modified 2007 Chevy Tahoe, finished the course about 20 minutes faster than Stanford's Junior--an autonomous driving VW Passat. Boss averaged about 14 miles per hour over about 55 miles. It crossed the finish line only minutes after Junior, but Boss started at least 20 minutes later in the challenge. Virginia Tech crossed the finish line third.

Red Whittaker, CMU's team leader, said that the win is all about inspiring people and changing beliefs. "Once the perception of what's possible changes it never goes back. This is a phenomenal thing for robotics."