Carnegie Mellon's Sandstorm--an artificially intelligent, robotic car with drive-by-wire modifications, GPS (Global Positioning System) and radar sensors, and remote control emergency safeguards--drove 200 miles, or 131 laps, autonomously last week. That's a milestone for the university's Red Team in the race to out-drive 39 other robotic vehicles and win the 2005 DARPA Grand Challenge, a desert race with a $2 million prize sponsored by the Defense Advanced Research Projects Agency, the research arm of the U.S Department of Defense.
"That doesn't sound like a big deal for a human-driven car, but it is a very big deal for the pioneering of computer-driven vehicles," Red Team leader and Carnegie Mellon robotics professor Red Whittaker said Wednesday in a statement.
It certainly surpasses Sandstorm's performance last year during theDARPA race. Sandstorm lost the challenge driving the fastest and farthest, making it only 7.3 miles on a 142-mile course before spinning its wheels. No team claimed the prize.
The 2005 contest, to be held Oct. 8 in a location yet to be disclosed in the Mojave Desert, is sure to prove difficult, despite advances from robotic leader Carnegie Mellon and other new contenders this year including Stanford University's Racing Team. DARPA recently told contestants that this year it would include artificial obstacles on the longer, 175-mile course.
Whittaker said that the recent test, at Pittsburgh's BeaveRun MotorSports Complex, helped the team learn that its hardware and software are reliable. The vehicle's exterior sensors help the car see the road and avoid obstacles, while the car's computers command the driving. During the test, the robot averaged 28 mph, with a top speed of 36 mph.
"Sandstorm ran a quick pace on this track, but the Mojave will not be so easy or forgiving. To finish first, you must first finish," Whittaker said.
Carnegie will also host another vehicle, the Hummer H1ghlander, in the semifinals, which take place Sept. 26 at the California Speedway at Fontana.