DARPA brings robotic race cars to city streets for US$2M prize

Military hopes that technology used in driverless cars can save lives on the battlefield

The US government is hoping that a 60-mile race among up to 20 driverless, self-guided vehicles will yield technological advances that can save lives on the battlefield.

The US government's military research arm, the Defense Advanced Research Projects Agency (DARPA), this week is gearing up for the Urban Challenge, to be held Saturday, November 3 at the abandoned George Air Force Base in California.

The race pits teams of researchers -- academics from the likes of MIT and CalTech, along with a few diehard hobbyists -- against one another, as they match up their creativity and technical savvy.

The goal? The vehicles must accurately navigate a difficult 60-mile course without human assistance. No one is in the car to turn the wheel, apply the brakes or figure out which way to go. The automobile must basically think its way through the course and do it all on its own -- from start to finish. DARPA is awarding prizes to the teams whose vehicles complete the course in less than six hours. The first prize is US$2 million, second prize is US$500,000 and the third prize is US$250,000, DARPA said.

About 35 teams are taking part in qualifying trials this week. The field is expected to be narrowed to 20 or fewer when the race begins at 7 a.m. Eastern time on Saturday. The Tartan Racing Team out of Pittsburgh is the only team that has moved past the qualifying round and into the finals so far.

"This is enormously helpful to us," said Jan Walker, a DARPA spokeswoman. "The automotive technology coming out of this will allow us to deploy [driverless] vehicles that now are being driven by human drivers who are at risk." Self-driven vehicles could be used for bomb-detonation tasks, in military convoys and for military reconnaissance missions, she added.

The Urban Challenge is DARPA's third Grand Challenge race. The first two were set on desert courses that spanned nearly 150 miles from California to Nevada. None of the teams participating in the first race, held in 2004, finished the course. The top team in that race only covered five miles of the 142-mile course. The next year, four vehicles finished a 132-mile course in less than 10 hours. The winning team, from Stanford University, took home a US$2 million prize.

This year, DARPA's planning team is less focused on the relatively straight and obstacle-free desert course and is looking to make things more interesting -- and more difficult.

Walker said the Urban Challenge course forces the cars to navigate through four-way stops and around buildings and through narrow streets. The vehicles also must merge into moving traffic, navigate traffic circles and obey driving laws.

The automobiles are equipped with GPS navigation systems, radar and even lidar, which is similar to radar but uses light lasers instead of radio waves. Walker said in the first year of the Grand Challenge race series, the vehicles had equipment and wires hanging all over them. This year, though, the technology is much more embedded within the cars, so they look far more like ordinary vehicles.

"This has demonstrated the art of the possibility," said Walker. "We've shown that it's possible for ground vehicles to travel long distances at relevant speeds. Once the government sees this is possible, they will be more willing to delve into this. It allows the... military services to start imagining how to use this technology. We're pushing the technology forward and involving a wide array of people."

Join the newsletter!

Or

Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

More about Defense Advanced Research Projects AgencyMITStanford University

Show Comments
[]