IE 11 is not supported. For an optimal experience visit our site on another browser.

Robotic race gets off to rocky start

Twenty robo-racing teams came together on the same track Monday to prepare for the Pentagon’s $1 million road race — a competition where even getting past the starting line was something to cheer about.
Denny Gudea uses a laptop to drive "Terrahawk" on the first day of testing and qualifying for the DARPA Grand Challenge.
Denny Gudea uses a laptop to drive "Terrahawk" on the first day of testing and qualifying for the DARPA Grand Challenge.Jim Seida / MSNBC.com

Twenty robo-racing teams came together on the same track for the first time Monday to prepare for the Pentagon’s $1 million road race — a competition where even getting past the starting line was something to cheer about.

The Defense Advanced Research Projects Agency launched its Grand Challenge more than a year ago, inviting all comers to develop autonomous vehicles that could navigate 210 miles through the Mojave Desert in 10 hours without human intervention. Monday marked the start of qualifying trials here at the California Speedway, leading up to the race itself on Saturday.

DARPA Director Anthony Tether told journalists that the task was similar to programming your car to drive itself to the grocery store. But blending electronic hardware, software and off-road muscle is clearly going to be harder than most of the teams may have thought.

Not one of the eight vehicles going through Monday’s trials successfully completed the Speedway’s 1.25-mile obstacle course. Most of them never even got started. But cheers went up when Team Caltech’s entry, a converted Chevy Tahoe SUV nicknamed “Bob,” made it out of the starting chute and began picking its way over concrete slabs and down a grassy knoll.

“C’mon boy, get through!” one of the team members yelled as they ran along the grandstands. The driverless Bob successfully eased its way through a narrow underpass, three-quarters through the course — but then it tailed off to the right, wheels squeaking on the concrete barriers. The SUV was stopped just before it would have run into a wall.

March 8, 2004 - California Speedway, Fontana, California - Team Caltech members Jeff Lamb (left), Jackie Wilbur and Haomaio Huang celebrate as \"Bob\", their autonomous 1996 Chevy Tahoe SUV leaves the starting line during qualifying trials for the DARPA Grand Challenge.
March 8, 2004 - California Speedway, Fontana, California - Team Caltech members Jeff Lamb (left), Jackie Wilbur and Haomaio Huang celebrate as \"Bob\", their autonomous 1996 Chevy Tahoe SUV leaves the starting line during qualifying trials for the DARPA Grand Challenge.Jim Seida

During a postmortem, team members determined that Bob lost its GPS navigation signal when it went through the underpass — and that the equipment designed to make up for the lost signal wasn’t working. They quickly set to work on a fix, and despite the setback, Team Caltech’s Lars Cremean said he and his teammates were “absolutely” pleased with Bob’s initial showing.

Another robo-car — a modified Honda Acura SUV called “Doom Buggy” that was entered by students from Palos Verdes High School in California — also made it out of the chute but veered into a barrier near the grandstands.

At least those two cars made a run at the course Monday — which is more than could be said for Team Phantasm, which is fielding a modified Kawasaki all-terrain vehicle nicknamed Ladibug. The contraption had to sit out its initial qualifying round.

“Our radar is on the fritz,” Phantasm team leader Warren Williams explained.

Team Phantasm and the others will get several chances to qualify this week. And 13 of the 21 official entrants still have to face their first DARPA trial. (One team didn't even arrive in time for Monday's trials.)

The first day of trials showed “just how difficult this technology is,” said Thomas Strat, deputy program manager for the DARPA Grand Challenge, but it also showed that the challenge was “doable.”

Actually completing the Speedway ’s obstacle course was not a “hard prerequisite” for moving on to Saturday’s finals, Strat said. The teams needed only to prove that their vehicles could be operated and stopped safely — and that they had a chance of running Saturday’s race from Barstow , Calif. , to Primm , Nev.

Competitors won’t be given the actual route until two hours before the race begins. If none of the vehicles finish the race in the prescribed 10 hours, DARPA’s $1 million prize will roll over to the next robotic challenge, probably in 2006.

March 8, 2004 - California Speedway, Fontana, California - Red Whittaker sits on \"Sandstorm\", a heavily-modified HMMWV before qualifying trials for the DARPA Grand Challenge.
March 8, 2004 - California Speedway, Fontana, California - Red Whittaker sits on \"Sandstorm\", a heavily-modified HMMWV before qualifying trials for the DARPA Grand Challenge.Jim Seida

The current favorite is Carnegie Mellon University’s Sandstorm, a converted military Humvee bristling with a stereo camera system, laser range-finders, GPS receivers, actuators and three flavors of computers.

Even Sandstorm has suffered setbacks: Just a couple of days ago, the vehicle flipped when it took a sharp turn too fast during a practice outing at the Nevada Automotive Test Center.

“We were kind of pushing the envelope when we were testing,” said Aaron Mosher, a member of Carnegie Mellon’s Red Team. But after some round-the-clock repairs, Sandstorm was rolling again on Monday.

“You never underestimate this challenge,” said Red Whittaker, a robotics professor at Carnegie Mellon University as well as an ex-Marine. He said he had no doubt that his students would get Sandstorm going again.

“If you get doubt, there’s a reason why you’ll never get it back together,” he told MSNBC.com. “Leaders don’t have doubt.”

Contest 'about saving lives'
Another entrant, Anthony Levandowski of the Berkeley, Calif.-based Blue Team, freely admitted he had no chance of winning this year with Ghostrider Robot, his gyro-stabilized motorcycle. If Ghostrider falls over, it can’t get up, at least not until its auto-rectifying system is perfected. More importantly, Ghostrider’s fuel range is only about 190 miles — which would appear to fall short of the finish line.

March 8, 2004 - California Speedway, Fontana, California - Blue Team leader Anthony Levandowski with \"Dexterit\", the world's first autonomous single-track vehicle. during qualifying trials for the DARPA Grand Challenge.
March 8, 2004 - California Speedway, Fontana, California - Blue Team leader Anthony Levandowski with \"Dexterit\", the world's first autonomous single-track vehicle. during qualifying trials for the DARPA Grand Challenge.Jim Seida

Rather than winning this year, Levandowski expects his research to spawn next-generation riderless cycles in the next five years. Such contraptions could serve as military reconnaissance robots or even Hollywood stunt bikes that don’t require stunt men, he said.

“The purpose of our vehicle is not to compete in the Grand Challenge,” he said. “The Grand Challenge is a field test for us. ... We don’t expect that this year there will be a winner.”

Even if the autonomous vehicles aren’t quite ready for the million  dollars this year, DARPA’s Tether said the exercise was worth it. He estimated that the agency has spent $13 million on the project, and said “the return on the investment far exceeds that amount” — perhaps by a factor of four or five.

The innovations developed for the Grand Challenge could well find their way over the next decade into robotic vehicles that could carry supplies to the front lines without endangering soldiers’ lives, Tether said. If such a vehicle were available during the war on Iraq, he said, it might well have taken on the kind of trip that led to the deaths of 11 U.S. soldiers as well as the capture of Pvt. Jessica Lynch and four other Americans.

“The DARPA Grand Challenge is not about money. It’s not about fame. It’s not even about winning. It’s all about saving lives,” Tether said. “I have no idea how this challenge is going to turn out, but the outcome is less important than the event itself.”