Robots Come Up Short in the Grand Challenge

IN THE MOJAVE DESERT OUTSIDE BARSTOW, CALIF., MARCH 13, 3:20 A.M. "I¿ve got the CD," Alex Gutierrez says into his radio as an official from DARPA (the Defense Advanced Research Projects Agency) hands him a CD-ROM containing the route for the first cross-country robot race in history. Gutierrez takes off at a sprint for a mobile office 200 yards away. He bursts through the door less than a minute later. The CD is popped into a computer, which immediately begins running sophisticated path planning software to plot a safe passage over the 229-kilometer (143-mile) course. Gutierrez struts down the center aisle of the trailer, patting the backs of 14 "route editors" from Carnegie Mellon¿s Red Team sitting at workstations on either side.

Within minutes, the computer has finished its processing, and the route editors begin poring over every meter of the course, looking at satellite images, aerial photographs, 3-D terrain models, and topographic maps to set optimal speed and directions for their robot, an autonomous Humvee called Sandstorm, to follow on its trip over the mountains to Primm, Nev. None of the 14 other robots competing today in DARPA¿s Grand Challenge race will carry such a detailed and carefully prepared set of instructions.

6:30 A.M. Sandstorm idles at pole position in the starting chute. Next to it is a robotic desert buggy built by engineers from Elbit Operations, an Israeli defense contractor. In the far left starting chute is Bob, a self-driving SUV constructed by Caltech undergraduates. Of the 21 robots that showed up at California Speedway on March 8 to qualify for the Grand Challenge, these three ran the short obstacle course fastest and with the fewest mistakes.

Right on cue Sandstorm takes off, quickly accelerating to 22 miles an hour as it rounds the first curve. Nine minutes later, it has come to the first road crossing. Using its long-range laser scanner, it detects the gate through which it must pass and adjusts its course, but not enough to avoid plowing down fence posts on both sides of the road. The Elbit vehicle passes cleanly through the gate five minutes later, but Bob takes out the other side of the fence and gets confused by the accident. That will be the end of the race for Caltech.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Meanwhile one of the dark horses in the race is closing in on the two leaders. Professional roboticists observing the competitors had doubted that DAD, a pickup modified by David and Bruce Hall at Velodyne, would get far at all. Its only means of sensing terrain ahead is a pair of color video cameras. Because cameras produce so much data, they are notoriously difficult to use for high-speed navigation. The Hall brothers optimized their software to run at blistering speed on digital signal processors, and, at least on the obstacle course, it worked surprisingly well.

As Sandstorm rounded the switchbacks on a rough mountainous trail near Daggett, the robot began hugging the left side of its route. Just under 12 kilometers into the race, the left two wheels went over the gravel berm on the shoulder. The berm grew higher, and at last the vehicle was left stranded on its belly, its wheels spinning frantically, churning up gravel, melting their rubber treads but failing to budge another inch.

Just around the same time, Elbit¿s buggy, which had also been tracking toward the outside of its programmed path, wedged itself into the face of the mountain. It was not far behind Sandstorm, and as the two teams spoke afterward they suspected both robots fell victim to a common problem with how they were relying on navigational information from the Global Positioning System. Indeed, at that point in the course, the rules of the race left them little option but to do so.

Of the 13 robots that attempted the race (two teams withdrew just before the start), the DAD pickup made it third farthest. "It came to a fork in the road," says Bruce Hall. "It had to choose the left or the right path, but instead it chose to split the difference and headed for the middle. A football-sized rock got stuck under the front wheel, and it didn¿t know to hit the gas hard in order to get over it."

Tony Tether, director of DARPA, pronounced the race a great success despite the fact that no robot was able to cross the mountains to the flatter, straighter and thus easier part of the route to Primm. "We will be holding the Grand Challenge again around this time next year," he said at a finish line reception, "and next year, the prize will be $2 million."

It’s Time to Stand Up for Science

If you enjoyed this article, I’d like to ask for your support. Scientific American has served as an advocate for science and industry for 180 years, and right now may be the most critical moment in that two-century history.

I’ve been a Scientific American subscriber since I was 12 years old, and it helped shape the way I look at the world. SciAm always educates and delights me, and inspires a sense of awe for our vast, beautiful universe. I hope it does that for you, too.

If you subscribe to Scientific American, you help ensure that our coverage is centered on meaningful research and discovery; that we have the resources to report on the decisions that threaten labs across the U.S.; and that we support both budding and working scientists at a time when the value of science itself too often goes unrecognized.

In return, you get essential news, captivating podcasts, brilliant infographics, can't-miss newsletters, must-watch videos, challenging games, and the science world's best writing and reporting. You can even gift someone a subscription.

There has never been a more important time for us to stand up and show why science matters. I hope you’ll support us in that mission.

Thank you,

David M. Ewalt, Editor in Chief, Scientific American

Subscribe