A driverless red Hummer that was involved in a rollover accident during practice a few weeks ago snagged the pole position Wednesday in a government-sponsored sequel race across the Mojave Desert that will pit 23 robots against one another.
The finalists were chosen after an intense, weeklong qualifying run at the California Speedway, where the self-navigating vehicles had to drive on a bumpy road, zip through a tunnel and avoid obstacles. No human drivers or remote controls were allowed.
The Hummer named H1ghlander, built by Carnegie Mellon University, flipped during practice a few weeks ago when it struck a rock. But it still managed to complete all four required semifinal runs.
Last year, only half of the 15 autonomous robotic vehicles that ran in the so-called Grand Challenge passed the semifinals. No team claimed the $1 million inaugural prize because all the contestants broke down within a few miles of the starting gate.
So this year, the sponsor, the Defense Advanced Research Projects Agency, doubled the purse with the hope that a vehicle would finish Saturday's $2 million race.
This year's finalists completed the hilly qualifying course littered with hay bales and parked cars at least once. Five of the vehicles finished it four consecutive times. Those included H1ghlander; a converted Humvee Sandstorm; a modified Volkswagen Touareg by Stanford University; a six-wheel truck; and a Jeep Grand Cherokee.
"I'm inspired by all the robots," said William "Red" Whittaker, a Carnegie Mellon robotics professor. "Never discount or diminish any of them."
The race is part of the Pentagon's effort to fulfill a congressional mandate to have a third of all military ground vehicles unmanned by 2015. The Defense Department envisions using robotic vehicles to bring supplies in combat zones.
DARPA, the Pentagon's research and development arm, spent $9 million on this year's event. The agency would award the prize to the first team whose computer-driven vehicle can traverse a rough and winding desert course of up to 175 miles in less than 10 hours.
There are several reasons why this year's field is more competitive. Teams had more time to prepare for the race. Many spent the past 18 months focused on the vehicles' computer "brain," beefing up their artificial intelligence through improved computer algorithms.
Teams also had the advantage of practicing in various parts of the Southwest desert under race-like conditions. Even before the semifinals, some robots had already driven hundreds of continuous miles during practice, including some that tested on last year's course between Barstow and Primm, Nev.
"Nobody was ready last year," said Bill Kehaly of Westlake Village-based Axion Racing, whose Jeep Grand Cherokee named Spirit is racing in the finals. "Everybody feels a lot more confident this year."
Because the vehicles must be self-navigating, they are equipped with GPS tracking. Mounted sensors, radar, lasers and cameras feed information to onboard computers to orient the vehicles and help them avoid obstacles and traps.
The exact route that vehicles must follow is kept secret until two hours before the competition. Organizers said the course, which will loop from and to the casino town of Primm will feature rugged desert and mountain terrain. Vehicles have to average 15 mph to 20 mph to finish in time.
"The worst vehicle we have is as good or better than the best vehicle last year," said DARPA director Anthony Tether.
Carnegie Mellon's workhorse, Sandstorm, traveled the farthest in the Mojave Desert last year despite trekking only 7 1/2 miles. It will start third in this year's race.
Of the vehicles that successfully coursed the speedway four straight times, an off-road, six-wheel truck built by Oshkosh Truck Corp., had the slowest time. But team leader Jim Fravert did not think that was a problem. The desert will be tougher and the truck was designed to handle 60 percent grades and push through five feet of water, he said.