It's almost time for the next DARPA "Grand Challenge" for autonomous vehicles. (We've written about previous events here and here.) This time, they'll be tested in a simulated urban setting:
The DARPA Urban Challenge is an autonomous vehicle research and development program with the goal of developing technology that will keep warfighters off the battlefield and out of harm’s way. The Urban Challenge features autonomous ground vehicles maneuvering in a mock city environment, executing simulated military supply missions while merging into moving traffic, navigating traffic circles, negotiating busy intersections, and avoiding obstacles.
In a story previewing the event, New York Times reporter John Markoff writes:
During the first two races the robots were basically very smart crumb followers. Just before the start of the race they were given a list of G.P.S. way points and simply had to stay within the road boundaries on the outlined course.But in the Urban Grand Challenge the cars will be given a set of missions to complete in a simulated urban setting (a military training city). They will have to figure out how to get from point A to B and then C, etc.
Better still, other robot competitors will be busily attending to their own competing missions at the same time.
Markoff took a ride in "Junior," one of this year's entrants, and compared it with "Stanley," a previous competitor:
Because it was in a parking lot, this year’s ride was in some ways less dramatic than my first tour. But in other ways it was more striking, and the pathfinding abilities of the car even had a spooky quality.Whereas Stanley was only able to follow its route, Junior actually “thinks” about the route and can choose alternatives. It can pass other cars. It can wait for traffic and it can even make decisions that “cheat” on the rules a bit.
At one point, for example, when both lanes of our course were blocked by cars, Junior made a three-point turn and got to its destination by a secondary route.
Very cool.
UPDATE - We have a winner!
November 4, 2007 -- The slow-motion finale of the 2007 DARPA Urban Challenge is at an end. After spending the night tabulating the run-times for the six robots that finished the 60-mile course yesterday, and weighing their overall speed against potential penalties for moving violations, the officials at DARPA have come up with a winner. Tartan Racing, the joint effort from Carnegie Mellon University and General Motors, beat out arch-rival Stanford University for the $2 million first prize. Stanford took second place, and Virginia Tech's Victor Tango team took third, winning $1 million and $500,000 respectively.Because of the secretive and unabashedly subjective nature of DARPA's decision-making process, we may never know exactly why Boss won, or how close Stanford's Junior did or didn't come to beating Tartan. Citing the fact that this contest was completely unprecedented, DARPA avoided posting a list of penalties, and reserved the right to adjust the weight of those penalties after the race. But according to Tartan leader William "Red" Whittaker, they won because of planning. "Boss didn't hesitate," Whittaker said. "When it came to planning—planning at intersections, planning on straightaways, planning throughout the course—we were unbeatable." By taking turns without an extra pause to get its bearings, and knowing when the coast was clear enough to gun the engine up to 30 mph (the max speed for the course), Boss may have shaved minutes off its run-time.

Tags: nanotechnology nanotech nano science technology ethics weblog blog
Comments