Slashdot Mirror
DARPA Challenge Prize Money Restored
antispam_ben wrote to mention that, some three months later, DARPA has been able to find the money to offer cash prizes once again. The DARPA Urban Challenge will go forward next November with more than $3 Million on the line. From the article: "The race will see as many as 90 teams 'drive' an unmanned robotic road vehicle through city traffic, competing to finish a 60-mile course within six hours. Set for November 3 of next year, the challenge will call on robots to safely obey traffic laws, negotiate busy intersections, merge into moving traffic, avoid obstacles and navigate traffic circles. DARPA has yet to disclose the race location, but has said it will be in the western United States. The government research group didn't unveil the 2005 Grand Challenge location in the Mojave Desert until weeks before that race, in order to avoid giving any team an advantage."
To prefix my comment, let me just say that I am currently on the Georgia Tech team, Sting Racing. According to the rules, we will be on the course simultaneously with the other cars. The other other vehicles allowed on the course will be professional drivers. AFAIK, this is not being done in an actual city, but a small-ish mockup is being constructed for the purposes of the event. I could be wrong, though.
No comment.
I've posted elsewhere in this story, but again to prefix my comments, I'm a member of the Georgia Tech team, Sting Racing.
The course plotting part of the challenge is actually probably the easiest part. It's roughly analogous to you reading a map beforehand -- we're given a file detailing all the aspects of the course (course segments, how many lanes are in the segments, etc., and zones where free driving is allowed) plus a mission file giving the different waypoints we have to reach. This is, relatively speaking, easy.
The difficult part is determining where the edges of the road and its lanes are (GPS is terrible at this; most of the time it's accurate to 10 or so feet unless you're using extraordinarily expensive differential units) which is mostly done using visual scanning. Of course, we also have to detect other vehicles or obstacles in the path (using LIDAR and vision) and also determine the correct "pose" of the vehicle. Then we have to take that information and use it to modify the path we've already decided to take. These problems as it turns out are far, far harder than just plotting courses.
No comment.