SMS
RSS feeds
Reddit
NewsvineThe Stanford Racing Team and its robotic car “Junior” begin competition today at the National Qualifying Event for the Defense Advanced Research Projects Agency (DARPA) Urban Challenge.
Enlarge
#gallery http://daily.stanford.edu/image/full/8071
“Junior,” a 2006 Volkswagen Passat wagon, enters the National Qualifying Event for the Defense Advanced Research Projects Agency (DARPA) Urban Challenge today. The Stanford Racing Team hopes that the autonomous car will advance to the Urban Challenge on Nov. 3.
Currently at the former George Air Force Base in Victorville, Calif., the Stanford team hopes to advance from the National Qualifying Event this last week of October to the Urban Challenge race on Nov. 3.
At the center of their work is Junior, a 2006 autonomous Passat wagon whose steering, throttle and brakes have all been modified to be completely computer controllable. Engineered at the Volkswagen of America Electronics Research Laboratory in Palo Alto, Junior is being accompanied by a replacement car on the team’s trek.
The upcoming Urban Challenge is more complex than previous robotic car races. The team’s previous car, “Stanley,” simply had to avoid hitting still objects and stay on the road — for which it won the 2005 DARPA Grand Challenge.
This year, however, Junior will have to be aware of fast-moving objects all around it, navigating through mock city environments that include busy intersections, traffic circles and other obstacles.
To tackle the challenge, Junior comes equipped with a range-finding laser array that spins to provide a 360-degree, three-dimensional view of the surrounding environment in near real-time. The laser array is accompanied by a device with six video cameras to cover all of the car’s surroundings.
Junior also uses bumper-mounted lasers, radar, Global Positioning System receivers and inertial navigation hardware to collect data about where it is and what is around it. Junior’s “brain” is composed of chips that have multiple processing units, making it approximately four times more powerful than Stanley’s “brain.”
“Junior collects much more data than Stanley did; its computational hardware must be commensurately more powerful,” said Mike Montemerlo, a senior research engineer in the Stanford Artificial Intelligence Laboratory.
The upcoming challenge is an outgrowth of two previous DARPA Grand Challenge autonomous vehicle competitions. The first Grand Challenge event was held in March 2004 and featured a 142-mile desert course. Fifteen vehicles attempted the course and not a single vehicle finished.
In the 2005 Grand Challenge, several vehicles were able to complete a 132-mile desert route under the required 10-hour limit. DARPA awarded the $2 million prize to Stanley, which has since been retired. Another $2 million will be awarded to the the fastest qualifying vehicle this year, with $1 million and $500,000 for second and third places, respectively.
“The only team that will be able to make up their costs will probably be the winning team,” said David Orenstein, public relations and communications manager for Stanford School of Engineering. “The fact is, we are not in it for the money.”
For team leader Sebastian Thrun, the coming day’s competition is yet another test of years of hard work put into the project.
“The upcoming race will be a lot of fun,” said Thrun, a professor in electrical engineering. “This type of race has never happened before. We have made a lot of changes to Junior in the past year; we have put in a lot more computers and written new software for the car.”
The Stanford Racing Team will be provided with course maps just 24 hours prior to the actual race, but Thrun said no major modifications will be made in that time. As with many other teams, much will be left up to the advances already made in robotics.
“We hope that many of these cars will be able to do well,” Thrun added, “and that this technology will have huge leaps forward in the field of robotics.”