Stanford Racing Team Robot 'Stanley' Named DARPA Grand Challenge Finalist
Stanley Heads to October 8 Primm, NV Desert Competition
STANFORD, Calif., Oct. 5 -- Stanley, the Stanford Racing Team's entry into the upcoming Grand Challenge race of robot cars, emerged from qualifying events this past week with a squeaky clean driving record. In four runs on three two-mile courses designed to test the skills of autonomous vehicles at the speedway in Fontana, Calif., Stanley was the only car that never hit an obstacle.
Stanley also consistently ranked in the top three in terms of finishing times, often constrained only by speed limits imposed by the Defense Advanced Projects Research Agency (DARPA). Based on the Volkswagen Touareg's performance at the qualifying events, DARPA today selected Stanley along with 22 other cars to compete in the Grand Challenge Oct. 8 in Primm, NV.
"It has been great to have made so many new friends and to have had the opportunity to collaborate with such an enthusiastic group of robotics innovators during the past week," said Stanford Racing Team leader Sebastian Thrun, an associate professor of computer science in Stanford's School of Engineering. "I'm proud of my team and the work that's been done. The much tougher part is yet to come, but we are more confident than ever that we can meet the challenge."
None of the 23 teams knows what lies ahead in Primm. DARPA won't reveal the exact route or nature of the roughly 150-mile desert obstacle course until two hours before the event starts in the morning of Oct. 8. Cars will have to improvise their way around the course, dodging obstacles yet staying on the rough dirt road and making the best time. The winning car will be the one that navigates the course, which also ends in Primm, the fastest in less than 10 hours.
Hard-driving technology
Stanley's performance in Fontana is the product of a year's worth of work on integrating artificial intelligence software and a rich mix of sensors to give the car the ability to spot obstacles and quickly plan the best route to get around them. The car uses a combination of lasers, radar and, video to "see" what's ahead, and GPS and inertial navigation to keep track of how it's progressing. The mostly custom-written software is run on six Intel Pentium M- based computers rack-mounted in the trunk.
"Since last fall, dozens of Stanford faculty and students have been working to write, test and refine the code that makes Stanley able to find the road and make sound decisions about how to follow it safely and quickly," says Mike Montemerlo, a Stanford computer science postdoctoral researcher. "Finding the road is easy for people but incredibly hard for a robot. Avoiding some obstacles can be hard for people and machines alike."
Along the way, Stanford's engineers have collaborated with nearby engineers at the Volkswagen Electronics Research Lab in Palo Alto, Calif. The team has also benefited from the hands-on involvement of venture capital firm MDV-Mohr Davidow Ventures. Additional team supporters include energy drink maker Red Bull, startup Android and a few other technology companies.
The team certainly hopes to win the Oct. 8 contest, but also hopes that the robotics breakthroughs achieved in developing Stanley will advance science and automotive safety technology. Intelligent driver assistance systems could help reduce the approximately 44,000 fatalities that occur from traffic accidents in the U.S. each year.
On the web please visit: www.stanfordracing.org www.vwerl.com www.mdv.com www.redbull.com www.grandchallenge.orgPhoto: NewsCom: http://www.newscom.com/cgi-bin/prnh/20051005/SFW108
AP Archive: http://photoarchive.ap.org/
AP PhotoExpress Network: PRN13
PRN Photo Desk, photodesk@prnewswire.com