PORTLAND, Ore. — The winners of the last Grand Challenge, an autonomous vehicle race sponsored by the Defense Advanced Research Projects Agency (Darpa), announced their entry for this year's Urban Grand Challenge.

The entries were announced Saturday (Feb. 17) during the annual meeting of the American Association for the Advancement of Science in San Francisco.

The Stanford University entry, called "Junior," is a a driverless Volkswagen Passant that will navigate over 100 miles of urban streets with over 85 other autonomous vehicles competing for the grand prize of $2 million for the fastest qualifying vehicle; $1 million for second place and $500,000 for third place. The race takes place on Nov. 3, 2007.

"The biggest challenge facing all the teams, I believe, is dealing with the other vehicles," said Mike Montemerlo, senior research engineer here at Stanford University Artificial Intelligence Laboratory where "Junior" is being built. "We can deal with fixed obstacles, like curbs, and can follow streets and stay in lanes, but when we come to a stop light and need to queue up behind other vehicles already there, or want to change lanes while avoiding collisions, that will be our biggest challenge."

To meet the task, Stanford enlisted Intel Corp. to donate its fastest dual- and quad-core microprocessors, as well as the software to program them, plus the help of some of Intel's own software engineers to handle some of the programming. Sensor streams from both laser rangefinders and video cameras constantly scan in all directions around the vehicle. From those data streams a "worldview" in created inside a computer model, then planning algorithms take over to fire the actuators that autonomously steer and brake the vehicle.

"Our multicore processors are very well suited to the problems facing Junior," said Scott Ettinger, researcher in Intel's Corporate Research Group who has been assigned to assist the Stanford Racing Team. "Many of the data streams, for instance, can be very conveniently assigned by programmers to separate cores, and since the programmers will be using our compilers too, some of the lower level code multithreading will be automatically generated."

Of the 85 teams competing this year, 60 made applications under Darpa's "fast track" program to qualify for up to $1 million in financial aid per team. Only 11 teams won awards, including all of the winners from 2005.