Look, Ma, no humans.

Such is the rallying cry of the Stanford Racing Team, composed of nearly forty Stanford faculty, researchers and graduate students. The team has been in collaboration with Volkswagen’s Silicon Valley lab to try to win the 2007 Defense Advanced Research Projects Agency (DARPA) Grand Challenge, seeking a car that can navigate a simulated urban environment for sixty miles in less than six hours, without any human guidance. To win the $2 million prize money, the Stanford team must do all this while finishing first in what will likely be a field of more than thirty teams.

The Racing Team is fresh off a first-place finish in the 2006 DARPA Grand Challenge, where its robot car, a Volkswagen Touareg dubbed “Stanley,” won a race through the desert.

Michael Montemerlo, senior researcher at Stanford’s AI Labs and software lead for the Racing Team, said the squad faces additional challenges this year.

“Last year, you had a robot that had to complete a course in the desert all by itself without any manual intervention,” Montemerlo said. “So you press the go button and either it makes it or it doesn’t. In the urban challenge it also has to act reasonably in the face of other traffic.”

According to Montemerlo — the new car, a Passat donated by Volkswagen — will have to reliably detect and track a variety of new obstacles. These will range from static hazards, such as curbs and holes, to dynamic hazards including cars, bicycles and pedestrians. And it will have to do so while obeying traffic laws, acting appropriately at intersections and doing everything else an experienced human driver would.

David Stavens, a fifth-year Ph.D. candidate in computer science and co-creator of Stanley, said the team is in it for the “humanitarian aspect”. With 43,000 deaths on the road every year, countless lives could be saved if human drivers were replaced by machines that could potentially have quicker reactions, better senses and never get tired.

“We have a tremendous amount of space on this campus devoted to parking,” Stavens said. “But an automatic car could drop you off and go park itself some distance away.”

Stavens said that an automated car could be a more efficient use of both time and money. For instance, three or four people could easily share one car, since it could drive itself between their locations. And commuting time would no longer consist of wasted hours stuck on either end of the working day. In a self-driving car, the passenger’s attention could be focused on more productive tasks, such as answering email.

“It could also affect traffic,” Montemerlo added. “With communication between vehicles, you could increase [traffic flow] substantially so you actually get places faster on the highway. And you can imagine people driving who can’t drive today, such as elderly people who can’t see very well. Or even kids. Their parents could put them in the car and send them to where they need to go.”

Montemerlo emphasized that the move towards autonomous cars like Stanley and the winner of the 2007 Grand Challenge would not happen overnight.

“I don’t think cars will become autonomous instantly,” Montemerlo said, pointing to “smart” anti-lock brakes that are already in cars. “You should think of that car as being just a tiny bit autonomous. Cars are going to have more and more of these adaptive systems and one day you’ll wake up, and you’ll have a car that’s able to drive itself.”

The Stanford Racing Team is led by Sebastion Thrun, an assistant professor of computer science and the head of Stanford’s 150-person AI Labs. Last year, Stanley was first out of 23 robot cars to compete in a 132-mile race through the Nevada desert and one of only four which successfully completed the drive.

The team took away $2 million from last year’s race. Some of the prize money was used to fund this year’s research, while the rest went to the School of Engineering to endow a scholarship.

“From now until pretty much indefinitely there will be a student funded by Stanley’s prize money,” Stavens said.

This year, the Racing Team sought money from a variety of sponsors, including Mohr Davidow Ventures, Intel, Google and NXP Semiconductors. Sponsors donated money in return for the right to put logos on the autonomous car and exchange information with the team. The team also applied for and received a technology development grant from DARPA, which may total up to $1 million. The team also includes members of Volkswagen Labs, based in Palo Alto.