By ROSE JENKINS

CONTRIBUTING WRITER

“Drivers not required.”

That’s the motto adopted by the Stanford Racing team and printed on the side of “Stanley,” the Volkswagen S.U.V. Touareg that’s been modified to drive without human guidance. But there’s much more to Stanley than its motto, which is a twist on the “Drivers wanted” slogan of Volkswagen. The drive-by-wire vehicle boasts seven computers in its trunk and a host of other safety and perceptual features that its design team hopes will propel its success in a 175-mile race called the DARPA Grand Challenge.

The second-annual race, hosted by the Defense Advanced Research Projects Agency, or DARPA, is designed to develop autonomous robot vehicle technologies for military purposes and is offering a $2 million prize to the first car to make it across a course in the Mojave Desert in less than 10 hours. In the 2004 race, the frontrunner, an entry from Carnegie Mellon University, went eight miles before crashing. However, this year more and better-prepared teams seem to be interested in the Challenge.

The Stanford Racing Team first started developing Stanley in June 2004, under the leadership of Sebastian Thrun, associate professor of computer science and head of the Stanford’s Artificial Intelligence Labs.

In early April of this year, Stanley’s creators were notified that they were one of 118 out of the 195 entries to receive a site visit. Today, they will find out if they are one of the 40 teams to pass the second-to-last stage of screening. Race day, October 8, will see just 20 cars competing.

To date, Stanley has logged 200 miles of autonomous driving time, some of it on the Mojave Desert route from the 2004 DARPA Challenge. This year’s path won’t be announced until two hours before the race, though DARPA organizers have said that it is in Southern California.

Stanley is a standard gray Touareg, donated by Volkswagen and customized for drive-by-wire navigation and steering by Volkswagen of America’s Electronics Research Laboratory in Palo Alto, one of the Stanford Racing team’s major collaborators. Its major physical additions are oversized wheels, reinforced bumpers and a protective plate on the underside of Stanley’s body. Unlike some of the golf carts, utility vehicles and other machines that are among Stanley’s competitors, it is completely legal for driving on streets.

“That was an important design constraint to facilitate the testing of the robot,” said Vehicle Lead Cedric Dupont of Volkswagen in a press release.

Members of the Stanford Racing team have equipped Stanley with a GPS system and sensors that can pinpoint its location within two miles. Lasers give the car a vision span of four car lengths ahead, while radar allows it to search for further obstacles as it reaches race speeds of up to 35 mph

Additionally, stereo-vision cameras take pictures of the course that are used to determine the optimal driving course by the seven Pentium M laptops in the trunk, which have their own shocks and back-up networks. The computers are also hooked up to an electronic break, throttle and steering system. And the whole car, electronics and all, is powered by a turbo diesel engine that will run on a single tank of gas.

The most important component of Stanley’s design is the sophisticated software, which essentially drives the car. A computer algorithm analyzes input from the sensors to evaluate alternative paths for the car as often as 10 times a second.

The first prototype of the car emerged from a computer science class offered in fall quarter of this year titled “DARPA Grand Challenge.” Most of the 40 some students involved in the Stanford Racing team, who range from undergraduates to doctoral candidates, got involved with the team through that course.

According to David Stavens, a doctoral candidate in computer science, Stanley also benefited from projects completed in the winter quarter computer science class “Introduction to Computer Vision.”

Aside from encompassing students from all over the School of Engineering, the Stanford Racing Team includes eight Stanford professors, a post-doctoral researcher and 17 external members from all over the Bay Area. The team is supported not only by Volkswagen, but also by Mohr Davidow Ventures (a local venture capital firm) and Android, as well as by secondary contributors Intel, Honeywell and TYZX.

While a strong performance in the Grand Challenge is the Stanford Racing Team’s first goal, the hope is that the expertise developed in building Stanley can be used in improving the systems of commercial vehicles.

“After the contest, we are interested in adapting Stanley’s technology into safety systems for commercial passenger cars,” Stavens said.

Many new cars already feature automated features such as anti-lock brakes, adaptive cruise control and electronic parking aids. The knowledge needed to create obstacle-avoidance and lane-keeping technologies could come from the team’s experience building Stanley. The car itself would have an extra set of reflexes.

“The result,” Stavens explained, “would be a vehicle that watches what a human does, reacting periodically and harmoniously with the driver to address specific perceived safety concerns.”