By Tom Foremski for SiliconValleyWatcher

I recently rode shotgun in a Volkswagen Touareg SUV. It's a great car. But this one was extra-special because it drove itself.

I was over at Stanford university conducting an exclusive interview with the engineering team that built the autonomous car, capable of avoiding obstacles and negotiating tough terrain in challenging weather conditions - without any need for human direction.

The car, dubbed Stanley, will compete with 19 others for a $2m prize on October 8, in DARPA's Second Grand Challenge race. The winner has to complete a 175-mile course in 10 hours across rugged desert terrain filled with unknown obstacles.

Pamela Mahoney, marketing director at veteran VC firm Mohr Davidow Ventures, has been working with the Stanford team. It is led by professor Sebastian Thrun, director of the Stanford Artificial Intelligence Lab. Pamela said MDV maintains close links with Stanford researchers because of its focus on early funding of startups. (Pamela also worked 10 years in the auto industry, including a spell at Chrysler, before coming to Silicon Valley.)

Itanium powered?

Volkswagen provided some engineers and the car, which bristles with gear and antennas on its roof rack. And the trunk has a rack with seven Pentium M PC boards, plus extra car batteries and various black boxes. Most of it is donated by Intel, yet another enthusiastic supporter.

Interestingly, Intel initially offered its high-end 64-bit Itanium microprocessors. These are power-hungry chips and Intel has been trying to boost the visibility of Itanium users within IT. But for cars, it's overkill, so Sebastian had to politely decline the offer and ask for Pentium M chips designed for laptops, which have low power consumption.

Apart from the computer and sensor gear, the car is a stock model that you could buy from a dealer. This is important, said Cedric Dupont and Mike Montemerlo, members of the Stanford racing team engineering group. It means that their approach is adaptable to many types of vehicles.

Another key aspect is the team's software-heavy approach to the challenge. This also makes the project more portable, and possibly adaptable to controlling other vehicles, ships or machinery.

Sebastian and his team said several times that they are not supporters of the war in Iraq, but that their technology could save lives. "You can imaging such vehicles being used by the military to transport supplies and avoid exposing people to roadside bombs," Sebastian said.

The gear on the car includes lasers, radar, GPS, and many types of sensors. So far, the team has taken the vehicle on secret trips to the Mojave desert. Secret because the mostly university teams competing for the prize are notorious for spying on each other. The Stanford team said they managed to keep their project so quiet that they heard on the grapevine that there was not going to be any Stanford entry.

Stanford's project could do very well. I can vouch that we didn't hit anything or anybody, successfully completed about eight or nine 50 meter circles, and that the car stopped exactly next to an orange cone Sebastian had placed at the start of the demo. I was impressed and a little dizzy.

(We hope to cover the race, which likely to be in Nevada this year because that will remove the need for an environmental impact report. It took 20 biologists to check out last years Grand Challenge in California's Mojave desert.)

A link explaining more:

http://media.vw.com/article_display.cfm?article_id=9583

San Jose Merc: http://www.siliconvalley.com/mld/siliconvalley/11685903.htm

UPDATE: The Stanford team has made it into the semifinals--only 20 teams will compete in the finals. DARPA press release.