The Competition That Made Google’s Automonous Cars Possible

Google just released a video of its new driverless car. For some time now, we’ve all been aware of Google working toward this. But where did all this innovation start? The following is an excerpt from my upcoming book, “A Curious Path: Creativity in an Age of Abundance,” on the crowdsourcing competition that put the driverless car movement in motion.

 

 

 

It was a sunny Saturday morning in 2004 when race teams from all over the U.S. converged on the Southern California desert to make automotive history. Their course began at Barstow—about a hundred miles east of Los Angeles—and headed 142 miles out into the rugged windswept terrain to the small town of Primm, just across the Nevada border.  Of the 21 teams that showed up to race, only 15 passed the arduous qualifying event the day before.[i] As the sun rose that morning, tensions were high, because promptly at 6:30am, vehicles were to begin crossing the start line in a race signaling a dramatic change in how we get from one place to another.

In many ways this race was unique. First, the teams had no advanced notice of the exact course. The way-points (digital coordinates) were disclosed just three hours before the start.  Another oddity was that the U.S. military sponsored it as a winner-take-all event in which the fastest vehicle to cross the finish line would get a million dollars.[ii] Unlike typical auto races, though, drunken fans didn’t fill the stands. Nor were there any high-profile drivers dressed in brightly colored fire suits.  Instead, high IQs, thick glasses and computers ruled the day. Surprisingly, there were few restrictions governing the types of vehicles that could enter. Anything short of flamethrowers or attacking competitors was allowed.[iii] Vehicles of all makes and sizes showed up, from small ATV-looking things to a 15 ton truck similar to the Oshkosh MTVR defense trucks used by the US military in the Middle East.[iv] However, there was one restriction that was set in stone, “no drivers allowed!”

In 2001, Congress enacted Public Law 106-398, stating “It shall be a goal of the Armed Forces to achieve the fielding of unmanned, remotely controlled technology such that… by 2015, one-third of the operational ground combat vehicles are unmanned.”[v] [vi]  The reasoning was that by reducing the number of soldiers in harm’s way, the risk of injury or death among troops would be lessened too. Unmanned vehicles—robots—are essentially to replace a majority of humans in combat. Having a large number of robotic military vehicles could pay dividends in other ways too. Being super strong and tireless, robots may increase the overall efficiency and effectiveness of the military. The only problem was that, in 2001, the technology didn’t exist. So, the Pentagon handed off the project to the Defense Advanced Research Projects Agency (DARPA).

DARPA, the main R&D agency of the Defense Department, is responsible for technological innovation. It was quickly formed in the late 1950s after it was apparent the Soviet Union could launch nuclear weapons into space. The U.S. was fearful another country could control the skies, and it needed to regain its technological superiority. Since its inception, DARPA has specialized in high risk projects leading to the development of new technologies reaching beyond the immediate or practical use of national defense.[vii]  Contrary to any other government agency, DARPA is lean, agile, and independent. [viii] Its impressive run of successes includes creating the setting that supports our modern computer environment. Things like virtual reality, stealth technology and computer networks all stem from DARPA inventions.[ix]

Creating a fleet of unmanned vehicles was a project tailor-made for DARPA. Their culture and resources were set up for just this type of thing. But instead of doing the job internally, they outsourced it. Not only that, they opened it to anyone willing to take a stab at it. The Grand Challenge, as it was called, was open to innovators in a vast array of fields, “includingadvertisersand corporatesponsors,artificialintelligencedevelopers,automanufacturersand suppliers,computerprogrammers,futurists,inventors,motorsports enthusiasts,movieproducers,off-roadracers,remote-sensingdevelopers, roboticists,sciencefictionwriters,technologycompanies,universities,video gamepublishers….”[x] Essentially, any team who could design an unmanned vehicle to complete the course was eligible.

So on that Saturday in the desert, a lot was on the line. By all media accounts, though, that day was a disaster. From the beginning, vehicles lost their way and rendered themselves useless. Others advanced skittishly as if they scared of their own shadows. Actually, one did see a shadow and swerved off the road to avoid it. The 15 ton behemoth shied away from a cluster of small bushes in an act of self-preservation. The best performance of the day was a car that burst into flames on an embankment seven miles down the course.[xi] None of the vehicles finished the trip to Primm. Thus, the million dollars went unclaimed. But that’s not the end of the story.

From his computer at Stanford University, SebastianThrun was checking in on the competition now and again to see how the challenge was going. What he saw was his field as a laughing stock. He felt the pain of the entire field of robotics. As the head of the artificial intelligence lab at Stanford, he was determined to right the wrong of the 2004 challenge. In doing so, he formed a team of researchers to compete in the next Grand Challenge.[xii]

Shortly thereafter, Thrun gained the attention of Volkswagen’sPaloAltoresearch and developmentteam which sponsored Stanford’s team with a VW Toureg. Building on lessons learned by teams in 2004, and with over 20 years of experience in the field, Thrun led Stanford’s team to victory in 2005. Their robot, Stanley, crossed the finish as the fastest of three autonomous vehicles to finish the course at six hours and 53 minutes.[xiii] In the span of four years, an open competition spurred a loose network of scientists, engineers and designers to work together to finally create the driverless car. That’s crazy.

The overarching lesson from the Grand Challenge is how creative projects are getting done today. A vast majority of creative endeavors these days are collaborative, like the Grand Challenge. Also, many of today’s creative solutions come from open competitions. Crowd funding and crowd sourcing are variations on this. And the speed of innovation is ever increasing.

DARPA knew what they were doing when they created the competition. They knew the creative potential of a loose network of curious people would far exceed what they could do internally.  They also knew that creativity is messy. They needed a lot of people taking risks and thinking outside the box to move this thing forward. Thrun and his team were able to accomplish so much in such a short period of time because the teams in the previous challenge failed in so many ways. Those failures opened the door for future successes.

After reading so many stories like the Grand Challenge where innovative problem solving is so important, I’m convinced that creativity is a skill everyone should develop. Creativity is deeply imbedded in so many things we do and experience. Our times are complex and change is everywhere. The number of choices we face on a regular basis is overwhelming. With this in mind, and the unbelievable access to information and technology present today, our time is indeed one of the creative mind. Plus, being creative is fun.

As for SebastianThrun (the guy who headed the winning team of the Grand Challenge), he is an executive at Google now, having logged hundreds of thousands of miles in driverless cars.

 

[i] Walton, Marsha. “Technology: 15 teams qualify for Mojave robot race.” CNN.com. May 5, 2004. http://www.cnn.com/2004/TECH/ptech/03/12/darpa.race/index.html (accessed 12 14, 2013).

[ii] Ibid

[iii] Drogin, Bob, and Aaron Zitner. “No Drivers Wanted in Race for $1 Million.” Los Angeles Times. February 13, 2003. http://articles.latimes.com/2003/feb/21/nation/na-race21/2 (accessed December 26, 2013).

[iv] Hanlon, Mike. “DARPA Grand Challenge 2004 autonomous ground vehicle competition.” http://www.gizmag. March 12, 2004. http://www.gizmag.com/go/4128/ (accessed December 23, 2013).

[v] “DARPA Challenge Overview.” DARPA. n.d. http://archive.darpa.mil/grandchallenge/overview.asp (accessed December 20, 2013).

[vi] Public Law 106-698 2000

[vii] “History.” DARPA. n.d. http://www.darpa.mil/About/History/History.aspx (accessed December 30, 2013).

[viii] Van Atta, Dr. Richard. “Fifty Years of Innovation and Discovery.” DARPA > about > history. n.d. http://www.darpa.mil/About/History/History.aspx (accessed December 26, 2013).

[ix] Ibid

[x] “Grand Challenge Press release: DARPA Plans Grand Challenge for Robotic Ground Vehicles.” DARPA. January 2, 2003. http://archive.darpa.mil/grandchallenge04/media/announcement.pdf (accessed December 27, 2013).

[xi] Davis, Joshua. “Issue 14.01, Say Hello to Stanley.” Wired. January 2006. http://www.wired.com/wired/archive/14.01/stanley_pr.html (accessed December 26, 2013).

[xii] Ibid

[xiii] “Grand Challenge 2005.” DARPA. December 31, 2007. http://archive.darpa.mil/grandchallenge05/ (accessed January 6, 2014).