Look, no hands! Test driving a Google car

MOUNTAIN VIEW Calif. Sun Aug 17, 2014 1:18am EDT

Chris Urmson, (L) director of Google's Self-Driving Car Project, and team members  Brian Torcellini, Dmitri Dolgov, Andrew Chatham, and Ron Medford (R), who is director of safety for the project, pose for a photograph in front of a self-driving car at the Computer History Museum after a presentation in Mountain View, California May 13, 2014.     REUTERS/Stephen Lam

Chris Urmson, (L) director of Google's Self-Driving Car Project, and team members Brian Torcellini, Dmitri Dolgov, Andrew Chatham, and Ron Medford (R), who is director of safety for the project, pose for a photograph in front of a self-driving car at the Computer History Museum after a presentation in Mountain View, California May 13, 2014.

Credit: Reuters/Stephen Lam

Related Topics

MOUNTAIN VIEW Calif. (Reuters) - The car stopped at stop signs. It glided around curves. It didn't lurch or jolt. The most remarkable thing about the drive was that it was utterly unremarkable.

This isn't damning with faint praise. It's actually high praise for the car in question: Google Inc.'s driverless car.

Most automotive test drives (of which I've done dozens while covering the car industry for nearly 30 years) are altogether different.

There's a high-horsepower car. A high-testosterone automotive engineer. And a high-speed race around a test track by a boy-racer journalist eager to prove that, with just a few more breaks, he really could have been, you know, a NASCAR driver.

This test drive, in contrast, took place on the placid streets of Mountain View, the Silicon Valley town that houses Google's headquarters.

The engineers on hand weren't high-powered "car guys" but soft-spoken Alpha Geeks of the sort that have emerged as the Valley's dominant species. And there wasn't any speeding even though, ironically, Google's engineers have determined that speeding actually is safer than going the speed limit in some circumstances.

"Thousands and thousands of people are killed in car accidents every year," said Dmitri Dolgov, the project's boyish Russian-born lead software engineer, who now is a U.S. citizen, describing his sense of mission. "This could change that."

Dolgov, who's 36 years old, confesses that he drives a Subaru instead of a high-horsepower beast. Not once during an hour-long conversation did he utter the words "performance," "horsepower," or "zero-to-60," which are mantras at every other new-car test drive. Instead Dolgov repeatedly invoked "autonomy," the techie term for cars capable of driving themselves.

Google publicly disclosed its driverless car program in 2010, though it began the previous year. It's part of the company's "Google X" division, overseen directly by co-founder

Sergey Brin and devoted to "moon shot" projects by the Internet company, as Dolgov puts it, that might take years, if ever, to bear fruit.

So if there's a business plan for the driverless car, Google isn't disclosing it. Dolgov, who recently "drove" one of his autonomous creations the 450 miles (725 km) or so from Silicon Valley to Tahoe and back for a short holiday, simply says his mission is to perfect the technology, after which the business model will fall into place.


Judging from my non-eventful autonomous trek through Mountain View, the technology easily handles routine driving. The car was a Lexus RX 450h, a gas-electric hybrid crossover vehicle - with special modifications, of course.

There's a front-mounted radar sensor for collision avoidance. And more conspicuously, a revolving cylinder perched above the car's roof that's loaded with lasers, cameras, sensors and other detection and guidance gear. The cylinder is affixed with ugly metal struts, signaling that stylistic grace, like the business plan, has yet to emerge.

But function precedes form here, and that rotating cylinder is a reasonable replacement for the human brain (at least some human brains) behind the wheel of a car.

During the 25-minute test ride the "driver's seat" was occupied by Brian Torcellini, whose title, oddly, is "Lead Test Driver" for the driverless car project.

Before joining Google the 30-year-old Torcellini, who studied at San Diego State University, had hoped to become a "surf journalist." Really. Now he's riding a different kind of wave. He sat behind the test car's steering wheel just in case something went awry and he had to revert to manual control. But that wasn't necessary.

Dolgov, in the front passenger's seat, entered the desired destination to a laptop computer that was wired into the car. The car mapped the route and headed off. The only excitement, such as it was, occurred when an oncoming car seemed about to turn left across our path. The driverless car hit the brakes, and the driver of the oncoming car quickly corrected course.

I sat in the back seat, not my usual test-driving position, right behind Torcellini. The ride was so smooth and uneventful that, except for seeing his hands, I wouldn't known that the car was completely piloting itself - steering, stopping and starting - lock, stock and dipstick.

Google's driverless car is programmed to stay within the speed limit, mostly. Research shows that sticking to the speed limit when other cars are going much faster actually can be

dangerous, Dolgov says, so its autonomous car can go up to 10 mph (16 kph) above the speed limit when traffic conditions warrant.


In addition to the model I tested - and other such adapted versions of conventional cars - Google also has built little bubble-shaped test cars that lack steering wheels, brakes and accelerator pedals. They run on electricity, seat two people and are limited to going 25 mph (40 kph.) In other words, self-driving golf carts.

Google's isn't the only driverless car in development. One of the others is just a few miles away at Stanford University (where Dolgov did post-doctoral study.) Getting the cars to

recognize unusual objects and to react properly in abnormal situations remain significant research challenges, says professor J. Christian Gerdes, faculty director of Stanford's REVS Institute for Automotive Research.

Beyond that, there are "ethical issues," as he terms them. "Should a car try to protect its occupants at the expense of hitting pedestrians?" Gerdes asks. "And will we accept it when

machines make mistakes, even if they make far fewer mistakes than humans? We can significantly reduce risk, but I don't think we can drive it to zero."

That issue, in turn, raises the question of who is liable when a driverless car is involved in a collision - the car's occupants, the auto maker or the software company. Legal issues might be almost as vexing as technical ones, some experts believe.

Self-driving cars could appear on roads by the end of this decade, predicted a detailed report on the budding driverless industry issued late last year by investment bank Morgan Stanley. Other experts deem that forecast extremely optimistic.

But cars with "semi-autonomous" features, such as collision-avoidance radar that maintains a safe distance from the car ahead, are already on the market. And the potential advantages - improved safety, less traffic congestion and more - are winning converts to the autonomy cause.

"This is not a toy," declared the Morgan Stanley research report. "The social and economic implications are enormous."

For video of a car similar to the one tested, see reut.rs/YfiJez

Paul Ingrassia, managing editor of Reuters, is the author of three books on automobiles, and has been covering the industry since 1985. The car he drives is ... a red one.

(Reporting by Paul Ingrassia; Editing by Frances Kerry)

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/
Comments (8)
loyalsys wrote:
I’m worried that since an all male set of engineers developed the car it might not drive the way a woman driver does. Chance of an accident definitely higher.

How about stopping at railroad crossings. They might not all look the same to an inexperienced car.

Also, it would have to take a driving test. Has to be old enough and be good at parking or otherwise couldn’t get a license.

Definitely not an easy road ahead.

Aug 17, 2014 9:28am EDT  --  Report as abuse
phalcon1 wrote:
I am looking forward to the day that these are allowed on all streets. I am more looking forward to the day when a majority of people start using them. Driver’s education the last couple of decades has dramatically dropped. Every year I see more and more people that should never be in control of these machines.

I get such a kick out of the people that don’t understand how these will be so much safer than today’s average driver. For one thing: the car is programmed with the rules of the road (not by a male brain as suggested by loyalsys). If every driver on the road followed these rules, there would be NO (ZERO) accidents except for mechanical failure – and even those would be reduced because some of them are addressed in the rules. Computers do one thing: they follow rules. They also don’t have egos. They don’t get bothered when some other car races ahead of them only to suddenly get in your lane and slow down to turn. And they don’t try to show off for the cute girl in the car next to them.
To the “problem” mentioned about liability: If they outfit all robo-cars with 360° cameras, I’m willing to bet that 99.9% of all accidents will prove to be the other driver’s fault. The only reason for the 0.1% chance of it being the robo-car is for unforeseen circumstances.
To the “problem” of choosing between protecting its occupants or hitting a pedestrian – easy: don’t ever hit the pedestrian. Rely on the safety restraints inside the vehicle to protect the occupants. Have a cascading set of rules: Choice: Going to hit a suddenly appearing pedestrian or turn into a parked car? Hit the car. If pedestrian could be given a ticket for Jay Walking, then great. If not, No-Fault insurance would take care of the collision since it was unavoidable.

Aug 17, 2014 4:32pm EDT  --  Report as abuse
JohnOfOnt wrote:
I think this will be just another means of destroying urban cores ; just another means of furthering big oil ; just another means of making drivers even lazier – see plenty of examples of that today walking through downtown Ottawa which has become unliveable – in part because of car alarms being used to lock vehicles so I hear beeping constantly in my area adjacent to several shopping streets – no possibility on a weekend of a quiet stroll here or in nearby residential areas ( I think there should be a hefty congestion tax however doubt that will ever happen here until congestion becomes extreme – at least then the cars move more slowly so there is less danger to pedestrians – I hope no google car comes near me – what North American cities need is proper transit , not more personal appliances ) .

Aug 17, 2014 9:53pm EDT  --  Report as abuse
This discussion is now closed. We welcome comments on our articles for a limited period after their publication.