Coming soon to your street: driverless taxis?

Driverless cars are becoming a thing.

Google runs about a dozen of them on the roads of California and Nevada. They’ve logged over 1.7-million miles, and have been involved in 11 minor traffic accidents. That sounds low, but it’s more than double the national average of incidents per 100,000 miles, according to a recent posting on the Wall Street Journal’s Digits blog.

(The driverless vehicles were apparently not responsible for the mishaps.)

And now, researchers in California are kicking the tires on one of the logical extensions of this technology – driverless taxis. And the benefits could be endless.

“A driverless car inherently has some advantages over a human-driven car, in terms of efficiency, because if it’s programmed correctly, it can accelerate and brake more smoothly, saving energy there,” says Jeffery Greenblatt, staff scientist at the Lawrence Berkeley National Laboratory, in a conversation with Yahoo Canada.

“It can plan its route perhaps more optimally – although that gap is becoming a little smaller with people having ready access to route-planning software on their cell phones. But still, a computer might be able to plan on the fly a little more responsibly than a human driver can.”

The former cab driver in me would gladly accept that challenge. I did two years on the streets of Toronto, and developed a sharply honed instinct for the timing, short cuts and inescapable traffic traps of Canada’s largest city. It’s not just knowing where to be in theory; it’s mastering where not to be in reality.

Greenblatt concedes the technology is not yet that far down the road. Instead, he points to significant advantages in efficiency.

“The vast majority of trips people take in the U.S. – and probably Canada and elsewhere as well – are taken alone, or with a single other passenger,” he explained.

“Most of those can be satisfied by a far smaller vehicle than the taxis we’re familiar with.”

Smaller cars also burn less energy, he notes, and reduce greenhouse gas emissions – particularly if the so-called “autonomous” cabs are electric.

“The image-processing software that they have, the ability to identify hazardous objects – it’s a problem that engineers have told me is tractable. It just requires great amounts of information, but that’s all available,” Greenblatt said.

But how do you flag a driverless taxi down? Who do you argue with if you don’t think it’s taking the right route?

And who’s legally responsible if there’s an accident?

“There is a great deal of concern, I think, about how the public will respond to an accident that is found to be caused by the computer control,” Greenblatt said.  

“And of course, it’s inevitable. These things cannot be perfect ahead of time. I think the consensus is that it may come back to the manufacturer of the vehicle being the liable party.”

It appears there’ll be no shortage of manufacturers. Greenblatt says a number of big car companies are developing commercial driverless cars, with a target market date of 2020.

So, I ask him, if he’s got to get across town right now, and there’s a choice between a traditional cab and a driverless one, which would Greenblatt take?

“If they’re both available, and both say they can get me there fast, I would probably go with the cheaper one. It’s up to economics to determine which one that’s going to be.”