Google Isn't The First To Dream Of Robotic Cars

IRA FLATOW, host:

You're listening to SCIENCE FRIDAY. I'm Ira Flatow.

A couple of weeks ago, Eric Schmidt, the CEO of Google, took the stage at a tech conference in San Francisco, and he rolled out his vision for the future of search technology, the future of mobile computing, the future of human life, right down to the way we get around.

Mr. ERIC SCHMIDT (CEO, Google): Computers will clearly be good at doing the things that we're no good at. Your car should drive itself. It's amazing to me that we let humans drive cars. Computers should drive cars. It's obvious, right? If you think about it, and all of a sudden it'll be much, much safer when we let the computers do the things that they're good at, and the humans can talk or eat or whatever they want to do in the car. You get the idea.

FLATOW: Well, you know, not everybody at that meeting did get the idea. It really wasn't clear at the time if this was some kind of utopian joke or if Schmidt was really serious. But then last weekend, Google announced that its engineers are indeed designing cars that drive themselves. You know, this isn't a revolutionary idea. It's been kicked around by tech enthusiasts for quite a while. If you've been reading science and tech magazines for the past decades, you've seen them.

For example, let me read from an issue of Popular Science way back in May of 1958. It says: The car that drives itself. Your car, in the future, will be run by black boxes while you watch.

Or we can go August of 1958 in Popular Mechanics, where it says: This car has electric brains.

So whatever happened to those 1950s cars with electric brains, or cars that would drive themselves? And could Google be the one that finally brings the robotic car to the road?

Harry McCraken is the founder and editor of Technologizer. That's a website about personal technology, and he's a columnist for Time.com. And he dug up those old magazine articles for a piece he wrote this week on robotic cars at Technologizer.com.

Welcome to SCIENCE FRIDAY, Mr. McCraken.

Mr. HARRY McCRAKEN (Founder and Editor, Technologizer): Hi, Ira.

FLATOW: Did you follow those old articles as well as I did? I use to love reading those things. I don't know if you were as old as I am to have read them when they first came out.

Mr. McCRAKEN: You know, it's a topic that comes up constantly in both Popular Science and Popular Mechanics. And I think probably as long as there have been automobiles, there's been a notion that maybe they should drive themselves. And big companies have invested effort in doing that, and wacky inventors working on their own have done it, too.

FLATOW: Yeah. And what is the new reaction to this Google idea and these ideas in general? Do people really want to do this, or do they feel fearful that if they're not driving the car, the car is gonna get into some kind of trouble or crash or something like that?

Mr. McCRAKEN: I think it's an odd excitement of excitement at the idea, and also nervousness. Anybody who uses a computer knows that computers make a lot of mistakes, in part because they're programmed by human beings who make a lot of mistakes. And...

FLATOW: Right.

Mr. McCRAKEN: ...it's just such an ambitious idea. I mean, there aren't many things a human does that involve dexterity and reading and thinking and the fast reaction to things that change millisecond by millisecond in the way the driving does.

FLATOW: Yeah. You were at that conference where Schmidt rolled out the vision. What was your immediate reaction as he was speaking about it?

Mr. McCRAKEN: Well, he had a sort of - a cryptic smile and was being mysterious. And at the time, I kind of wondered whether Google saw this as something they might be involved in. But I did not expect to learn shortly thereafter they actually have put a lot of resources into this for quite a while.

FLATOW: A couple of years ago, we talked about a DARPA race with cars to try to come up with a race that the car could drive itself around a closed course. And if I read what Google is doing, they basically they've hired all those folks who worked on that race to make a car for them.

Mr. McCRAKEN: They kind of have a dream team of a lot of the people who won those races. Obviously, the Defense Department is very interested in the idea of self-driving vehicles. And there's actually a bill that says it should be the goal of the Defense Department to have one-third of the ground vehicles in the military be self-driving by 2015.

So DARPA has given away millions of dollars to scientists who have built vehicles that have driven themselves. And for that contest, it's been done in the Mojave Desert and on Air Force bases and places where there aren't gonna be any random pedestrians, as opposed to what Google has done, which is to send these cars out on roads in California.

FLATOW: But they're not did they run I mean, through traffic, through all kinds of traffic with no problems?

Mr. McCRAKEN: You know, apparently, the most serious issue was something when they were rear-ended by somebody, and so apparently it wasn't Google's fault. Most of the driving they've done, there has been a human being there who's intervened occasionally when something has happened, like a bicycle that's running a red light. They've done 140,000 miles with only a little intervention, and they say about 1,000 miles with no intervention.

FLATOW: You know, you have you can actually now document the first time that a robotic car got into trouble and said, it's not my fault, so...

(Soundbite of laughter)

FLATOW: As opposed to people always saying, it wasn't my fault. Now their car can say it's not my fault. What makes this car work so much better when all those cars in the '50s, in Popular Mechanics and stuff like that, did not get produced?

Mr. McCRACKEN: Well, in the past, a lot of the time the basic notion was to essentially sent cars along a track - and sometimes where there was a physical track, and sometimes it was an electronic one. You know, GM tried embedding radio cables in the road and having control boxes along the side of the road that would issue commands to cars. And Google's system is a lot, of course, sort of replicating what the human being does. It uses radar and machine vision to see the road and see other cars and people and signs. Obviously, Google already has a tremendous amount of data about the rules of the road and where the one way streets are, and so forth, which it uses. So, it's much closer to looking around, thinking, understanding what you should be doing, and reacting appropriately.

FLATOW: Mm-hmm. Let's go to the phone. Sharif(ph) in Burke, Virginia. Hi, welcome.

SHARIF (Caller): Yeah, hi. So, actually, I'm an engineer here in the D.C. area. And I have been following this technology for quite awhile. I was in grad school when some of my friends, my colleagues, where working on that DARPA project. In fact, one of my colleagues at my company now, was formerly on the Case Western team for that DARPA challenge.

FLATOW: Hmm.

SHARIF: And I'm a firm believer in self-driving cars. I feel that if we're not going to make a solid investment in high-speed rail or other kinds of transportation that sort of go beyond the vehicle, then perhaps we should look to self-driving cars. But my sense is that policy - in this case, it's the policy that lags far, far behind in technology. So are you aware of any policymakers or potential policies that sort of being tossed around right now...

FLATOW: What kind of policies...

SHARIF: ...that are trying to move this forward?

FLATOW: What kind of policies would we need, for example? We need traffic laws that are different, or...

SHARIF: You know, I think it's more like insurance policies.

FLATOW: I see. Let me ask Harry. Any...

Mr. McCRACKEN: Yeah. I mean there's a huge amount of infrastructure and laws and things like insurance policies that relate to cars. And they all have one thing in common. They assumed a human being is driving. And clearly, if this becomes a reality at any point, a lot of it has to change. Google says that they're confident that what they have been doing was legal. But even then, they had to they said they get a heads up to the police departments in the areas where they drove. Because, you know, for a century now, we've assumed that cars are driven by a human rather than by artificial intelligence.

FLATOW: All right, Sharif. Thanks for calling.

SHARIF: Thank you.

FLATOW: You're welcome. And I guess you could foresee that. If this idea catches on, it might do away with some cars, right? I mean people why would people who need to go to the grocery store and they'd - you know, maybe there'd be services like taxis without people in them. You just dial up, a car comes, picks you up and takes you where you'd like to go. Actually it would be good.

Mr. McCRACKEN: Well, right now, I'm also be fascinated to see how it changes people's reactions to that cars, because, you know, Americans love to drive their cars. And it's partially about getting places, but it's also about a relationship with the vehicle. And if you're not steering and accelerating and enjoying the experience, maybe you'll drive less.

FLATOW: Mm-hmm. You wrote an article in Technologizer of going through the history of these ideas of cars that drive themselves. Give us the flavor of three or four of those ideas.

Mr. McCRACKEN: Sure. Well, there had been a bunch of them, and a lot of them kind of remind me of things I saw in Tomorrowland at Disneyland...

FLATOW: Oh, yeah.

Mr. McCRACKEN: ...when I was kid. I mean my favorite I think it was one called the Urbmobile, which the Cornell Aeronautical Laboratory is working on the '60s. And the Urbmobile is a car that you can drive traditionally. It was electric, but you could do it yourself on roads. But they also anticipated that maybe, by 1985 or so, there would be these tracks, these highways that were made of tracks. And if you don't feel like driving, you could drive your Urbmobile onto tracks, and it would sort of become a self-driving car then, kind of like a loosely coupled railway car.

FLATOW: Mm-hmm. 1-800-989-8255. A tweet from Ted Pablo(ph) came in. It said, there's a stimulus money funding cyber physical system research in autonomous urban vehicles. Are you aware of that?

Mr. McCRACKEN: I haven't heard of that. But over the years, a lot of research that has been done has been funded by the government, dating back to the '60s, at least.

FLATOW: Mm-hmm. Do you think we are close to this kind of car? Or, you know, research, we always say, we're always 10, 20, 30 years away of something like this.

Mr. McCRACKEN: Compared to some of the organizations who have done this in the past, I think Google has been smart enough not to say this will all be reality by a given date. John Markoff had a very good story on the Google car in The New York Times. He said that the most optimistic predictions are maybe where eight years away from something happening with this.

FLATOW: Mm-hmm.

Mr. McCRACKEN: But it's absolutely true. It's not just about the technology. It's about legislation and all kinds of stuff. And even once the technology exists, do the big car companies start building in the cars automatically? What happens to all the non self-driving cars out there? Do you have to take them off the road or can they interact with the self-driving cars?

FLATOW: Mm-hmm. Yeah. So it's not something that's going to happen overnight. And it's going to take a lot of infrastructure to be put into place before any of these happens.

Mr. McCRACKEN: Absolutely.

FLATOW: Well, thank you very much, Harry, for taking time to be with us today and going down memory lane.

Mr. McCRACKEN: My pleasure.

FLATOW: Harry McCracken is the founder and editor of Technologizer, that's a website about personal technology. And he's a columnist for Time.com.

Copyright © 2010 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.