Scientific American technology editor Larry Greenemeier talks with Ken Washington, vice president of Research and Advanced Engineering at Ford, about self-driving cars, some of which are already on the road.

Steve Mirsky:    Welcome to Scientific American’s Science Talk posted on March 28, 2017. I’m Steve Mirsky. Larry Greenemeier is our technology editor. Last month he found himself in Spain where he filed this automotive update. You’ll here mention of DARPA, that the Defense Advanced Research Projects Agency of the Department of Defense.

Larry Greenemeier:  I’m here at the Mobile War Congress 2017 in Barcelona. I’m with Ken Washington. Can you tell me what your role is at Ford and we can get into some of the stuff here?

Ken Washington:       Sure—I’m the vice president of research and advanced engineering at Ford.

Greenemeier:    Talking about self-driving cars but I think it’s helpful to first define what that is before we get further into the discussion.

Washington:      Right, and I agree. It’s very important to be precise in defining autonomous vehicles because autonomous vehicles, that term is often used to describe driver assist technologies which helps the drive be a better, safer, more enjoyable drive. And that’s not what we mean when we say fully autonomous vehicle, a self-driving vehicle. We define that as a level 4 autonomous vehicle that can operate as a replacement to the human driver.

Greenemeier:    So level 4. I mean I hear that a lot but I’m hearing it enough to actually start using it myself.

Washington:      Yes.

Greenemeier:    And it’s helpful to distinguish also what level 4 is.

Washington:      So level 4: it’s actually defined by the Society of Automotive Engineers. And what level 4 means is that the technology in the vehicle and the vehicle work together to the degree that does not requirement the human to engage from the journey, from the starting point to the destination point. So the vehicle does the entire driving task for the entire journey. That’s what a level 4 vehicle means within a defined area. And so, the difference between levels 3 and 4 is that a level 3, you may encounter some situations on the journey that requires the human to reengage.

And by definition then, that means the human has to be there to be engaged in the driver seat, okay, and there has to be controls in the vehicle so that the human can reengage and drive the car. And then they can go back to not driving. That’s level 3. Level 4 says you don’t have to do that. The vehicle and the software that’s in the vehicle knows how to handle the situation. And if it comes across something that is—would be challenging, it will know how to deal with that, too.

So it can slow down when it needs to. It will read the red lights. It will read the stop signs. It will detect what pedestrians are doing around it. It will detect other vehicles and it will navigate safely to where it needs to go. It will stop when it needs to and it will provide you with that entire journey without you engaging. So you can be in the back seat. You can be totally disengaged from the driving task.

Greenemeier:    So then what makes it a level 5 which would be the highest?

Washington:      So level 5 is like level 4 but everywhere at any time. So you don’t have to say it’s in this area. You could say I’ve got this vehicle and it can drive me wherever I want to go.

Greenemeier:    Does that mean no steering wheel? No steering wheel?

Washington:      Wouldn’t have to have a steering wheel. You wouldn’t have to have brake or gas pedals. So just to be clear in level 4.

Greenemeier:    Right.

Washington:      You also don’t need controls.

Greenemeier:    Okay.

Washington:      Because if you only operate that vehicle in the geofenced area.

Greenemeier:    Oh okay.

Washington:      You don’t need controls in that area. Now if you want a level 4 vehicle that you could operate outside the geofenced area, you could have a level 4 vehicle with controls.

Greenemeier:    So how far away are car companies from level 4? Is that what the five-year promises a lot of car companies are making, Ford and others?

Washington:      Well, I’ll only speak to Ford and our timeline is 2021 which is now four years from now.

Greenemeier:    Right.

Washington:      And so, we’re on a trajectory to deliver the first level 4 self-driving vehicle in 2021.

Greenemeier:    Okay. And I read that Ford’s for 40 self-driving Fusions.

Washington:      We do. Our first generation—actually, it’s not even our first. It’s our next generation, our second-generation autonomous vehicle, prototype autonomous vehicle is based on the Fusion. It’s a hybrid Fusion. And we are testing our software now on public roads. And a few weeks ago, we announced an investment in a start-up company called Argo AI. We’re really excited about that because they’re bringing to our more than 10 years of experience of developing autonomous vehicles, some additional robotics and artificial intelligence expertise and a start-up culture that’s going to allow us to continue to advance that technology. But with the built-in relationship with Ford so that they’ll know how to bring that to market and scale as they manufacture the vehicles.

Greenemeier:    So one thing I’ve been curious about—I’ve been covering autonomous cars for a couple of years now. This idea that we, that car companies or engineers want to move in this direction. Where did it come from? Is this something that car companies are saying you know what? We need something new. We need to sort of reinvigorate the car market. Let’s add these driver assist technologies and let’s progress them to an autonomous vehicle ultimately. Or is it something that people were asking for, people were saying you know what? It would be a lot safer if we could have machines at least helping us. So where did it come from?

Washington:      Well, it started with a challenge that came out of DARPA. And DARPA said what’s one of the hardest problems—because that’s what DARPA is all about is solving really hard problems. And they issued the challenge of developing a car that could drive itself. And first they did it in the desert, the desert challenge, and then they did it in a city setting, the urban challenge.

Greenemeier:    Right. Nobody finished the first one. They didn’t make it through the desert.

Washington:      That’s right. And Ford interestingly enough was the only automaker that participated in both of those. But more to your point, you question about well, where did it come from and today’s timing. I think it’s a combination of trying to do something really innovative and challenging that came from the DARPA days and people recognizing that the software and computers can actually help the vehicles be safer.

And that started with driver assist technologies increasingly getting better and better on the vehicles. And so even starting with simple things like cruise control. If you think about it, cruise control is a very rudimentary level of autonomous driving. It gives you the ability to not press the gas pedal and it sets your speed. And now we’ve got adaptive cruise control and it will set your speed and adjust your speed based on the vehicle in front of you and how close you are to that vehicle.

And so, it was very natural then to say well, if we can adjust your speed longitudinally we can do it laterally and help keep you in your lane. So lane-keeping assist came about. And then we realized well, sensors can actually help you identify things that might be in your blind spot so that we can make your drive safer by removing your blind spot with indicators on the mirror.

Greenemeier:    So when you put it that way it makes a lot of sense—

Washington:      So these features kind of got added and as we started thinking about autonomous vehicles, we started thinking about it in the context of adding more of these features. And then we kind of hit a wall and realized that we couldn’t get there. You just can’t add enough features to driver assist technologies to get to full self-driving. That’s when we did a reset and said okay. We now have to do self-driving more top down which requires a different sensor set and requires a prior map. It requires more computational power than you would put in a typical passenger car. And that’s when we started accelerating our autonomous vehicle project and it got us to where we are today.

Greenemeier:    And that would be what year where you sort of made that—

Washington:      That’s about four years ago we made that pivot from walking up from driver assist to top down.

Greenemeier:    Cybersecurity is emerging as a big issue with regard to automobile safety now. That’s another thing I don’t know if it’s coming from the carmakers or the tech companies. There’s a lot of tech companies that specialize in that and they like to sort of say the sky is falling. But I’m also hearing people, the drivers are worried about this. Of course, nobody cares about their computer or their smart phone. They’ll download anything on that. But of course, when it’s your car, you’re worried. So where do you see cybersecurity fitting into this? Is this one of the top concerns or—

Washington:      It’s a very serious issue and we take it very seriously at Ford. And we’ve started looking very seriously at cybersecurity from day one putting computers into our vehicles. That’s why when we introduced sync we introduced it in a safe way by providing a firewall between the safety critical functions and the connected functions and the infotainment systems that—because we recognize that if you bring your phone in, you connect it to the vehicle then that creates a potential vulnerability. So we had to firewall that off of the safety critical functions.

And now we do penetration testing and we—and we’re very serious about addressing the issue of cybersecurity because we want our customers to feel safe and we—we realize that we’ve been entrusted to managing and protecting their data. And so, we take that extremely seriously. It is a top priority for us.

Mirsky:               That’s it for this episode. Get your science news at our website www.scientificamerican.com where you can check out the collection of Scientific American eBooks including Evolution Versus Creationism, Inside the Controversy there at books.scientificamerican.com/sa-ebooks. And follow us on Twitter where you’ll get a tweet whenever a new item hits the website. Our Twitter name is @SciAm. For Scientific American Science Talk, I’m Steve Mirsky. Thanks for clicking on us.

[End of Audio]