Koch Foundation

NOVA Science Trust Logo

Draper

<div class="rich-text">Tiny, battery-powered cars called EN-Vs, now in R&D, can talk to each other, come when called, even drive themselves.</div>

Robotic Cars

<figure class="richtext-image full-width"><img alt="Sebastian Thrun" class="richtext-image full-width" height="419" width="672"
 src="https://www.pbs.org/wgbh/nova/media/images/cars-drive-themselves-01.width-800.jpg"/><figcaption>Stanford computer scientist Sebastian Thrun has long been driven to make smart machines. Now a smart machine named Stanley can drive him.
			<div class="credit">Courtesy Sebastian Thrun</div></figcaption>
 </figure>
 
<h2>
	Dreamers wanted</h2>
<h3>
	NOVA: How sure are you that we'll be driving—or driven by—autonomous cars in the future?</h3>

Sebastian Thrun: It's a no-brainer for me that at some point our cars will have the ability to drive themselves. Cars are young. They're just 120 years old, 100 years old, depending how you count. The car as a social phenomenon is maybe 80 years old. Just think 50 years ahead. There's no question that 50 years from now we'll have the technology for cars to drive themselves. In fact, I think we're amazingly close, and I believe this specific competition, the Grand Challenge, pushes us tremendously in that direction.
<h3>
	Is that part of the thrill of the competition?</h3>

	I draw a lot of motivation from the dream that we all share here, which is to build cars that drive themselves. It's been fantastic to see a thousand people at the same location who all share this wonderful dream.
<h3>
	How do you respond to critics who say it's impossible?</h3>

	There are always people who doubt the dreams, who have no imagination. I'm sure that the entire Industrial Revolution was a dream to people 500 years ago. Two hundred years ago, nobody would think that we'd plaster our country with pavement and filling stations to support a new infrastructure [for today's vehicles]. I prefer to spend my time with people who dream.


 <figure class="richtext-image full-width"><img alt="1935 photo" class="richtext-image full-width" height="480" width="368"
 src="https://www.pbs.org/wgbh/nova/media/images/cars-drive-themselves-02.width-800.jpg"/><figcaption>State of the art, 1935. A century before this, no one could have imagined that filling stations serving "horseless carriages" would one day dot the planet.
			<div class="credit">&copy; Hulton-Deutsch Collection/Corbis</div></figcaption>
 </figure>
 

	Maybe autonomous cars have failed so far. It's an idea that's 30, 40 years old. But the Grand Challenge should change that. Already, in test runs leading up to the race, vehicles have traveled hundreds of miles over unrehearsed terrain. That's never happened before. I think when people read the history of the automobile 100 years from now, this event will stand out as a really significant step towards making the dream a reality and communicating to people that we can actually do it.
<blockquote>

		"I think the race could push forward tremendous change in the automotive industry."
</blockquote>
<h3>
	Why pursue the dream?</h3>

	Why is it needed? Well, we have 43,000 people die every year in traffic accidents in the U.S.; 6,000 in Germany, where I'm from. And they die mostly because of human error. If you build cars that drive themselves, eventually they'll be safer than human driving. And they'll free us from the burden of focusing on driving as opposed to being more productive while we commute.
<h3>
	Is saving lives a big part of what inspires you?</h3>

	Yes. I lost a really good high school friend. He was in a car with another friend. They borrowed one of their parent's cars, and they took a turn too strong. They were 18 years old, hit a big truck head-on, and they were dead instantaneously. It was really tragic. And I wouldn't go so far as to say that Stanley [Stanford's robotic vehicle] could have assessed this and predicted this, but the type of technology that's being showcased in the Grand Challenge can actually anticipate something like this and interfere and say, "You know, you're going too fast." It can make people aware of dangers.

	Maybe my friend would still be alive if the same thing happened 30 years later, because we will have cars that can detect these things at the onset. It's a very simple thing. Suppose the car realizes you're going too fast. The car could slow down, or it could just tip the brakes very briefly and the person wakes up and sees there's a threat coming up.

	We humans usually feel that we are the best at everything we do, that we can safely drive ourselves. But tens of thousands of people die every year. We need to be open to having technology assist us, to find ways in which technology makes us safer.


 <figure class="richtext-image full-width"><img alt="VW Touareg" class="richtext-image full-width" height="446" width="672"
 src="https://www.pbs.org/wgbh/nova/media/images/cars-drive-themselves-03.width-800.jpg"/><figcaption>Driver not required: Stanley, a modified VW Touareg, constantly gathers information about the road ahead using radar, stereo and monocular vision, and an array of five lasers.<div class="credit">Courtesy Stanford Racing</div></figcaption>
 </figure>
 
<h2>
	A turning point</h2>
<h3>
	Where does the vision of cars driving themselves come from?</h3>

	We've been surrounded by cars that drive themselves for a long time. Batman's car drives itself, you know, and we've kind of taken it for granted that it's going to happen in the real world. I've developed my passion for cars that drive themselves from being stuck in traffic for many, many, many hours of my life. I don't know what it adds up to, but I feel like I've lost a year or two just in traffic. That's big to me. That's a lot of time, a lot of money that I just lose on the road. And more importantly, as I've said, I've had friends who've lost their lives in accidents.

	This is really a pivotal time. Up to now, we have had vehicles in which the intelligence is entirely provided by people and the horsepower is provided by the machine, and we're at the point where we can redefine this. We can say, "Let's also have the intelligence defined by the machine and make the device safer." In avionics it's happening already. There are aircraft that can be flown with computer control. In the automotive industry, it's the next logical step.


 <figure class="richtext-image full-width"><img alt="Batman and Robbin" class="richtext-image full-width" height="480" width="612"
 src="https://www.pbs.org/wgbh/nova/media/images/cars-drive-themselves-04.width-800.jpg"/><figcaption>In the future, everyday folks&mdash;not just superheroes&mdash;might enjoy hands-free driving.
			<div class="credit">Bettmann/Corbis</div></figcaption>
 </figure>
 
<h3>
	What interests you most about the DARPA Grand Challenge? [DARPA is the U.S. government's Defense Advanced Research Projects Agency.]</h3>

	I think the Grand Challenge is ingenious. Tony Tether [the Director of DARPA] and DARPA are incredibly smart to do this. They threw out a goal that looked seemingly simple, but it's a huge step in terms of technology. If we can show that the principle is possible, then we'll have to make it more robust and so on. But in many things—the first air flight, the first telephone—the moment you show it's possible, it's just a matter of time until it's going to be a widespread reality.
<h3>
	Are military applications an important motivation for you?</h3>

	I'm certainly not in it to build better ways of fighting wars, although I'm saddened by recent events, and if I can help to save lives, be it on the battlefield or not, I'd like to.

	But I honestly believe the impact that we're going to make will go far beyond the military domain. Take another example from DARPA: DARPA created the Internet. The Internet was called ARPANET originally. It was for military people to talk to each other. Today the Internet has revolutionized the way people interact with each other in ways that are fundamentally different from even 10 years ago, and I think the same will happen with transportation.
<blockquote>
	"The day in the future when my car commutes for me—that will be the ultimate victory."</blockquote>

	I personally think cars that drive themselves will have their greatest impact in everyday life. For instance, for elderly people, typically when they stop driving, their social networks go away because they can't see their friends anymore. They can't go shopping. They become dependent. And it often correlates with physical deterioration. What if we give them a car that drives itself? Maybe they'd live longer. They'd certainly have a better quality of life. I think just looking at the military alone would not do justice to the many enthusiasts who want to make autonomous vehicles a reality.
<h3>
	What technological advances are needed to make such vehicles a reality in everyday life?</h3>

	I honestly at this point don't really know, and I'd love to spend a year thinking about this when we analyze what actually has happened in the Grand Challenge. We've been busy just building the machine.

	To make this technology a reality, you don't just drive two or three miles, you have to drive 20,000 miles without any accident, 50,000 miles, before people can even think about adopting it. And this kind of robustness really requires a fundamentally different understanding of the road surface. I think some of the steps have been taken for this Grand Challenge.
<h3>
	What else is needed for widespread use of robotic vehicles?</h3>

	There are two big factors at this point. One is price. You have to scale down your platform from $40,000 to $2,000. That's just a matter of how much you produce. I think it's feasible. And the second is infrastructure. We have no structures right now to deal with autonomous cars. We have no special lanes on highways that say "For Autonomous Cars Only." We need a whole evolution of systems. We won't have autonomy tomorrow morning. We'll have a driver-assist system that helps a little bit to avoid obstacles, and then we'll go further, and at some point in the future, maybe 30 years from now, we'll have autonomous cars.


 <figure class="richtext-image full-width"><img alt="Sticker" class="richtext-image full-width" height="325" width="506"
 src="https://www.pbs.org/wgbh/nova/media/images/cars-drive-themselves-05.width-800.jpg"/><figcaption>Robotic vehicles are already a reality for the competitors in the Grand
Challenge, as this bumper sticker makes clear.
			<div class="credit">&copy; NOVA/WGBH Educational Foundation</div></figcaption>
 </figure>
 
<h2>
	A pure machine race</h2>
<h3>
	What's your main motivation for entering the competition?</h3>

	This is a fantastic scientific challenge. For me, it's not just about building Stanley; it's about understanding the nature of driving. We humans have much to learn about what it really entails to drive at high speeds safely. Preparing for this race, we already have made a number of fairly significant scientific discoveries, I believe, about how to perceive environments and how to drive that I think will have an impact beyond the Grand Challenge.
<h3>
	Is the Grand Challenge an important historical event for robotics?</h3>

	I think this is a watershed moment of immeasurable magnitude. This is the first endurance race where the technology has to make all the decisions. It's not just about a component like running fast or thinking fast. It's about the integration of it all. The race is substantial. It's long. It's difficult. And every single decision is made by the things we've created, not by ourselves anymore. Humans—even though the teams are important—humans are out of the loop the moment the race starts. It's a pure machine race. And that, I think, is really new and unprecedented. The moment the race starts, we'll celebrate, we'll go to a bar and have a bottle of champagne, even though the race is still ongoing. When did that happen before?
<h3>
	So it's different from other technology races in the past?</h3>

	Yeah. We've had a number of recent races—the circling of the globe in a balloon, for example, or going into deep space—that involved significant technology components. But in these races, it was always people who made the ultimate decisions. The Grand Challenge to me is entirely new. It's about the machine making every single decision. We have to build a machine that not just has the physical power to endure but also the brainpower to make all the decisions along the way. The machine becomes the actor itself. I think it's a fundamental step in the history of robotics, of humankind so to speak. I think this race will be heralded as an historical race.
<h3>
	What will completing the course successfully mean for robotics?</h3>

	The implications of this race are much broader than some people understand. It's not just about building a car that drives a couple of miles. It's about replicating one aspect of human intelligence, in this case the ability to drive, to the point that the machine can do it entirely by itself. And there's no precedent for that in history.
<h3>
	Who do you see as the main competition?</h3>

	You know, for me the competition is not against Carnegie Mellon or any other team. We're all in the same boat. The competition is to make the world's roads safer, and that's a competition that everybody in this contest is in together. Even if some other team wins, I think it's going to be a victory for the field. I'm going to be proud of having been part of the event. I'm going to be proud of DARPA for having pushed it. I think we lose if this year the best team does seven miles [the distance covered in the first DARPA Grand Challenge in 2004, in which no vehicle finished the course]. That would be a profound loss for us as a whole community of researchers.
<h3>
	What would constitute a win for you?</h3>

	To finish the Grand Challenge and cash the $2 million check! But much more so, I think the race could push forward tremendous change in the automotive industry in terms of environment perception and driver-assistance systems that can make driving safer but still keep it pleasurable.

	A couple of months ago, I spent about two hours every day commuting. The day in the future when my car commutes for me—and I can sit there, read the newspaper, do e-mail in the car while the car is driving itself—that will be the ultimate victory.


 <figure class="richtext-image full-width"><img alt="Big Check" class="richtext-image full-width" height="438" width="672"
 src="https://www.pbs.org/wgbh/nova/media/images/cars-drive-themselves-06.width-800.jpg"/><figcaption>On October 9, 2005, Stanley was the first vehicle to finish the 132-mile Grand Challenge course, averaging over 19 miles per hour over rugged terrain. (But could Stanley endorse the check?)
			<div class="credit">Courtesy DARPA </div></figcaption>
 </figure>

Roboticist Sebastian Thrun shares his excitement about real-world applications for autonomous vehicles.

Cars That Drive Themselves

How do we program moral decision-making into autonomous vehicles?
Self-driving cars will be confronted with ethical dilemmas. How do we code those into the machines?
Ethics and Self-Driving CarsPublished May 16, 2018Onscreen : Can driverless vehicles make ethical choices? The stakes are high…News video : An Uber autonomous vehicle was driving about 40 mph when it struck a pedestrian.Onscreen : Consider the “Trolley Problem.”Matthew Botvinik : There’s a trolley coming down the tracks.Onscreen : Towards 5 workers.Botnivik : And you have to decide whether you’re gonna pull the switch and move it onto other tracks.Onscreen : Towards just 1 worker. What do you do? Nothing? And let 5 people die? Or take action? And kill 1? Now imagine coding this decision-making directly into driverless cars.Edmond Awad : This is a big challenge.Onscreen: The dilemma of life and death remains, but now a computer is at the helm.Awad : So, for one example: we want always to give preference for humans over animals, and you try to apply this to every situation that you come across. The car can identify the object as human and the other as animals to avoid the object that is human even if that means it pays the cost of sacrificing the animal.Onscreen : This is a “top-down” approach. There’s also the “bottom-up” approach.Awad : You can imagine on one side you have a big truck and on the other side you have a cyclist. The car could choose to become a little bit closer to the cyclist or could choose to be closer to the truck. Now in each of these two situations, we are introducing more risk to one side.Onscreen : In a “bottom-up” approach, a car might make case-by-case decisions based on optimizing things like passenger comfort or manufacturer liability or observing preferences of human drivers. Most likely, car designers will opt for a top-down/bottom-up combo.Awad : By training the cars on different examples with some feedback about how the car should resolve this kind of scenario.Onscreen : But it’s not easy. What if a trained car determines that to minimize loss of life the driver must be sacrificed?Awad : While most people approve of self-driving cars, they’re not willing to buy self-sacrificing cars.Onscreen : And what about that Trolley Problem? Most people claim they would pull the switch to kill one person if they could save five. But most of the time, driverless cars won’t have to make these kinds of decisions.Awad : Self-driving cars will be minimizing accidents. You’re actually minimizing risk to your life.Onscreen : Today, 94% of accidents are caused by human error. A programmed car can’t get drunk, angry, tired, or distracted.Awad : So, every day that that this adoption is being postponed, that means loss of more lives.Botvinik : The future of AI technology is not just an engineering story, it’s also a story about how we decide to design our societies. We have to continually decide what kind of world we want to live in.

How do we program moral decision-making into autonomous vehicles?

Ethics and Self-Driving Cars

How should a self-driving car behave? It’s not as simple as you might think.

Can Autonomous Cars Learn to be Moral?

There’s a lot going on ‘under the hood’ of a self-driving car.
To match human drivers, an automated vehicle needs to learn how to handle three distinct tasks: seeing everything around the car, understand what it’s seeing, and planning the car’s path.
Three Tasks Driverless Cars Need to LearnPublished: October 18, 2019Chris Gerdes: Trying to develop an automated vehicle that can do everything that a human driver can do is a huge problem, and it requires an awful lot of interlocking pieces. Narrator: To even match human drivers, a self-driving car needs to learn to handle not just one, but three distinct tasks nearly flawlessly. The first: seeing everything that’s around the car; second, understanding what it’s seeing; and third, planning the car’s path and controlling its acceleration, braking, and steering along the way.First, seeing: most self-driving cars rely on cameras, radar and LIDAR. Each has weaknesses. Cameras work poorly at night, radar doesn’t distinguish very well between different types of objects, and LIDAR can fail completely in rain, snow or fog. But even when sensors are working nearly perfectly, there’s still the second big challenge: understanding what they are seeing.For us, taking in sensory information and making best guesses about the real world are second nature. For a computer, making meaning out of visual data is fiendishly difficult. In each case, it’s the task called “perception.”Amnon Shashua: The human brain is an expert in perception, but for machines, it’s not natural. So this is one big challenge, that if you want to drive autonomously, you need to perceive the world just like humans do. Narrator: Until recently, no computer could come close, but since 2010, a big leap forward in a branch of A.I. called “machine learning” has come a long way to closing the gap.Daniela Rus: Machine learning, it’s about, uh, giving machines the ability to look at data, identify patterns and make predictions. Narrator: Even if the car can see and make sense of its surroundings with unmatched perfection, it could still be a lethal hazard if it fails to handle its third crucial task: planning; yet another daunting challenge, because the planning software has to anticipate what’s likely to happen, then plot the car’s pathway and speed, ready to change both in a split second if the sensors spot trouble.Just like the car’s perception software, its planning software also needs to be trained.

Three Tasks Driverless Cars Need to Learn Hero

There’s a lot going on ‘under the hood’ of a self-driving car.

Three Tasks Driverless Cars Need to Learn

<section class="entry-content instapaper_body clearfix" itemprop="articleBody">
<div class="add-container clearfix" id="add-block-placer">
</div>

 Christie Wilcox, writing at Science Sushi:
 
<blockquote>

 Long wings are a benefit in certain environments. Longer, pointier wings provide increased lift and less drag, making it easier to spend lots of time in flight — a huge plus if, like the cliff swallows, you migrate long distances seasonally. A lot of birds, including most swallows, have long wings for their body size because they spend a lot of time catching flying prey like insects, and longer wings reduce the effort needed to glide and maneuver in the air. But while short wings are less efficient once in flight, they provide a bigger boost from the ground, allowing birds a near-vertical takeoff. When you’re trying to jump out of the way of a speeding car, the faster you can get out of range of the vehicle, the better.
 

 Not only did the team find that the roadkill birds had longer wings, they found that over time, the overall wingspan of the cliff swallow population in Nebraska has been declining. While it’s possible that other factors are influencing this change, the research pair think the selective pressure to avoid cars is the real driver.
 
</blockquote>
</section>

Cars Driving Change in Cliff Swallows

Driverless cars are being tested in private and public spaces.
To test self-driving cars without putting lives at risk, most companies do a lot of their training in environments they can fully control. But ultimately, the only way to know how a car will perform on public roads is to test it in the real world.
Testing Self-Driving Cars in the Real WorldPublished: October 23, 2019Onscreen: To test self-driving cars without putting lives at risk, most companies do a lot of their training in environments they can fully control, like this private training facility in Northern California.It’s run by Waymo, the company that Google spun off to develop self-driving cars.Stephanie Villegas: And so, because the site has so many different types of roads, from residential to expressways to arterial roads going between the two, cul-de-sacs and things like that, we’re able to stage, basically, an infinite number of scenarios that you would encounter on those types of roads in the real world.Tony Kum: Great! Let’s go for run, three, two, one…Narrator: Today’s test: on a street where the view is blocked, a car suddenly backs out into the Waymo car’s path.Kum: Great job, everyone.Villegas: We can go a little spicier. We can change the speed at which the auxiliary vehicle exits the driveway. Does it rip out of its driveway like a bat out of hell, really coming out of the driveway ahead of the Waymo vehicle? Or is it slowly kind of meandering down the driveway and taking its time?Kum: Great job, guys. Let’s restage and stand by.Narrator: Next, the Waymo car has to handle what’s called a “pinch point,” a two-way street made narrow by parked cars.Gunjan Yagnik: Waymo rolling.Villegas: You encounter this really often on public roads, two oncoming vehicles have to negotiate who will assume the right of way and who will have to yield.Narrator: The other car arrives at the pinch point first, so the software should tell the autonomous car to yield.Yagnik: Yielding, yielding right here. Butter going around. [sings] Butter…Narrator: These are just two of the thousands of scenarios the company uses to train the software, and they introduce new ones all the time.But ultimately, the only way to know for sure how a car will perform on public roads is to test it in the real world, where the stakes for wrong decisions are much higher.Shai Shalev-Shwartz: We don’t just, “Let’s go for a test drive just for the, for the fun of it,” because a test drive by itself may be dangerous, okay? But there are things that are very difficult to check without actually doing a test drive. And these are things that involve negotiation with other drivers, because in order to check if you are negotiating normally and properly with other drivers, you need other drivers.Narrator: Today, Mobileye engineers are testing a part of their planning software that handles merging in traffic.Amnon Shashua: So, so we merged fine, but at the last point we’re a bit slow.Merging into traffic, this multi-agent game that we are playing, requires sophisticated negotiation. You are negotiating. Your motion signals to other road users your intent, so you are negotiating. And this negotiation requires skills, and those skills don’t come naturally. Those skills need to be trained.You need to now change two lanes, because the exit is a few hundred meters from us.Narrator: The software is trained to be assertive when it has to be. Here, the car signals its intention to exit by speeding up so it can merge.Shashua: And you saw that we changed two lanes without obstructing the flow of traffic and there are many vehicles here. You need to provide agility that is as good as humans, if you want to be safe.

Testing Self-Driving Cars in the Real World Hero

Driverless cars are being tested in private and public spaces.

You searched for: cars