Less is more.
Next: Autonomous vehicles receive a lot of coverage, as does electric mobility. But rarely are the two analysed in combination. Why did you choose to look at the disruptive potential of both technologies together?
Greenblatt: Most people don’t realize the synergies when you combine electric vehicles (EV) and autonomous vehicles (AV). Both are enabling technologies that amplify each other to achieve broader adoption. It’s an inevitable trend. Autonomous vehicles can use any drive train and their appeal is not about lowering emissions, energy savings or even cost savings. This technology is really about offering a better service. Electric vehicles are a completely different conversation, largely about the environment. But it’s still a hard sell to convince the majority of the market to buy an EV because it has to be as cheap or cheaper, and has to have at least the same performance of a conventional gasoline-powered car. When you look at both technologies in combination, they show enormous disruptive potential in terms of energy savings and environmental benefits.
What did you discover happens when the two come together?
Saxena: It’s a sea change. We ran simulations for the year 2030 and found that if you combine electric and autonomous vehicles in the form of robotic e-taxis, emissions per passenger mile are 90 percent lower than a conventional vehicle. Why is this effect so dramatic? One reason has to do with the shift to electric drive, which is an economic consequence of taxis driving much further in a year than a typical car—electric vehicles are more expensive to purchase, but cheaper per kilometre travelled. The second reason is “rightsizing,” meaning that an autonomous vehicle can be smaller, even down to a one-seater since you don’t need a second seat for the driver. If someone is taking a trip alone, a robotic one- or two-seater can be dispatched instead of a giant, empty car. That’s crucial because almost 90 percent of rides today carry only one or two passengers. We think that one-seaters will comprise two-thirds of all these vehicles. Finally, there is a third reason driving the drastic drop in emissions. Unlike gasoline or diesel, electricity will become a lot greener over the next few decades. If you combine those three factors, greenhouse gas emissions per mile driven drop by a factor of 16 compared to today. That’s a big deal.
' Recharging is markedly simpler, because it is left to a robot. '
AVs are at the experimental stage and EVs have only garnered a small part of the car market. How will we get to millions of autonomous, electric vehicles?
Saxena: Market data shows that the adoption rate of EVs is on par if not slightly higher than it was for hybrids. They will gain marketshare. Autonomous vehicles can be an enabler for more EVs. You don’t even need to have an autonomous vehicle for this effect to happen, you simply need the vehicle to be shared. But it’s far more likely a car will be shared if it’s autonomous because this technology gets rid of the driver’s responsibilities. While it’s true that shared rides make up only a small percentage of driving today, we think AVs can break down those barriers in important ways. For one thing, recharging electric batteries becomes much less of a hassle when it involves a robot’s time instead of a human’s. A well-managed fleet of EVs can also optimize which vehicles are dispatched for each trip, so the limited battery range becomes less of an issue. Also, when you share a car, you tend to drive it a lot further in a year than a privately owned car. The average car in America drives around 12,000 miles a year, whereas the average taxi drives anywhere from 40,000 to 90,000 miles. When you’re driving five times more in a given year, an alternative fuel vehicle — be it with hydrogen fuel cells or an electric battery — becomes the most cost-effective option. It is the most economical choice for a taxi platform.
That positive effect, you argue, holds true even without looking at autonomous driving…
Greenblatt: Exactly. So far, there is a big trade-off between a gasoline versus electric car. But the autonomous piece is critical because it enables a car to drive much more. An autonomous vehicle fleet manager will choose what lowers his or her costs, including electric vehicles. Even though you have the limited battery, range, and long recharge times, it doesn’t matter. You can still serve customers because those cars don’t need to run around the clock. For 70,000 miles per year, you only need to travel 200 miles per day. At an average speed of 20 miles per hour, ten hours per day are sufficient, and that leaves plenty of time for recharging. They can also make several trips before they have to stop to recharge because urban trips tend to be quite short. In New York, each ride is about 2.5 miles, and a typical U.S. trip is about ten miles. If the battery range averages 60-80 miles, you can still fulfil quite a few rides before you have to recharge.
People still want to own and drive their own cars. How many of today’s vehicles do you expect to be replaced by automated e-taxis?
Saxena: Knowing how many autonomous cars there will be in the future is hard to predict. We modelled it with two scenarios for 2030. If only five percent of all new vehicle sales were autonomous taxis, it would come to around 800,000 cars per year, which is quite a conservative number. The higher number we used was 10 percent of the 2030 vehicle fleet, or about 25 million cars, which is not inconsistent with what some analysts predict. Owning a personal car has had a lot of benefits, but those advantages are eroding because autonomous technology makes a shared vehicle much more appealing, convenient and cheaper to use.
Can you describe some advantages of a shared autonomous fleet?
Saxena: You have a car that’s there when you need it, on both ends. You don’t have to worry about where to park. It’s always new, clean, fuelled and ready to go. Oh, and by the way, you don’t have to drive it, you can do other things. When used as a taxi, you can also take the driver’s salary out of the mix. Even when you factor in expensive autonomous technology, which runs as much as $150,000 today, it still costs less in total operating costs. And there’s something else to consider. Some experts estimate that up to one-third of a city’s area is used for parking. If parking is freed up, it can transform cities. You could widen sidewalks, put in cafés, street gardens and trees.
Greenblatt: It does require a shift in how we’re looking at mobility. This is not a scenario where you personally are choosing between purchasing a regular car or an autonomous vehicle. If I’m willing to accept shared mobility, an expensive autonomous vehicle makes a lot of sense economically and environmentally. There is already a mind shift happening in urban areas where people choose not to have their own cars. It’s also a generational shift because young people are not as attached to the idea of ownership.
What consequences do your projections have for car manufacturers?
Greenblatt: Car companies are already trying to move from selling cars to providing access. All of their shared fleets will probably need to evolve, autonomous or not, to the point where they offer a variety of sizes. Booking a car will become more and more convenient until it’s essentially an instant decision.
' Owning a car of your own has had many advantages up to now, but these are gradually disappearing. '
A key part of your model depends on “rightsizing” the electric, autonomous cars of tomorrow. What does that mean for vehicle design?
Saxena: Rightsizing is the idea that a car is appropriate for each particular trip. There are large barriers to adopting small vehicles when they’re privately owned, because very few people will choose a one-seater unless they live like a hermit. Nevertheless, smaller cars are already gaining market share. When you introduce autonomous vehicles into the mix, you don’t need the driver, so the car can be even smaller. The other important factor is higher efficiency through predictive technology. If your car knows what type of ride and terrain to expect, your peak power and peak torque capabilities can be lower. You don’t need such a powerful engine and battery for a small trip in the city. And if cars don’t even crash, they can be very light and realize even bigger energy savings.
Greenblatt: If you suddenly don’t need the car to travel even 80 miles, you design differently. You might also be focusing your battery research differently. If cars are used much more heavily, components need to last longer. A new car will go 350,000 miles in five years, whereas most current cars don’t go half that far in 15 years. That alone will drive a rethinking in car design.
To what extent does it matter whether the electric drive is powered by a battery or fuel cells?
Greenblatt: Fuel cells are harder to model from a cost and environmental perspective, because the varied ways in which you can make hydrogen. Given our data sets, EVs performed better. Fuel cells were an interesting intermediate technology but were surpassed by an EV once you get beyond 70,000 miles per year. Fuel cells currently have the advantage of a rapid refuelling time. But in the case of an autonomous vehicle, especially if it’s shared, is that really a big deal?
Your time horizon for this shift to happen is 15 years. Why did you pick 2030?
Saxena: It’s a consensus number looking at what manufacturers are doing and announcing. Autonomous cars are very close to being a commercial reality by 2020. Give it another ten years, and this will be pretty routine technology. Even if those cars are expensive, if you deploy them in a fleet they make sense. And if fleet operators buy them at regular intervals and in large numbers, the transition will happen faster than we think, driving new generations with constant upgrades. We are careful to say that not all cars will be autonomous by 2030, but by 2050 it’s possible that almost all cars could be autonomous, with the majority being part of a fleet.
If shared mobility becomes as cheap, ubiquitous and convenient as you say, won’t people ride more and thereby cancel out the environmental benefits?
Greenblatt: We looked at this scenario in two ways. What if people choose autonomous trips but aren’t rightsizing the car? Even if the vehicle was two sizes larger than needed, emissions would grow but still not cancel out the 90 percent reduction per mile. Second, what if we all drive more because it’s so convenient and more people who have not been able to drive can suddenly get around without a license? It turns that even if we increase the number of miles travelled per person fivefold from today, we’d still cut energy use and emissions because of the benefits of rightsizing and electrification.
One aspect missing in your models is the role of public transport. Where do buses and subways fit in?
Saxena: What if in the future public transportation was all autonomous vehicles, and the transit agencies were the fleet operator? AVs could help or hinder public transit, but that might not be a bad thing in either case because what we’re really trying to do is get the most people moved for the least amount of energy. So far, public transit has been a good solution, but buses and trains are pretty inefficient if only few people are riding them. The nice thing about a fleet is the engines only run when there are people in them. It’s more modular and more efficient.
On the brink of extinction: Only if future fleets of self-driving cars take actual user demand into account will emissions decrease significantly.