Text Size AAA
Sep 9, 2014

Energy Implications of Autonomous Vehicles: Imagining the Possibilities


Autonomous vehicles have captured people’s imaginations for decades. At the 1939 World’s Fair in New York, GM’s pivotal Futurama exhibit presented its vision for 1960s autonomous highway infrastructure to 30,000 visitors a day. Two decades later, during the construction of the interstate highway system (largely based on Futurama, sans vehicle autonomy), the Central Power and Light Company, an electric utility, employed the autonomous highway dream in a newspaper advertisement to demonstrate the vital role power companies could play in our driving future.

Now autonomous vehicles are no longer a utopian dream, capturing the attention of many, including some of my colleagues at RMI. Google recently made headlines by announcing it has started to manufacture its own autonomous car prototypes that lack steering wheels. Almost every major automaker is investing significant R&D capital in vehicle autonomy, including Audi, BMW, Ford, General Motors, Honda, Infiniti, Lexus, Mercedes-Benz, Nissan, Tesla, Toyota, Volkswagen, and Volvo. It’d be a shorter list to note which companies aren’t working on autonomous vehicle technology.

Although fully autonomous cars aren’t yet commercially available, vehicles with autonomous features are already on the market. Self-parking features have been available since 2007. The 2014 Mercedes-Benz S-Class can drive itself on the highway—as long as the driver’s hands remain on the steering wheel, the car maintains its lane and accelerates and decelerates according to speed limits and the locations of surrounding vehicles. A San Francisco start-up, Cruise, manufactures a kit that allows Audi A4 and S4 owners to drive autonomously on the highway. These semi-autonomous cars will likely be the first stepping stones to fully automated driving. Nissan and Google both expect to market fully autonomous cars early next decade, and they’re not alone.

Experts vary widely in their predictions of when fully autonomous vehicles will become ubiquitous—some say all cars on public roads will be autonomous by 2030, while others say the transition will take more than thirty years. We at RMI believe that widespread deployment of autonomous vehicles is inevitable within the 21st century—and will likely occur sooner than many anticipate. But will autonomous vehicles help or hinder our transition to a less carbon-intensive economy? Some fear that autonomous cars will bring about a more carbon-intensive world due to people (or, more accurately, their cars) driving more and using less public transportation.

Additional Driving

Autonomous vehicles could encourage additional driving, leading to more energy use and more emissions. If you’re able to work, nap, or read the newspaper on your daily commute, why bother choosing to live near the workplace? Autonomous cars could give rise to an “exurbia,” a new development layer farther from urban centers that could undo years of progress on smart, sustainable urban growth and transit-oriented development. Further, many people today make transportation decisions by optimizing based on both time and money. Autonomous vehicles could eliminate the need to optimize based on time, leading to more weekly supermarket trips, for example.

Less Use of Public Transportation

Public transit commutes today have the advantage of being passive, not active, letting travelers work, read, and sleep. But if your car can drive you, public transit’s productivity edge over car commuting is removed. And autonomous cars could alleviate traffic congestion, too, removing another benefit of public transit. Today human error causes over 90% of accidents. When human error is removed from the equation, the number of accidents is likely to decrease significantly, eliminating a major source of congestion. Additionally, autonomous vehicles can cope with much smaller following distances, permitting traffic flow where gridlock exists today. Connecting the dots, it’s not too difficult to imagine how autonomous vehicles could lead to the decline of urban public transit.

An autonomous car future could be pretty bleak. Fortunately, I believe our future is much brighter. Regardless of how they’re deployed, autonomous vehicles could permit emissions reductions by enabling more-efficient driving patterns, reducing drag, and actually increasing the use of public transit by solving the first/last mile problem. In autonomous fleet systems—for personal or commercial transit needs—the efficiency benefits are compounded through reduced operating costs, quicker payback periods, and increased asset utilization.

Vehicle Automation Benefits

Constant accelerating and braking greatly lowers fuel efficiency. Even the most efficient drivers are simply not as good as a computer at hypermiling. Autonomous cars can consistently drive far more efficiently than people can, even if you switch your Nissan LEAF or Honda Civic into “eco” mode.

These benefits are compounded if vehicles are able to communicate with each other (V2V) and surrounding infrastructure (V2I). While cruising towards an intersection, your car could intelligently coordinate with other vehicles and infrastructure to time its passage perfectly, avoiding significant waste in energy and time.

On the highway, around twenty percent of a vehicle’s gasoline is burned just to combat aerodynamic drag. Today, safety concerns limit following distances, requiring each car to make its own aerodynamic “puncture” through space. Autonomous vehicle communication permits closer following distances, allowing multiple vehicles to take advantage of the same “puncture.” So-called “platooning” is similar in aerodynamic principles to an archer’s arrow.

Trucks traveling through the Australian outback already realize many of the benefits of platooning—one cab can transport four tractor trailer-sized storage compartments in what’s dubbed an Australian Road Train. Tests conducted by SARTRE in Germany have successfully platooned several communicating Volvos with four meters of separation, and further testing could permit cars to get even closer.

Remember my earlier concern about autonomous cars potentially wiping out public transit? Well, the opposite could happen, too. If implemented correctly, autonomous vehicles could lead to increased use of public transit by solving the first/last mile problem. If you can coordinate your arrival at a subway station with train schedules and public transit is cheaper than driving, you may choose to commute via train. Public transit could carry additional benefits for longer-distance trips—faster speeds permitted by high-speed rail, no wasted time for bathroom stops, access to compartments with ready-made food and drink, and more.

Fleet Benefits

At least in urban areas, autonomous vehicles would likely arise in a fleet system, similar to how Zipcar, Car2Go, and other car sharing services work today. Autonomous cars will have lower operating costs per mile, achieving better fuel economy, reduced wear and tear on brake pads and the engine, and sophisticated matching of available taxis to consumer demand.

Inherent economic differences of fleet vehicles compared with individually owned vehicles unlock additional efficiency improvements. For personal and commercial fleets, upfront capital costs are a small portion of lifetime ownership costs. In particular, upfront costs are less than ten percent of lifetime costs for commercial fleets similar to those owned by Comcast, FedEx, and the U.S. Postal Service, which still operates a vehicle fleet from 1987. Such companies have a vested interest in prioritizing operating costs over upfront capital costs.

This means that fleet-based autonomous electric cars have several advantages over fleet cars that are only autonomous, only electric, or neither. Expensive battery packs hinder widespread electric vehicle adoption today, and Google’s autonomous cars rely on $80,000 spinning laser LIDAR systems. But dramatic decreases in operating costs could make the increased upfront investment worthwhile, significantly shortening payback periods. This concept is already employed when individuals consider purchasing a hybrid—if a hybrid hypothetically costs $5,000 extra, but saves $1,000 a year, that’s a simple payback of just five years.

Further, shared fleet vehicles are likely to travel farther each year, since today’s vehicles sit unoccupied over 90% of the time. Increased vehicle miles traveled (VMTs) per year further shortens payback periods, important when considering the time value of money. This is why 80 percent of new New York City cabs are hybrids.

Because operating costs become so much more important, previously uneconomic efficiency improvements could be realized. Expensive carbon fiber, championed by engineers due to its strength despite its low density but hindered by high costs, could finally comprise the core of new cars due to the dramatic efficiency improvements that result from lightweighting. Today, few manufacturers employ carbon fiber for weight reductions simply because it is too expensive and results in long payback periods. But increasing VMTs per year, compounded by the increased emphasis on operating costs, could change that.

Increased VMTs per year that arise in a shared fleet model carry another benefit: technological turnover. As fleet car components would wear out more quickly than their individually owned counterparts, fleet cars could more quickly take advantage of technological advances. Instead of being replaced every ten years, batteries might be replaced every two, enabling fleet owners to quickly reap the benefits of new battery chemistries, advances in energy density, increased range, and higher cycle tolerance.

Car sharing, too, is simplified in a fleet ownership model. Because individuals would rely on software to hail an autonomous vehicle, it’d be relatively easy to agree to share a ride with a stranger in exchange for lower fees. Uber, exclusively summoned via smartphone, is already piloting a similar project dubbed “Uberpool,” so it’s not difficult to imagine a similar scenario with autonomous cars. (Indeed, last year Uber received a $258 million investment from Google. Coincidence? You decide.)

Fleet vehicles also enable users to match vehicle and purpose, reducing waste. Are you headed to the grocery store? Order a car with enough trunk space to fit your haul. Headed home from Ikea? Order a larger car to fit the futon you just bought. Shared vehicle fleets already enable this, with companies like Zipcar including SUVs and pickup trucks in their offerings.

The Bottom Line

With autonomous vehicles, it’s not a question of whether, but when—and their impact on our energy use could either be incredibly good or incredibly bad, or somewhere in between. The way we get around could be significantly different in the next 50, 20, or even 10 years, and the environmental impacts of this new transportation economy will ultimately depend on several different factors. But I believe the autonomous driving future looks bright … and green.

Images courtesy of Shutterstock. Second image Dong liu / Shutterstock.com


Showing 1-10 of 10 comments

September 11, 2014

How sad, carbon pollution is the problem and we pretend the first new cost effective, three times more energy efficient engines since the 1700’s (ICE’s are just advanced steam engines) that runs on hydrogen (not carbon based fuel) isn’t the solution. GE announces that it has built a pilot fuel cell plant to supply fuel cells that are three times as energy efficient as diesel engines to replace commercial building emergency generators with peak shaving and combined heating and power fuel cells but there’s no excitement. Figure out that these fuel cells can be configured to be fueled directly with say, cellulosic ethanol (100% renewable) from one of the proposed 40 planned cellulosic ethanol plants and yet still no excitement. Then figure that GE may learn to displace train engines with these fuel cells still fueled directly with cellulosic ethanol but still no excitement. And then acknowledge the first very exciting fuel cell car https://www.youtube.com/watch?v=RqLpqR0SPnQ with 920 hp, that goes 215 mph, accelerates from zero to 60 mph under 3 seconds while getting three time the mileage, and when mass produced would cost about the same as today’s cars of the same capacity but still no excitement. No, all the excitement is about a much dumber than human car (doe doe bird car) that may drive itself someday knowing that hackers will figure out how to make them run into each other but nevermind that.

September 11, 2014

Perhaps I missed this in the article, but there is a likely severe downside to autonomous vehicle operation. People will be sending their cars here and there with no one in them. Or to further such nonsense as delivering their kids to schools like little packages. Sort of like abuse of remote starting functions. I expect the wealthy to be more likely to avail themselves of this "opportunity", and I suspect it will be difficult to regulate it.

For better or worse, autonomous vehicle operation will put a whole class of employees out of work. Truck drivers, taxi drivers, etc. Society needs to start planning now for this employment disruption.

September 11, 2014

Personally, I think that one of the largest secondary benefits will be the ability for EMTs and first responders to more easily navigate through traffic to get to fires and deliver patients to emergency rooms more quickly. Just think of the lives saved.

Google User

September 11, 2014

In "Energy and Autonomous Urban Land Vehicles" (IEEE Technology and Society Magazine, Summer 2012) I pointed out that the possible benefits of self-drive go way beyond the obvious. If we do get to a point where almost no humans are doing the driving, safety increases to the point where a motorcycle is almost as safe as an SUV. Instead of using 4,000 pounds of vehicle, the vehicle can weigh less than the riders. Coupled with efficient operation, energy savings can increase by an order of magnitude beyond what we currently consider highly efficient. A light, aerodynamic urban vehicle can be powered by a 25 pound battery (using current technology). A battery that light can be swapped by robot in less time than it takes to fill a gas tank. Electric vehicle range anxiety goes away. With a large bank of batteries needing recharge, we can harness energy whenever the sun shines or the wind blows.
Yes, public transportation as we know it will disappear. Private transportation may also disappear. If the backbone of urban transportation is automated taxis, there is little reason to own a car. The distinction between public and private transportation blurs. A platoon of self-driving taxis to handle the passengers on a bus will have about the same capital cost as the bus. Since the main transit operating cost is driver salary, public transportation can make money instead of requiring subsidies. See "The Modular Bus" in Traffic Technology International, Aug/Sep 2013.

September 11, 2014

@Steve, good point. The "additional driving" section addresses that a bit. When you remove the need to optimize based on time, sending your car to pick up groceries, drop kids off at school, grab your lunch, and more becomes possible, and really does have the potential to increase emissions through increased vehicle miles traveled (VMTs).

But consider potential sources of decreased VMTs: increased ease of carsharing or sending one car to run errands for many people a la Google Shopping Express (https://www.google.com/shopping/express/). These definitely have the potential to negate some of the possible VMT growth. Further, if AVs significantly increase the rate of adoption of EVs, lightweighting, and more, significant VMT growth wouldn't necessarily mean additional emissions.

And you're right, society must begin planning for shifts in employment demand, though the pace of AV adoption will determine the urgency.

September 11, 2014

@Carl, this post doesn't delve into the arguments for or against hydrogen vehicles, and I won't get into this much here. Personally I have significant reservations when considering the viability of hydrogen as a transportation fuel, principally due to the high cost of platinum that would only increase with increased demand. I also believe battery technology (with significantly higher round-trip efficiency than hydrogen) will rapidly improve, decreasing charge times and cost and increasing density, alleviating range anxiety concerns.

No, AVs don't have human intelligence. But that can be a benefit. Computers would have better reaction times, for one. Computers can quickly process information from other vehicles and decide to steer, brake, accelerate, etc more quickly than a human.

Good point regarding the security implications of autonomous vehicles. I'm not a computer scientist so my technical expertise on this is limited, but today airplanes are in autopilot much of the time, and our military effectively controls UAVs on the other side of the globe. Whatever security concerns exist, I'm sure we'll be able to address them in time.

September 12, 2014

@JB - more rapid response to emergencies is definitely a huge benefit. Plus, there'd hopefully be fewer accidents in general without human error.

@Tyler, thanks a ton for your comment and pointing me towards your articles on the topic. I love the prospect of using battery banks to provide grid support during charging.

September 18, 2014

Good point by Steve Grant: A Vehicle Mile Traveled with an empty car is still a VMT, and uses fuel. So the real question is whether autonomous cars will increase fuel consumption, or decrease it, and whether the effect will be significant.

I suspect the answer is that they will slightly decrease fuel consumption, because people will feel less possessive of a car that you "call to pick you up." It will feel more like a taxi ride than a typical early-21st-century car commute.

But in the larger picture, the effect of autonomous cars is likely to be dwarfed by other considerations.
The sprawl of the late 20th century was a unique event, with racism ("I don't want to live in a mixed neighborhood!") intersecting with massive federal earmarks (urban freeways, which even Eisenhower opposed) and cheap gasoline, to create a perfect storm of sprawl. It's not likely to be repeated. Already, freeway traffic jams are limiting sprawl, and urban living is "cool" again. So sprawl has reached the point where it is "self-limiting" itself, and by extension, limiting Passenger MT.

On the other side, the more alternative fuels catch on, the greater the pressure will be to move toward a VMT tax, which is really just code for "toll roads." This will result in a higher visibility of the variable costs of driving, which will encourage more ride-sharing, and further act to decrease total VMT.

September 19, 2014

What I looked for in this article and didn't see was the impact of autonomous cars on urban parking. Once cars drive themselves, it's like a limousine service: You are dropped at your destination and the car parks itself based on what is cheap and least congesting. However, where huge numbers of people are gathering or leaving, say, a sports event, the lines of cars coming to the closest drop-off/pick-up points will be like a private elementary school on steroids.

April 20, 2015

"However, where huge numbers of people are gathering or leaving, say, a sports event, the lines of cars coming to the closest drop-off/pick-up points will be like a private elementary school on steroids."

I don't agree that there's necessarily a big problem. Suppose the entrances to the stadium were serviced by buses that then took the passengers to separate lots where cars pulled up to take away the bus passengers. Then, instead of 50,000 cars pulling up to the entrances to the stadium, there would be only 1000 buses (or even 500 buses, if the buses could fit 100 passengers each).

Then, at those satellite parking lots, there would only be 50-100 cars pulling up to each bus. It would actually be tremendously more efficient to fill and empty a stadium in that way than in the present way (where people have to walk to 20,000+ cars in massive parking lots around the stadium).

And I can imagine that, for example, large restaurants that might be several miles from the stadium would welcome buses to drop people off in their restaurant parking lots (to be met by autonomous cars), because that might generate more business for the restaurant after the games.

Again, I think it will be much more efficient to fill and empty stadiums with autonomous vehicles than without autonomous vehicles.

PAGE: 1 
Show Subscribe