clock menu more-arrow no yes mobile

Filed under:

The transformative potential of self-driving electric cars

If you buy something from a Vox link, Vox Media may earn a commission. See our ethics statement.

The Toyota i-road, a single-occupant urban electric vehicle concept.
The Toyota i-road, a single-occupant urban electric vehicle concept.

Public discussions about electric vehicles, self-driving cars, and the future of transportation seem weirdly circumscribed to me. The value of alternative vehicles always seems to be judged against their ability to swap out for today's internal combustion engine (ICE) vehicles.

(By the way, a massive new report from the Electric Power Research Institute finds that swapping out ICE vehicles for electric vehicles would bring enormous benefits indeed, in both greenhouse gas reductions and local air quality.)

But viewing electric vehicles and self-driving cars simply as substitutes for ICE vehicles misses the forest for the trees. The enormous bulk of the potential lies not in replacing units in today's transportation system, but in generating new systems. It is the system benefits that are at once most difficult to predict and most rich with possibility.

Just as in the power sector, in transportation we are about to witness the collision of an irresistible force (new technology) and an immovable object (existing infrastructure). Political realism and lack of imagination — often one and the same — can prevent us from thinking and talking about what kind of world might be possible if the immovable object could be ... moved.

So I want to get a little utopian. Autonomous electric vehicles (AEVs) have the potential to completely transform urban landscapes, which in turn can transform how those who live in cities interact. It's heady stuff. Let's get excited!

Traditional cars are overengineered; AEVs can right-size

Back in 2010, I reviewed a book called Reinventing the Automobile: Personal Urban Mobility for the 21st Century. It is super, super nerdy, written by engineers, but it opened my eyes to one of the key advantage electric vehicles have over traditional ICE cars. Here's how the authors put it:

A traditional [ICE] car requires elaborate systems of reservoirs, tubes, valves, and pumps to distribute the gasoline, oil, water, air, and exhaust gases, but a battery-electric automobile replaces most of these complicated distribution systems with wires connecting the batteries to the wheels.

The basic system for electric mobility is simpler than an ICE engine, by a wide margin. For an electric vehicle, you need four components:

  1. A wheel or wheels, each containing an electric motor
  2. A platform of some kind, to carry a battery and the passenger(s)
  3. A user interface
  4. Wires to connect 1 through 3
basic electric vehicle platform

A basic electric vehicle platform, from a company called Trexa. This one's got the electric motors in the central tube, which allows different suspension/wheel combinations to be swapped out.


That's it. What becomes clear when you think about it this way is that there is an almost unlimited array of form factors an electric vehicle could take. It could have one or two wheels with a gyroscope (e.g., a Segway), three wheels, four wheels in a diamond shape or the traditional rectangle, six wheels, retractable wheels. It could be a modular vehicle on which wheels can be added or removed. It could be the traditional long, low chassis, an upright chassis, or any other shape or orientation of chassis you can imagine. The user interface could be a touchscreen, a joystick, pedals, levers, or, in the case of a self-driving car, nothing.

Electric vehicles are intrinsically more efficient than ICE vehicles:

Electric vehicles convert about 59%–62% of the electrical energy from the grid to power at the wheels—conventional gasoline vehicles only convert about 17%–21% of the energy stored in gasoline to power at the wheels.

But they are more efficient in another way as well: Their simplicity and flexibility mean they can be right-sized, designed for a fine-grained array of transportation uses.

Transportation infrastructure is overengineered; AEVs can right-size

ICE vehicles are designed for peak use — driving long distances at high speeds. And because all ICE vehicles are overengineered, they're also big and heavy, which means they must be highly up-armored against collisions from other big, heavy, fast-moving vehicles. All that armor makes them even bigger and heavier. And so on.

Basically, we're all driving around in quasi-military vehicles, designed to go 100 miles an hour for 300 miles. Yet most of the time, we drive to work and the grocery store, slowly, in traffic. Actually, scratch that: About 95 percent of the time, we aren't driving at all; we leave our cars and trucks sitting, parked.

Just like ICE vehicles themselves, the urban transportation system of roads and parking is designed for times of peak use. Parking is scaled for maximum traffic — covering up to one-third of urban land in some cities — which means most spaces are sitting empty most of the time.

parking lot

A dumb use of valuable land.

(Flickr: Matt' Johnson)

Self-driving vehicles open up different options. And since we're in utopian mode, let's not think about how current urban infrastructure could be repurposed; instead, let's imagine what urban structure might be built from scratch around AEVs.

The signal direction of modern technology is to replace machinery, infrastructure, and commodities with intelligence, i.e., with computing power, which is on an inexorable march downward in cost. If vehicles become much smarter, it enables a concomitant reduction in machinery, infrastructure, and commodities.

Begin with the fact that AEVs would radically reduce (and eventually all but eliminate) accidents. If AEVs were piloting through a landscape of slow-moving AEVs, bicyclists, and pedestrians, there would be far less need for up-armoring against collision. That means AEVs could be far, far smaller and lighter, increasing their range and reducing their land-use needs, their impact on surfaces, and the environmental impact of their manufacture.

Consider how AEVs might transform transportation into a service. Imagine a fleet of shared AEVs, of varying sizes and purposes, that can be summoned with a smartphone app — dare I say it, an Uber, but for self-driving cars. (This is not Jetsons futurism; Tesla is already pondering it.) If right-sized transportation were available to everyone within minutes, there would be no need for every individual to own a vehicle that is parked 95 percent of the time.

When they are not in use, AEVs could drive themselves to specially designated parking garages, where they could be stacked much closer than human-piloted vehicles and plugged into the electrical grid, where they can serve as distributed energy storage. There would effectively be no need for any other parking, especially on-street parking, and there would be no distracted, frustrated human beings driving around looking for parking.

stackable citycars

Stacked citycars.


Small, lightweight AEVs vastly expand the possibilities for shared use of urban spaces. In 2010, hotshot architect Bjarke Ingels won the Audi Urban Future Award for his concept of a driverless city. Here's the capsule summary:

We imagine that inner cities that are currently banning cars through taxation or tolls to relieve congestion will simply become driverless rather than car-less. Driverless cars will combine individual mobility outside city limits with collective mobility within; as selfdriven cars move in coordinated concert with their fellow commuters, they occupy a quarter of the space human-driven cars require. As the new generation of cars will additionally be noiseless and pollutionless it will mean the end of the apartheid that currently separates cars from pedestrians and bicyclists for comfort, health or safety reasons. The result is an elastic urban space that can expand and contract to accommodate peak traffic hours or allow a park or plaza to invade the car lanes to fit the demands and desires of its citizens. Picture a future city in 25 years where the vertical facades appear unchanged, but the city pavement is transformed into a reprogrammable surface, replacing the fixed elements of driveway, sidewalk or square into a digital street surface that is completely re-animating a familiar city.

driverless city

The future is groovy.

Driverless City Concept © Bjarke Ingels | Audi Urban Future Initiative

In effect, the city would reconfigure itself in real time around changing patterns of land-use demands. It's like the Dutch concept of woonerfs [CORRECTION: I am informed that the plural of "woonerf" is "woonerven," not "woonerfs"; I apologize to the Dutch] — "living streets" shared by pedestrians, bikes, and slow-moving vehicles, devoid of traffic signs or lane dividers — taken to its space-age conclusion.

Also, it would look really cool.

driverless city

Driverless cities would presumably also have techno soundtracks.

Driverless City Concept © Bjarke Ingels | Audi Urban Future Initiative

This is all a bit science fiction-y for now, but the technologies to make cities smart are already in furious development. With current technology, or at least technology plausibly within reach, city infrastructure could be infused with sensors and communications that enable every AEV to know exactly where demand is high, congestion is developing, and the fastest routes are. There may be things that humans can do better than computers, but traffic is not one of them.

This world is also likely to feature less air pollution. It is true, as we're incessantly reminded, that EVs are only as clean as the electricity that fills their batteries. But the US grid is getting cleaner, and that process is expected to continue. (In fact, electrifying the vehicle fleet could help get more renewable energy on the grid faster, as it could serve, in effect, as dispatchable demand to soak up excess solar or wind energy.)

Transportation utopia is a long way off, for these dream-crushing reasons

Were one the type to crush utopian dreams, one could cite one of the following families of concerns:

1) Path dependence and the risk of more sprawl

The Chinese get to build futuristic cities from scratch, but in the developed world we no longer have that luxury. There's a whole bunch of built infrastructure already in place, and, especially in the US, it is designed for cars. Where there is built infrastructure, there are sunk costs and vested interests. That creates an enormous amount of path dependence, which will shape future developments.

That helps explain why many urbanists are suspicious of or even hostile toward self-driving cars. They worry that by meliorating some of the current problems with cars — accidents, congestion, pollution — they will sap any political will to solve the more fundamental problem, which is that any system that relies primarily on individual vehicles will prevent the kind of urban density that's desirable for environmental and many other reasons.

More prosaically, they point out that it will raise demand for individual vehicles, opening them to whole populations, including those under 16 and older people with fading vision and reflexes, to whom they are now inaccessible. And it will make for easier and more pleasant commutes, enabling housing to sprawl farther outward.

google self-driving car

Google's adorable self-driving car prototype.

(Google Self-Driving Car Project)

David Edmondson has a good rundown of the issue here and makes the point that AEVs are just as likely to transform suburbia as they are to enable it. No parking lots, no gas stations, no big eight-lane arterials, storefronts up against the sidewalk ... it's not very sprawly, because "this transportation cloud functions much more like the streetcars of the old days than personal cars of today."

My educated guess is that AEVs will make dense city centers more, not less, appealing, by enabling transportation as a service, radically reducing noise, air, and aesthetic pollution, and opening up new mixed-use urban streetscapes. But it's probably just too early to tell.

2) Big Brother

The discussion of smart cities has been dominated by technophiles and big corporations, which means the substantial privacy issues involved have not exactly gotten a full airing. But "smart" means information, and a city filled with sensors and trackers will accumulate a lot of information about every citizen within it. Who owns that information? Who can access it? How much will basic services like transportation hinge on the surrender of personal data? How will all that data be protected from the copious cybersecurity threats that face smart cities?


Look how smart this city is.


Focusing in on AEVs, there's the simple fact that humans will be surrendering control of the hundreds of second-to-second micro-decisions involved in driving. In many cases, those decisions have an ethical dimension. How will the AEV make them? What programming and algorithms are involved? Who decides how priorities are set?

Much like the internet, the connected transportation net of the future opens all sorts of utopian possibilities, but it is also subject to capture by corporations and nosy government surveillance types — who, the historical record shows, have a way of boxing out the utopians.