Published May 12, 2014
BY JOSIE GARTHWAITE
Driverless vehicles are coming as surely as winter is coming to the House of Stark. Automakers, technologists, and academics alike say it’s only a matter of time until self-driving cars hit the roads. Indeed, many autonomous features, such as lane switching warnings and adaptive cruise control are already available. Some forecasters say cars that are mostly or entirely autonomous could number nearly 54 million by 2035 (mostly in North America, Europe, and China) and take over the world’s roadways sometime after 2050.
Today, self-driving cars in research and development projects are making the leap from highways to city streets, where cyclists, pedestrians, school buses, construction barriers, and double-parked vehicles require more sophisticated software modeling. In Ann Arbor, Mich., work began this month on a new site designed to simulate a cityscape for testing driverless, or connected, vehicles. In Gothenburg, Sweden, Volvo is launching a pilot project that will see a fleet of 100 autonomous vehicles tested on public roads by 2017.
And Google, which has logged nearly 700,000 miles of autonomous driving since it began developing self-driving cars about five years ago, has refocused its efforts from highways to dynamic urban environments. First up are the relatively calm streets of Mountain View, Calif., but more complex, bustling areas will likely be tackled down the road. Google’s Self-Driving Car Project Director Chris Urmson explained in a recent blog post: “Thousands of situations on city streets that would have stumped us two years ago can now be navigated autonomously.”
As autonomous and semi-autonomous vehicles become a real choice for our neighborhoods, transit hubs, entertainment districts, and downtown cores, there is an opportunity for them to bring environmental benefits. But realizing those benefits will require us to take a hard look at how personal cars—even the driverless kind—fit into urban life.
Like bus rapid transit systems, in which dedicated lanes and priority at green lights help buses avoid traffic, self-driving cars raise questions about who has right of way on city streets.
Safety First
The key driver so far to develop autonomous vehicles and policies has been safety. After all, self-driving vehicles never get sleepy, enraged, impatient, drunk, or distracted, so they can avoid many of the accidents that are commonly caused by human error. But reducing traffic congestion, pollution, and energy use are often cited as important perks.
“Automation, in sort of big-scale thinking, is about making the city more efficient,” said Dave Miller, a doctoral researcher at Stanford University who is studying how in-vehicle displays can encourage more efficient driving behavior. But at this point, how autonomous vehicles will affect overall emissions from transportation remains uncertain.
In one scenario, this technology could help unlock a virtually driverless city by filling gaps in mass transit, facilitating greater density, and making it easier to move people and goods around with less energy and fewer cars. “On-demand origin-to-destination mobility could dramatically reduce the need or impetus to own a car in the first place and make life in the city much more appealing,” said Greg Rucks, manager of the Rocky Mountain Institute’s transportation practice. In the long term, he said, this could inspire people to leave the exurbs for the urban core.
However, driverless cars also have a dark side. They could enable more car-centric, energy-intensive lifestyles if long commutes become less frustrating and sprawling development goes unchecked. “Mega commutes may become more tolerable as people nap or eat or work while their car drives them to work,” said Rucks. “People may also be tempted to send their car on every errand that strikes their capricious fancy, filling roads with a profusion of zero-occupant vehicles that would make today’s single-occupant-vehicle congestion look like a drive in the park.”
This much is clear: If autonomous driving is where automobiles are heading, it’s a critical part of the climate equation. Whether the technology ends up being a boon or a bane for climate efforts will depend greatly on planners, technology developers, and citizens having the vision to harness it not only for safety and convenience, but also social and environmental good. “The individuals behind technology always shape it,” said Susan Shaheen, director of innovative mobility research at University of California at Berkeley.
Rethinking the Role of Cars
The most recent report from the United Nations’ Intergovernmental Panel on Climate Change predicts that emissions from transportation may rise faster through 2050 than those from any other major source. Freight transport and car sales are expected to boom as incomes and demand for consumer goods rise in China, India, Brazil, and other nations. So technology that changes the way we drive ourselves and ship our stuff can have a vital impact on climate change.
“Environmentalists should be on notice,” said Ethan Elkind, climate policy associate on faculty at the U.C. Berkeley and U.C. Los Angeles law schools. “There’s an argument for developing more connected communities that don’t rely on automobiles, and self-driving cars don’t change that core argument.” But they will change the calculus of the proper role for cars in sustainable urban and regional planning, he said.
One example he cites is that of shared-use vehicles. According to Elkind, fully autonomous cars could support a “Netflix-type subscription approach,” in which people hire a self-driving taxi to pick them up on demand (like a Lyft or UberX, but without the car’s owner at the wheel) The idea is that this type of service could be more convenient for many trips than traditional cabs or car-sharing services like Zipcar, which require vehicles to be picked up and returned at the same location. The fact that a car isn’t ready and waiting in the garage could be enough to encourage people to “bundle” trips, Elkind added, which is more efficient than making multiple round trips.
Shaheen sees further potential for automation to reduce reliance on personal cars beyond urban cores. “If you are in the suburbs and you aren’t too distant, could this be an on-demand, shared-use shuttle service? It serves as a form of public transit instead of building out more fixed-route transit,” she said. That notion has particular appeal at a time when the Highway Trust Fund is approaching insolvency.
“It’s becoming more challenging for local governments to find the funds to extend local transit services,” Shaheen said. “Could this be a way of creating more public-private partnerships to help fill the gaps that exist in the transportation network, without relying on the government to do it all? It could make the areas between suburban and urban areas not entirely car-dependent.”
Ultimately, the sweet spot for autonomous vehicles may look less like a personal pod in every garage and more like a mesh of integrated services. Because when it comes to maximizing the benefits of vehicle automation for both the environment and quality of life, highways flowing with synchronized chains of robotic personal cars come up short.
Imagine instead smart cities that are crisscrossed by driverless taxis; connected to each other by semi-autonomous shuttles; supplied goods by intelligent fleets of downsized trucks; and importantly, still served by buses and light rail, while privately owned passenger cars fade to a much smaller role in city life. “We should always be looking at non-[car] ways of moving people around,” said Elkind. “That should be priority number one.”
Efficiency on Autopilot
For trips where we continue to use personal cars occupied by a sole individual, Shaheen said the environmental benefits of autonomous and connected cars will depend in large part on just how autonomous we allow vehicles to become. The key, Shaheen suggested, is to prioritize efficiency — basically hard-coding “eco driving” into the system.
In theory, connected cars can organize themselves in such a way that minimizes waste and congestion. But that means giving up some independence at the individual car level, allowing a computer to block or override certain driving behaviors.
On the highway, autonomous vehicles can deliver relatively straightforward efficiency gains by reducing speed variation and unnecessary acceleration, Miller said. But these techniques are an “incremental improvement” relative to the gains that are possible with drafting on the highway. Like cyclists locked into formation in the Tour de France, autonomous vehicles equipped with sensors can wirelessly link together in “platoons” that increase fuel efficiency by reducing aerodynamic drag.
But on city streets, greater efficiencies could be realized by using automation to reshape our behavior. “A lot of ‘eco driving’ [tools] today are responding to your human behavior,” informing drivers about their performance rather than actively discouraging specific behaviors in advance, Shaheen said. The information display in a Toyota Prius, for example, will show your miles per gallon and might motivate you to avoid jackrabbit starts. But this kind of “eco driving” is passive and easily overridden. “When you’re late for work, you throw it out the window,” Shaheen said.
Vehicles that can communicate with each other and the infrastructure around them could create a more binding model. It’s hard to see this approach accepted for all routes at all times. Can you imagine gearheads and anxious commuters giving up the option to floor it when space allows?
But today there are already zones where special rules apply. In London, most motorists must pay a daily fee to drive in the city center, while plug-in cars including the Tesla Model S and Chevy Volt are exempt from the charge. With connected cars, such zones might favor vehicles in their lowest-carbon or most pedestrian-aware mode, or a setting that gives up the most power to the network to choreograph smooth traffic flow.
“This form of eco driving could be very technology based and put decisions regarding how the vehicle flows through traffic in the hands of the system,” Shaheen said. “So the vehicle slows down when traffic is approaching. It’s interfacing with all the traffic signals.”
According to Google’s Urmson, the company has encountered thousands of different situations on surface streets and “built software models of what to expect, from the likely (a car stopping at a red light) to the unlikely (blowing through it).” A scene that “looks chaotic and random on a city street to the human eye is actually fairly predictable to a computer,” he wrote.
But fully realizing the potential for automation to relieve congestion, improve safety, and reduce emissions presents challenges beyond the vehicle itself.
If data is to flow between vehicles and infrastructure, it must be secure and reliable; privacy must be maintained. And equipment must be upgraded to send and receive signals, such as the phase and timing of traffic lights. Interoperability across borders, makes, and models is key.
Morality comes into play, too, as engineers will have to program cars to take action in life-or-death scenarios: If the options are to hit a child that runs into the road or swerve and mow down a crowd of people, what should the car be programmed to do? Does it matter if the car is carrying one passenger or five? “All the cars basically have to work the same way,” Miller said. “And for the engineer, that kind of has to be figured out before writing the code.”
Shaping Cities of the Future
Policy makers also have important — although perhaps not quite so grim — decisions to make. If automated cars indeed reduce emissions, should they join electric cars, carpools, and buses in getting toll-free access to special lanes or discounts on congestion pricing? Should they earn a tax credit? Should they receive dedicated lanes on surface streets, as bicycles, bus rapid transit, and streetcars do?
In California, air regulators have recognized the need to determine how autonomous vehicles figure into the state’s climate goals, which include reducing greenhouse gas emissions to 1990 levels by 2020. In the latest proposed update to California’s plan for reaching this target, air regulators wrote that, while more research is needed, an increasingly connected and automated transportation system “could have dramatic efficiency and emissions benefits” for the state’s transportation sector.
“We see it as a potentially beneficial technology, especially in the area of congestion reduction,” said Dave Clegern, a public information officer for California Air Resources Board climate change programs. “At this point we have no clear idea how these vehicles would be deployed or to whom. If they were to become available to businesses, the general public, or a portion thereof, it might well make sense to incentivize,” Clegern said.
Rucks, of the Rocky Mountain Institute, agreed. “It does seem it would make sense to provide dedicated lanes, similar to today’s HOV lanes, for autonomous freight platoons,” he said. Access needn’t be exclusive. “Imagine a heads-up display flashing an alert: ‘Platoon approaching. Would you like to join it?’ You put in your final destination, accelerate to sync up with the platoon for several miles, then disengage and move toward your exit as your destination approaches.” Access could be used as a carrot for cleaner cars, Rucks said. Spots in a platoon might be allocated based on the number of passengers in a car or the efficiency of the vehicle, with more efficient vehicles enjoying higher priority.
Such preferential treatment is sure to be a lightning rod for public debates. Those are debates worth having sooner than later, however, because robotic cars are coming. Indeed, in prototype form, they’re already among us. But as they begin to master city streets, their biggest benefit to climate and urban life may be the opportunity to re-orient our communities around on-demand access and low-carbon mobility, ultimately dethroning the personal car as king of the road.