Imagine if nine out of every ten cars suddenly disappeared from city streets. On-street parking would be eliminated entirely, giving way to restaurant seating, bike paths, and green space. Parking garages would all but disappear, making room for additional housing, retail, and office space.
Now imagine if cities around the world sprawled for hundreds of miles, their centers clogged with twice today’s car traffic, and daily commutes lasted for hours. Both scenarios, it turns out, are entirely plausible futures with the rise of autonomous vehicles, the self-driving cars now navigating city streets from Silicon Valley to Singapore.
“There is a huge potential for change,” says Philippe Crist, an economist with the Organization for Economic Co-operation and Development (OECD). “It creates new possibilities that we are only starting to get our minds around.” The OECD is one of a growing number of think tanks and institutions peering into the crystal ball of urban mobility’s future and, more often then not, conjuring rosy expectations. Fleets of shared, self-driving vehicles could indeed remove nine out of every ten vehicles on city streets, eliminating the need for all on-street parking and 80% of off-street parking, according to a recent study by the group.
The Paris-based think tank’s assessment comes from simulations of Lisbon, Portugal, a mid-sized European city, based on real trip-taking activity and assuming that people rode in self-driving cars that can be shared simultaneously by several passengers. The findings were comparable to similar recent studies of shared autonomous vehicles in New York, New Jersey, Ann Arbor, Michigan, and Singapore.
The significant reduction in vehicles—a 90% drop in cars over a 24-hour period and 70% fewer vehicles during peak travel periods—comes through sharing of cars that today typically sit unused 23 hours per day. “The arrival of this technology allows consumers, or companies, to extract value from these stranded assets the rest of the day,” Crist says.
While autonomous vehicles may sound futuristic, self-driving cars are increasingly part of the present. Last January, Mercedes-Benz introduced the F 015 Luxury in Motion self-driving concept car with front seats that can swivel away from the windshield to face the rear seats. In March, industry supplier Delphi Automotive drove from San Francisco to New York in an Audi Q5 piloted almost entirely by radar, cameras, and laser sensors. And Google, which has already logged over 1.3 million driving miles with autonomous vehicles, is now building a fleet of 100 next-generation prototypes. In June, the vehicles, each of which looks like a large computer mouse, began navigating city streets from Silicon Valley to San Francisco.
“The pace at which autonomous cars are coming on is dramatically faster than what people had imagined,” says Robin Chase, former CEO of car sharing company Zipcar.
Utopia or Something Else?
While autonomous vehicles are coming, it’s still entirely unclear what their arrival will bring. If everyone rides in an autonomous vehicle, there will be few if any accidents, which will allow for lighter, more fuel-efficient vehicles. And if people share these vehicles as they move about the city, the total miles driven each day will also decrease. Combine lighter cars with fewer miles traveled, and the amount of energy used to get people where they need to go drops by 80%. This, at least, is the “utopian scenario” in a recent study by researchers at Lawrence Berkeley National Laboratory.
Yet in the study’s “dystopian scenario,” energy consumption more than doubles as larger, privately owned vehicles travel longer distances. “You could imagine a consumer wanting to be able to cook dinner on their way home or to watch movies while the car is driving, and they could then not care if they lived close to their work place,” says lead author William Morrow.
“Why do we still insist on buying a car?”
Luís Bettencourt, a professor of complex systems at the Santa Fe Institute in New Mexico, who was not involved with the study, thinks a dystopian outcome is more likely. “I think the expectation is they would tend to make cities bigger and less dense,” he says. Bettencourt, whose research focuses on cities and urbanization, cites Marchetti’s constant. Named for the Venetian physicist who devised it, Marchetti’s constant states that throughout history, no matter where people lived or what form of transportation they used to get about, they have always spent and will continue to spend an average of 30 minutes each way getting to and from work. Autonomous vehicles could speed up commutes, allowing people to live farther apart, Bettencourt says.
“Once you have autonomous vehicles and algorithmic ways of coordinating traffic, then the obvious next step is you start coordinating vehicles into convoys. You can imagine 200 mph corridors where these trains made out of cars are just moving across the city, and those would be the future highways.”
Bettencourt questions why the reduction in vehicles, predicted by many, hasn’t already occurred. “We’ve had taxi services for a while, and we have Uber and new incarnations of those same services,” he says. “Why aren’t these things taking over cities entirely? Why do we still insist on buying a car?”
Fewer people, however, are insisting on buying cars. Recent studies suggest car ownership in the U.S. peaked during the last decade and is now on the decline. One study conducted by researchers at the University of Michigan’s Transportation Research Institute in Ann Arbor found more than 30% of households do not own a car in six out of 30 of the largest U.S. cities. In New York City, 57% of households do not have a vehicle, up from 54% in 2007.
A separate study by the same institute found a significant decrease in recent years in the number of young drivers across 15 countries. The decrease correlated strongly with internet access, suggesting that virtual contact via electronic devices may be reducing the need for actual contact.
Rise of the Taxibot
Electronic devices may not only be reducing interest in car ownership but are enabling the sharing of vehicles that make shared, self-driving vehicles possible.
“You couldn’t have imagined this ten years ago when people didn’t have smart phones and mobile computing was not available,” says Emilio Frazzoli, head of Future Urban Mobility for the Singapore-MIT Alliance for Research and Technology. “Now you have this ability to connect and book a car. You see it with Uber and the proliferation of taxi booking apps or public transportation schedule routing apps, and this is at the same time you have autonomous vehicle technology that is evolving. You can marry the two.”
Sharing of non-autonomous vehicles already offers a distinct advantage over traditional car ownership, says Chase, author of the book Peers Inc and a champion of the sharing economy. “If you are financially smart and you are living in the city and you don’t need a car to get to work, you are insane to own one,” she says. “You will always be saving money by renting them when you need them.”
Further savings come when people not only share a car but also share the ride. Ride-sharing startups Uber and Lyft now offer UberPool and Lyft Line, carpooling services that allow people headed in the same direction to share a ride. “In San Francisco today, UberPool is $7 to go from anywhere to anywhere. That’s probably one-third the cost of a taxi,” says Xavier Mosquet, a senior partner at the Boston Consulting Group.
If a private company could use an autonomous vehicle rather than paying a driver, the cost of transporting people across the city would be even less. A recent study by the Boston Consulting Group found the cost of conveying one passenger by an autonomous vehicle would be 35% less than by conventional taxi at the average taxi occupancy rate of 1.2 passengers. Increase an autonomous vehicle’s rate of occupancy to just two passengers and the cost per passenger becomes competitive with mass transit.
Mosquet says such shared, autonomous vehicles may not work everywhere but will be well suited for the world’s 50 to 100 largest cities where population density is greatest. “In Manhattan, you have ten people looking for the same cab and mostly likely headed in the same direction on every corner,” he says. “With a smartphone now collecting that information, using it is extremely easy.”
Where demand is high, larger, driverless buses could give autonomous vehicles an added advantage. Today, the Exclusive Bus Lane (XBL), a 2.5 mile stretch of highway connecting the New Jersey Turnpike with midtown Manhattan, is the busiest bus lane in America, with the capacity to transport up to 41,000 people per hour through the Lincoln Tunnel. Self-driving buses could reduce the time between vehicles from five seconds to one second, allowing the tunnel to accommodate more than 200,000 passengers per hour, according to a recent study by researchers at Princeton University.
Singapore and a number of European countries are now testing or preparing to test fully autonomous buses that could complement fleets of self-driving cars. “You can combine different types of transportation to provide both the convenience that people need with the scalability that a city needs,” Frazzoli says.
Before fleets of self-driving cars or buses can replace the cars of today, a number of challenges still need to be addressed. Among a range of sensor technologies for autonomous vehicles that includes radar, ultrasound, and cameras, laser-based “LiDAR” sensors shows the most promise. The technology, however, has been incredibly expensive. The LiDAR sensors that Google used on its first generation prototype cost $80,000 per vehicle and $8,000 per vehicle for its current prototype, Mosquet says. Costs continue to drop quickly. Just last week, automotive startup Quanergy announced a solid-state LiDAR system that will cost under $250 when it enters mass production later this year.
Apart from technical issues, a significant challenge may simply be gaining widespread adoption. The best case scenario of the recent OECD study—where shared autonomous cars take nine out of ten vehicles off city streets—is based on the assumption that everyone rides in an autonomous vehicle and shares each trip with others headed in the same direction. If half of a city’s population continues to drive their own cars, the number of vehicles—and the amount of time people spend idling in traffic jams—could actually increase.
Lower costs alone may not be enough to get people to give up car ownership. If any city is likely to embrace shared autonomous cars it will likely be Singapore, an authoritarian city-state that has pioneered a number of regulations designed to limit car ownership and ease traffic congestion. For the past quarter century, Singapore has imposed heavy taxes and licensing fees for car ownership that would be difficult to pass elsewhere. A BMW 435i that sells for $55,000 in the US costs $220,000 in Singapore after all taxes and fees. A $16,000 Ford Fiesta would set you back $108,000 in Singapore. In 1998, the city also introduced the world’s first electronic road pricing, or “congestion pricing,” which charges vehicle operators a fee to enter the city’s central business district.
In spite of these and other efforts, including free subway rides for early morning commuters, ridership on the city’s subway system has declined slightly, from 67% of all commuters in 1997 to 66% in 2014.
“They have tried to implement almost every measure that people have thought of in terms of encouraging public transportation use versus cars,” Bettencourt says. “But the fact is—what has been happening in parallel—is they have had such strong economic growth that a lot of people can afford a car even under those conditions. It’s actually a fairly congested city for all the wonderful public transportation it has.”
The cost of shared transport could drop so low that car ownership wouldn’t be worth it.
Frazzoli, who plans to submit a bid in Singapore’s recent call for an autonomous vehicle pilot program, says he thinks the technology has a bright future in the city-state due to strong government support. “They see this as a critical technology for their city, and therefore they are supporting its development,” he says.
Elsewhere, private companies like Uber, which recently began test-driving its own autonomous vehicle, could drive down the cost of shared transport to such a point that car ownership wouldn’t be worth it.
Bettencourt, however, remains skeptical whether enough people, in Singapore or elsewhere, will buy into shared self-driving cars to make them a success. “The past may already be giving you all kinds of hints to the future of what people may and may not prefer,” he says.