Follow by Email

Sunday, August 3, 2014

Why self-driving cars will mean more traffic


Bloomberg Businessweek

A future based on driverless cars could mean big changes to the way cities are shaped. Given that there are plenty of things wrong with our relationship to cars today, it’s tempting to fantasize about how much better things would be be once self-driving vehicles become the norm.

But things could get worse, too.

“U.S. history shows that anytime you make driving easier, there seems to be this inexhaustible desire to live further from things,” said Ken Laberteaux, the senior principal scientist for Toyota’s North American team, in an interview with Bloomberg at the Automated Vehicles Symposium in San Francisco last week. “The pattern we’ve seen for a century is people turn more speed into more travel, rather than maybe saying, ‘I’m going to use my reduced travel time by spending more time with my family.’”

Laberteaux thinks that self-driving cars could re-create a dynamic that turned American suburban in the years after World War II. Back then, big, smooth highways made longer commutes bearable, opening the countryside up to McMansions and creating a host of sprawl-related traffic and environment problems that are still around . Cities have been enjoying a rebirth in recent years, but the exurbs might look pretty attractive if sitting in a car resembled hanging out on a moving couch.

That’s the naysayer’s view, anyway. Self-driving cars would also create all kinds of positive possibilities in urban design. Tech-enabled transportation companies such as Uber and Lyft believe they can undermine the idea that every household should own a car or two. While the drivers for these companies are all human now, automated cars would be even easier to share because they could spend all their time on the roads. This could be the future—after all, Google (GOOG) is a major investor on Uber.

Cars that aren’t beholden to a single driver would mean cities wouldn’t have to use up so much valuable land on parking lots. A fleet of robotic drivers could use narrower lanes, shrinking the proportion of city’s surface area taken up by roads.

Sven Beiker, the head of the Center for Automotive Research at Stanford University, says that the future is likely to have aspects of both dynamics. “People say, ‘When will we have automated vehicles?’” he notes. “The question is really where will we have them.” Wherever they do proliferate, he says, there will almost certainly be more traffic, because personal vehicles are likely to become even more attractive when compared with public transportation such as buses and trains.

Tyler Folsom, a professor of computer science at the University of Washington who studies self-driving cars, think automated vehicles will be far more controllable than human-driven cars because the government can use data collected from automated vehicles to set rules and enforce them absolutely. “The state can simply say that no vehicle is allowed to drive more than X automated miles per day, and they can enforce it,” he predicts. “No increased sprawl.”

That kind of hyper-logical argument is much more likely to pass muster at a university’s computer science department than in the real world, where people are not going to love the idea of the government tracking their every move and sending them fines.

Beiker thinks that a more likely solution is the establishment of specific lanes for automated vehicles that come at a cost. Maybe the price could rise for cars depending on how many miles they’ve already driven that day. This resembles congestion pricing in places such as London, where drivers have to pay to enter center cities. “It’s basically the same mechanism,” he says. “You want to avoid people using their vehicles without thinking about it.” Even if they’re not actually driving them.