This summer, Google will begin testing its custom-made self-driving car on the streets of Mountain View, California. That’s a big step forward for autonomous technology, but there’s arguably a bigger one brewing in Singapore, at least as far as the future of cities is concerned. Officials there are expected to authorize an on-demand driverless taxi trial on public roads—a concept that could change the very nature of urban mobility, with shared autonomous vehicles operating as a sort of point-to-point transit system.
“For me, really the big benefit of this technology is essentially making car-sharing as convenient as private car-ownership, but also as sustainable and scalable as public transportation,” says Emilio Frazzoli, lead investigator for the urban mobility component of the Singapore-MIT Alliance for Research and Technology (SMART), a research consortium that’s applied to run the taxi pilot.
The SMART team is still sorting out lots of details; right now, for instance, the team has only one Mitsubishi i-MiEV, an electric city car, in its “fleet.” But Frazzoli says the basic idea will be that people can book a driverless ride via a smartphone. Initially service will be limited to the “one-north” district of Singapore, a massive business park full of biomedical, digital media, and technology companies. The first round of rides will be free, he says, and might be restricted to “one-north” employees.
Frazzoli says SMART was the first applicant for the trial, and hopes to get its testing permit soon. When approval does arrive, the consortium plans to role out a modest robotaxi trial in “one-north” within about six months. But Frazzoli believes Singapore has much bigger plans for driverless cabs; he thinks officials will broaden the program to the whole city-state within about a year.
“The objective is to use that [trial] to figure things out, then expand,” he says.
Will people embrace the service?
Strictly speaking, the “one-north” program wouldn’t be SMART’s first foray into a shared driverless service. Last fall it held a brief public trial for visitors to Singapore’s Japanese and Chinese gardens. The stakes weren’t quite the same—the vehicles were golf carts, and they didn’t share roads with private automobiles—but hundreds of users still booked the car via an app, entered their destination into a tablet that replaced the steering wheel, and got a driverless ride.
“Yes, it was in the park, but it was in the wild—not some reserved area,” says Frazzoli. “We didn’t segregate the vehicles from the general public. So we were driving along the path where children were running and playing, and dogs were doing their thing.”
One of the problems the SMART team encountered during the garden pilot was how the service responds to someone who doesn’t show up for a ride. If a user booked the car but wasn’t around precisely when the vehicle became available, the service moved down the list to the next user, since the queue was long. That made for a lot of unhappy folks, says Frazzoli. The “one-north” pilot would aim to improve the system’s ability to handle these types of service issues.
More broadly, SMART investigators want to make sure the technology is safe. That means not causing accidents, of course, but also avoiding those caused by other vehicles. They want to see if the taxis are reliable in all situations, from bad weather to heavy traffic conditions. And they want to get a real-world measure of public acceptance. It’s one thing for people to say they’d use a driverless car service, says Frazzoli; it’s another for them to actually do it.
“This is one of the big questions: If you put this technology in front of people, will they embrace it or stay away?” he says.
Trading parking for parks
Singapore seems as well-positioned to embrace driverless taxis into its transport network as any city out there. It’s got a great transit system in the MRT railway. Officials aren’t afraid of innovation: Singapore has had congestion road pricing for decades, and an early-bird fare program was recently tested out to relieve subway crowding. Plus, as Neil Swidey explains in this Boston Globe piece on Frazzoli’s work, private car travel is very, very expensive:
Residents must pay $60,000 or more just for permission to own a car, and that certificate lasts for only 10 years. Then there are crazy import taxes. A basic Toyota Prius can cost a Singaporean more than $150,000.
On paper, at least, the benefits of a shared autonomous cab system would be enormous for Singapore. Frazzoli and some collaborators recently estimated that a fleet of 300,000 robotaxis would be required to serve the city’s travel needs, assuming no one waited longer than 15 minutes during rush-hour. That’s less than half of the 780,000 passenger vehicles on the road in 2011, not to mention all the parking spots that could be converted into other uses.
“Instead of devoting that to storing metal and rubber, you can give it back to people and make parks or whatever else,” says Frazzoli. “If you can get rid of all the cars, you can actually get your city back.”
The government seems to recognize the broad social benefits that a shared driverless vehicle network might deliver. Last year the Ministry of Transport formed a committee to study self-driving cars not just for safety reasons but “to make Singapore an even more sustainable and liveable city.” (Frazzoli is a member.) Just this month officials announced a desire to incorporate driverless technology into the city’s mass transit network, perhaps as autonomous buses.
Whereas U.S. regulators have largely stayed off to the side and followed the lead of technology companies like Google, Singapore’s hands-on approach recognizes that an advance this significant will require strong planning well before it arrives in perfect form. “It’s clear from the government why this is useful and why they should support it,” says Frazzoli. “This is perceived as more of a societal need—something that is strategic as they plan for the future.”