California Loosens Rules For Driverless Cars, Clearing The Way For Robot Taxis
By Russ Mitchell, LOS ANGELES TIMES
February 28, 2018
Here come the robot cars.
Sometime this spring, self-driving test cars will begin appearing on California highways, with no people inside, the result of new regulations issued by the state Department of Motor Vehicles this week.
The rules, first proposed last October, will also allow ride-hailing companies such as Uber and Lyft to begin selling rides in driverless cars, removing expensive human drivers from the equation. Technically, that could happen this year, although no ride-hailing companies have yet announced such plans. (It’s unclear whether robot taxis would require approval by the California Utilities Commission in addition to the DMV.)
“We do intend to do that eventually,” said a spokesman for Waymo, the driverless-car arm of Google parent Alphabet. Lyft said it was “encouraged” by the regulations. Uber called the new rules a “significant step.”
For now, the California regulations bar driverless trucks, motorcycles and cars with trailers. Uber said it “look[s] forward to working with California as it develops regulations applicable to autonomous trucks.” (Uber owns Otto, an autonomous truck company.)
Until now, driverless cars were allowed on California roads only with a human behind the wheel. The new rules — which were sought by automakers, Silicon Valley technology companies and many safety advocates — loosen restrictions on testing and, crucially, set standards to allow the sale or lease of robot cars and their operation by ride-hailing fleets.
“Those deployments will start small but grow fast,” said Bryant Walker Smith, a specialist in autonomous vehicle regulation at the University of South Carolina School of Law. Already, Waymo plans to deploy a robot-car ride-hailing service in the Phoenix area later this year.
“In general, these regulations move California in the right direction and align California more closely with what the rest of the nation is doing,” said Nidhi Kalra, co-director of the Center for Decision Making Under Uncertainty at the research group RAND Corp.
Not everyone is happy with the new regulations. John Simpson of Consumer Watchdog in Santa Monica has long argued that robot cars should not be allowed until they’re proved safe, and that the new regulations in some ways make things worse.
“A remote test operator will be allowed to monitor and attempt to control the robot car from afar,” he said. “It will be just like playing a video game, except lives will be at stake.”
He’s referring to a new DMV rule that requires remote monitoring of robot cars, allowing a human to use communications networks to take over the car in case of trouble. Companies such as Nissan are already testing remote systems that can, for example, reroute cars around unexpected obstacles.
Other systems, like one provided by Silicon Valley startup Phantom Auto, use a human in a remote operations center with a steering wheel and brake that could route a driverless car around uncharted or confusing obstacles, such as an accident scene or an unexpected construction site.
“Everyone understands, although this tech has gotten to an incredible point, it’s not 100 percent yet,” said Elliot Katz, a Phantom co-founder.
The need for remote operators is debatable – Arizona and Florida don’t require them, for example. But safety is bound to remain a central issue as robot cars are deployed in increasing numbers.
RAND’s Kalra, who specializes in safety statistics, said robot cars have proved less likely to get into minor crashes than human drivers. But “when it comes to injuries and fatalities, we won’t be able to know until we’ve had hundreds of millions or billions of miles” of driving history, she said.
Banning driverless cars until the answer is clear would mean robot cars would never be deployed, she said. The United States suffers nearly 40,000 traffic fatalities a year, with 94% caused by distraction and other human error, according to the U.S. Department of Transportation.
The key, Kalra said, is early deployment “in locations that offer fair weather, well-marked streets, in neighborhoods where it makes sense.”
Waymo, for example, began its Arizona deployment in the Phoenix suburb of Chandler, where the streets are on a neat grid, there is little foliage to block the view, and few pedestrians brave the heat.
California took the lead in driverless car regulation in 2012, when the Legislature ordered the DMV to come up with rules.
Those rules, issued in 2014, set up a permit process that some companies complained was too cumbersome. States such as Michigan and Florida implemented rules that were far less restrictive.
Arizona pointedly created no new rules at all. In 2016, Uber took its Volvo driverless test cars off the streets of San Francisco and sent them to Arizona.
Waymo enrolled Phoenix-area residents in an “Early Rider” program last year for free rides in driverless Chrysler Pacifica minivans. On Jan. 24, Arizona officials green-lighted Waymo’s application to become a ride-hailing company. Later this year, Waymo plans to compete with Lyft and Uber in Phoenix, but will not hire human drivers.
Silicon Valley has been joined by traditional carmakers in the transition to autonomous driving. On Tuesday, Ford said it has begun introducing a fleet of driverless test cars in Miami that will grow into the “thousands.” The fleet will test out the automaker’s “Transportation Mobility Cloud” that could provide a computerized ride-hailing platform for its development partner Lyft and others. Already, driverless Ford cars are delivering Domino’s pizzas in Ann Arbor, Mich.
What happens to human drivers is one big issue not under the DMV’s purview. During breakfast at the Four Seasons hotel in San Francisco on Tuesday, Uber Chief Executive Dara Khosrowshahi was overheard talking to seven other people at a large table about how to keep Uber’s human drivers happy during the transition to robot control.