One of the biggest potential pitfalls for developers of autonomous vehicles is psychological: Will most people trust the cars enough to ride in them?
It might not be easy to win people over, but it’s possible—if a Boston startup’s recent tests are any indication. On Tuesday, NuTonomy co-founder and president Karl Iagnemma shared early reactions from people who have ridden in cars controlled by his company’s software.
“The feedback has been really interesting, and I would say overwhelmingly positive,” Iagnemma said at a press briefing, during which NuTonomy and its parent company, Aptiv (NYSE: [[ticker:APTV]]), announced plans for a new Boston office focused on autonomous vehicles and other mobility technologies. (Aptiv, formerly known as Delphi Automotive, acquired NuTonomy for $400 million-plus in October.)
“We’ve found,” Iagnemma continued, “that once people get into one of these cars, typically there’s a little bit of maybe nervousness or apprehension because it is a little surprising to see that wheel turn by itself for the first time. But if you’re able to deliver a ride that’s safe, that’s comfortable, and that [provides] the right [amount] of feedback to the rider, we’ve found that people will quickly gain acceptance of the technology.” (More on what he means by feedback in a moment.)
Iagnemma’s comments should be taken with a grain of salt. He didn’t say how many people have ridden in cars equipped with NuTonomy technology, but it’s not that many. And the rides have so far only taken place in one neighborhood in Boston.
This fall, NuTonomy began letting employees’ friends and family summon a driverless car with an app developed by the company, and then take a short ride from the firm’s offices in the Seaport neighborhood to nearby South Station, and back. And this month, NuTonomy and ride-hailing app company Lyft launched a limited pilot program through which some users of Lyft’s service can ride in a NuTonomy driverless vehicle, if the pickup and drop-off locations are in the Seaport. (During all of these tests, an employee is at the wheel and can take over driving if anything goes wrong.)
Despite the small sample size, the rider feedback NuTonomy shared provides an interesting snapshot of what people think of this emerging technology. In a video played at the press briefing, a couple of riders said they were surprised by how smooth the ride was. “I feel like it’s a better driver than most people I know,” one person said. (To be fair, that bar is not set very high in Boston. Yeah, I know, stereotype.)
Here are three of Iagnemma’s initial takeaways from the pilot program:
1. Cautious is better—for now. Most people have told NuTonomy that they felt like the autonomous car is safe and cautious, Iagnemma said. “There are a few people who say, ‘We think the car is a little bit too cautious,’” he said. “We are OK being in that spot.”
At this stage, NuTonomy wants to err on the side of careful driving—making riders feel safe and comfortable is the priority. Nevertheless, NuTonomy has programmed its software so that vehicles don’t sit at intersections too long before taking off, which could cause traffic slowdowns and aggravate human drivers, Iagnemma said.
“If you’re too conservative, people are either frustrated or will take advantage of the car,” he said. “There is a balance there.”
Driverless vehicles could someday become capable of customizing their driving styles. “We’ll eventually get to a point where the technology is adapting to your personal preference, perhaps to your own driving style,” Iagnemma said.
2. Riders want transparency, but perhaps not too much. NuTonomy sets up a screen near the dashboard of its driverless vehicles that displays real-time information about the ride and how the car’s software is interpreting its surroundings.
“The interface to the car is really important,” Iagnemma said. “The information that’s being shared between the vehicle and the rider turns out to be the way that the rider gains confidence in the technology.”
But if the display were to share all the information about the complex calculations made by the vehicle’s software each moment, it “might make them a bit nervous,” Iagnemma said. “We’ve done quite a bit of experimentation in trying to figure out what type and quantity of information we should be passing back to the rider,” he said.
3. Accessibility is critical. It sounds obvious, but operators of driverless taxi services must make it easy for riders to find and get into the vehicles. For example, if a senior citizen has to cross the street to get into the car, that could be “a bit of a hardship,” Iagnemma said.
“We all experience trying to call a Lyft and sometimes having the challenge of finding the driver,” Iagnemma said. “That challenge gets even more difficult when there’s no one behind the wheel.”
