NEW YORK — My first ride in a driverless vehicle was a surprisingly human experience.
Standing on a sidewalk in the Brooklyn Navy Yard, where battleships such as the USS Arizona were once built, I spotted a small black vehicle stopped at a signpost. It looked a bit like a stretch golf cart, but with the label “Self-Driving” on the back and the logo of Optimus Ride, a Boston startup, plastered on the side. There were two people in the front seat.
“What time do you leave?” I asked. The guy in the driver’s seat told me the vehicle was carrying people to an East River ferry dock every five to seven minutes. He told me that I could either get in right away, or they’d let me know when they were about to depart.
There was no app to use to interact with the vehicle — just two people to talk to through the open window. After a few minutes of standing around, I got a wave from the fellow in the side seat, who was balancing a laptop on his knees and holding a tablet. I hopped in. The computer operator gave his colleague in the driver’s seat the go-ahead. He pressed a button on the dashboard, and we started on the half-mile journey.
Optimus Ride is one of a few Boston companies that are developing technologies for driverless, or autonomous, vehicles. (The company doesn’t actually build them; they are made by Polaris, the Canadian maker of snowmobiles and all-terrain vehicles. Optimus outfits each one with cameras and sensors.) Optimus has its roots at the Massachusetts Institute of Technology, and cofounder and chief scientist Sertac Karaman is a professor there. The company has raised more than $40 million in venture capital funding.
Unlike competitors such as Uber and Waymo (a spin-out from Google), which are focused on the open road, Optimus says it is primarily developing vehicles that will operate in “geo-fenced” areas, like university campuses or housing developments — not on the turnpike during rush hour.
The Brooklyn Navy Yard is one such contained environment, a 300-acre complex that is home to coffee roasters, sculptors, prototyping shops, and import businesses. The benefit of such environments is that they are far less challenging for driverless vehicles to navigate, and the routes are essentially pre-planned, like those of shuttle buses.
On my ride, the first thing I noticed was that the driver was triggering the turn signals. I got two answers when I asked about that: that turn signals hadn’t been woven into the company’s system, and that it helped the driver stay alert if he was responsible for a task.
The vehicle was set with a top speed of 15 miles per hour through most of the Navy Yard, but on a final stretch of pier heading to the ferry dock, that was trimmed to 10 miles per hour.
Sometimes, I was told, lots of pedestrians and cyclists used the pier road — though none were in evidence on my trip at about 4 p.m. (I rode just as any other passenger would have, without pre-arranging my visit as a journalist.)
When the vehicle arrived at the dock, the driver took full control of the vehicle and executed a three-point turn to orient it properly for the return trip. That, he explained, was because the vehicle had a big buffer zone — it would stop if it got too close to people — and there were often many people on this part of the pier. A human driver could see that no one — save for an Optimus Ride employee who was overseeing the fleet — was within 30 yards, but this generation of driverless vehicle doesn’t have the same situational awareness.
On my return trip, I noted that the vehicle stopped at an intersection where there was no red sign — just the word STOP painted in white on the pavement. While that might have confused some self-driving vehicles, this one was programmed for a specific route, so it paused properly at the right spot. Its digital cameras also identified a pedestrian about to cross the road, and the vehicle waited before proceeding. It stopped to let one car go through an intersection, but seemed confident that it had the right of way when another car was waiting to make a left turn.
More than 1,000 people have ridden the vehicles in Brooklyn since the service opened to the public on Aug. 7. It’s paid for by the Brooklyn Navy Yard Development Corp., so it’s free. At 15 miles per hour, with two people in the front plus sensors and software monitoring our progress, it felt exceptionally safe. I considered offering a tip, but wondered how I might split it between the humans and the algorithms.
How long will it be before Optimus is offering truly driverless trips around such sites? Karaman hedges a bit when he says that 2020 is the goal — at least for “certain sites, certain routes, at certain times of day.”
Because driverless vehicles are still in their infancy, it’s easy to be skeptical about their potential to be safer, more efficient, and less expensive to operate than cars with drivers. Yet many new technologies require a decade or two of development before they’re truly better than the things they’re designed to replace.
So far, says robotics pioneer Rodney Brooks, predictions about the imminent arrival of driverless cars have been “way off.“ Brooks is a cofounder of the robot maker iRobot and an MIT professor emeritus. “There’s a general impression of things being tested autonomously” — without people at the wheel — “and I have not seen one yet,” he says.
I wondered: What would be Brooks’s criteria for hopping into a truly driverless vehicle and driving on a regular road or highway — versus one in a contained, low-speed environment like the Brooklyn Navy Yard? Would he want to know about the maker’s safety record?
“It’s a hypothetical so far in the future that you can’t say anything sensible, because you don’t know what will have gone down,” he says. “I just don’t think anyone is going to deploy them for a long time.”
Karaman, the chief scientist at Optimus, doesn’t disagree. That’s why, he says, his company is starting with geo-fenced environments; a second phase could be vehicles without drivers that are monitored remotely by a “team of 10 people that could operate a fleet of 50 vehicles,” he says.
But hopping into a car in Harvard Square, in a blinding snowstorm, and being taken to Times Square with no driver on board? “That’s not going to happen anytime soon,” Karaman says.
Scott Kirsner can be reached at kirsner@pobox.com. Follow him on Twitter @ScottKirsner.