- Waymo, Google’s self-driving car spinoff, showed off its technology to reporters on Monday at its California testing facility. Waymo officials declined to say when they expect to offer the technology commercially or where it will first show up. The technology seems mature, but questions and challenges remain.
MERCED COUNTY, California – California’s Central Valley is known for its miles and miles of farm orchards, fights over water rights, and, these days, high unemployment.
High-tech? Not so much.
But I was here on Monday to see in action a cutting-edge technology that has the potential to greatly reshape our economy and society – the self-driving car. Waymo, Google’s autonomous vehicle spinoff, has its testing facility at a decommissioned Air Force base. The company invited several dozen reporters to see its cars in action, get a ride in one, and hear its case that its technology is all-but ready for the real world.
“We’re pretty excited about where we are right now,” in terms of developing fully autonomous cars, John Krafcik, Waymo’s CEO, told the assembled journalists. He continued: “We’re getting to the point where we’re really close.”
Just how close, though, is anyone’s guess – anyone outside Waymo, that is. When asked when we might see Waymo roll out its technology commercially in a product the public at large can see or interact with, Krafcik declined to offer a forecast.
It’s also unclear exactly where Waymo’s technology will first start showing up. Krafcik said the company is pursuing several opportunities at the same time: in trucking, ride-hailing, and ride-sharing, in making “last-mile” connections between consumers homes and the closest public transit stops, and in personal vehicles. But he declined to say which opportunity the company thinks is closest.
Krafcik did say that the company plans to work with partners such as Fiat Chrysler, Lyft, and the rental-car company Avis to deploy the technology, rather than going it alone. Waymo has spent the last 12 to 18 months building out such partnerships, he said.
“We see our role not as disruptors but as enablers,” Krafcik said.
Waymo has been focusing on so-called Level 4 autonomy. At that level of sophistication, a car can drive itself without any driver input, but can only do so in a relatively confined area or in a relatively circumscribed set of conditions. The company is already offering fully autonomous rides in certain areas of Chandler, Arizona, outside Phoenix, as part of a test of its systems.
The company is hoping to gradually progress up to Level 5 autonomy, where self-driving cars can basically go just about anywhere a human-driven car could go.
Riding in a car with no driver
We reporters got to see first-hand how just far along Waymo’s Level 4 technology is. Part of the outing involved an approximately 10-minute, two-mile drive around the company’s testing facility with nothing but Waymo’s technology behind the wheel.
Waymo’s testing site is at Castle Air Force Base, a decommissioned military facility where pilots trained for World War II and in the Cold War, was the home of a wing of the nation’s Strategic Air Command. On its 91-acre section of the old base, the company has set up a network of streets, traffic circles, driveways, crosswalks, and even a railroad crossing to create a wide range of scenarios for its cars to contend with.
Accompanied by a Waymo engineer who was sitting with us in the back of one of the company’s specially equipped Chrysler Pacifica minivans, another reporter and I could do little but watch and hope for the best as our vehicle made its way around the facility. To show what the minivan could do and how it might handle real-world situations, Waymo had some of its employees drive around in other cars, cross streets or bicycle beside the vehicle.
At least in this fairly staged environment, the vehicle handled everything with aplomb. I was never concerned about my safety or those around us. And while it was a bit strange to not have anyone in the driver’s seat at first, I quickly grew accustomed to it.
Waymo has put a lot of thought into how its cars will reassure and communicate with riders, using visual and verbal cues. On the back of the front-row seats there are a pair of screens that show an image of the vehicle you’re riding in along with some of the things in the environment that it senses, such as other cars, pedestrians, traffic cones, stoplights.
The display is designed to show riders that the car is aware of what’s happening around it, and view changes noticeably depending on what the car is doing. If the car is attempting to turn right, for example, the angle of view will shift to show more of the road to the left of the car and any oncoming traffic from that direction.
But the screens also show messages to provide more information. They might show the speed limit in the area or flash a quick message to explain that the car is “yielding to pedestrians.” And when you’ve arrived at your destination, the screen will tell you, “We’re here.”
The information on the screens is designed to build trust, so “people will feel Waymo is a safe and reliable and trusted chauffeur,” said Ryan Powell, Waymo’s head of user experience design.
The challenges ahead
However, as well Waymo’s car performed, it’s clear that the company still has plenty of challenges ahead of it, both technical and otherwise.
Much of Waymo’s autonomous vehicle testing has been done in relatively mild climates, including the San Francisco Bay Area. The cars don’t have a lot of experience driving in winter weather conditions. The company is trying to address that. It announced last week that it will begin testing its cars in Michigan specifically to see how they handle ice, sleet, and snow.
Another key question for self-driving cars is how they will handle unusual situations that may occur in the real world. Waymo’s been trying to address that multiple ways, not only by having its cars drive on real streets, but also by running through potential scenarios at its facility here, and then using data from those scenarios to run computer simulations back at its Mountain View headquarters.
There are also numerous non-technical questions with which the company has to contend, not only about how it will market its technology but also about how it will convince government regulators that it is ready for the road. And there are the societal questions about how the rollout of autonomous cars will impact jobs. What will happen to taxi and Uber drivers, truck drivers, and the like?
And then there’s the public. It’s an open question about whether the public is ready to adopt robotic cars. Even if they’re open to the idea – and Waymo’s data suggests about half of them are – they’re likely to have plenty of questions about how they will use such cars, how they will work, and how they will handle particular driving scenarios.
In that area alone, Waymo seems to have a lot of work to do. When asked by reporters how its cars would handle particular situations like obstacles in the road that might block traffic, representatives gave conflicting answers. One even suggested that the cars might not move until an obstruction – a moving van that was double parked, say – were cleared out of the way.
I’m guessing Waymo’s engineers have a better answer than that. But they may find navigating the non-technical obstacles to our self-driving car future are as difficult as the technical ones.