Skip to Content

My Self-Driving Uber Needed Human Help

A test ride in a semiautonomous taxi in Pittsburgh shows the technology is not quite ready.
September 14, 2016

A handful of people who order Ubers in Pittsburgh this morning will discover that their driver is less chatty than usual. And like those lucky riders, I got to experience being chauffeured around town in one of the company’s experimental self-driving cars. I also got to sit behind the wheel and try driving—or, rather, supervising—one of Uber’s new vehicles.

“We’re going to slowly start inviting our most loyal Pittsburgh riders to experience the future,” says Raffi Krikorian, director of the Uber Advanced Technology Center in Pittsburgh. “If they call for a ride, then a self-driving Uber might turn up.”

Uber is exploring the technology in the face of competition from automakers and tech companies investing in automated driving, like Google and Apple. The company’s hope is that the technology will eventually allow it to do away with drivers altogether—something that would help it save a lot of money, even if it were only rolled out in some areas. But for now Uber’s vehicles come with a driver who is trained to take over in an emergency, something I was thankful for on my test ride earlier this week.

During my ride, along a few miles of road near Uber’s testing facility in an old industrial neighborhood, the car performed admirably in many difficult situations—reacting to pedestrians darting into the road, for example—and I mostly felt pretty safe. However, several times the person behind the wheel needed to take control: once so the car didn’t become stuck behind a truck, and once to avoid another vehicle making a sudden turn.

Uber’s trial is unusual in that it involves members of the public, so it will be a good chance to gauge people’s reaction to the technology. Inside the company’s cars, a screen in front of the back seats shows passengers a 3-D picture of what the car’s sensors see and indicates what actions the vehicle is undertaking, such as steering or braking. There is also a button that passengers can press to stop the vehicle stop, perhaps if they get too nervous. A fisheye video camera embedded in the roof will record passengers’ reactions for later analysis.

The display was fascinating to watch, but it didn’t always offer a clear explanation of what the car was planning to do. Besides, it was more interesting to look out the window and see how the car was performing. Sometimes it passed parked cars at a distance that felt a little unnerving.

Uber's fleet of automated taxis.

I jumped at the chance to sit behind the wheel and see how drivers would experience the ride. It was possible to retake control by moving the steering wheel or pressing either the gas pedal or the brake, and there was even a large red button to disengage the system immediately. The biggest challenge, however, was staying absolutely vigilant. Sometime the car drove so adeptly that I had to be reminded to pay attention, and a couple of times I was asked to retake control—once when it seemed as though the car might travel extremely close to a parked vehicle, and again when it mysteriously began turning to the left while we sat in traffic on a bridge.

Pittsburgh is certainly a challenging environment for self-driving vehicles, one reason why Uber is testing its cars there. The city’s roads are narrow, winding, and often filled with pedestrians and cyclists. The city also has hills and bridges that can play tricks on a self-driving car’s sensors, and its climate will pose challenges. To date, most self-driving cars have been tested in states, like California and Nevada, where conditions are usually bright and sunny. Uber says its self-driving cars will be able to drive in rain and snow but will need extra supervision.

“Pittsburgh is the ideal environment for us to be testing in,” Krikorian said. “In a lot of ways, [it] is the double black diamond of driving.”

An engineer monitors systems from inside the car.

Uber’s vehicles are arrayed with sensors. There are seven lasers, including a spinning lidar on top; 20 different cameras at the front and sides of the vehicle; two radars that provide 360° coverage; GPS; and inertial measurement units.

Herman Herman, who directs Carnegie Mellon University’s National Robotics Engineering Center, a subunit of CMU created to help commercialize automated vehicles and other robotic technologies, is skeptical that Uber will be able to do away with drivers altogether in short order. Herman says Uber has an advantage in being able to choose where it sends its driverless cars, but he adds that guaranteeing safety and reliability will be particularly challenging. “The last thing you want if you’re a passenger is for it to stop, or to crash into something,” he says.

Virtually all automakers and a number of other companies, including Google and Apple, are developing automated driving systems, inspired by the potential impact the technology could have on the trillion-dollar transportation industry (see “Uber’s Pittsburgh Project Is a Crucial Test for Self-Driving Cars”). Uber’s ambitious effort reflects the upheaval being felt across the entire transportation industry. With automakers and other companies exploring the ride-sharing business, Uber, which is valued in excess of $50 billion, clearly feels the need to stay ahead of the technology curve to ensure it does not get left behind.

Uber is testing this automated taxi in Pittsburgh.

The company has developed its automated cars very quickly. Uber itself was founded in the same year that Google began developing its own self-driving cars, in 2009, and the company’s Advanced Technology Center in Pittsburgh was founded only 18 months ago. Uber spent $680 million to create that center, luring dozens of engineers away from CMU’s prestigious robotics department in the process.

As my drive handily demonstrated, it will take time for Uber and others to perfect fully automated driving. In fact, it remains unclear what needs to be done before it can be considered safe to remove humans from the driver’s seat. And while riding in the cars is a futuristic thrill, for the time being I think I’d prefer to hail a human driver, no matter how chatty.

This story has been updated from its original version to include more details about the reporter’s hands-on experience.

 

Keep Reading

Most Popular

Scientists are finding signals of long covid in blood. They could lead to new treatments.

Faults in a certain part of the immune system might be at the root of some long covid cases, new research suggests.

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.