Remote ready: Is teleoperation the missing link for driverless cars?
Hello, automated driving community! Remote controlled cars for grown-ups, GM avoids a court trial and a Japanese company is on a shopping spree: we bring you this week’s key stories from the world of automated driving!
No doubt: autonomous driving is evolving at an incredible speed. But even Silicon Valley’s eternal optimists would agree that it’s not yet safe for mass deployment – like Shai Magzimof, CEO of Mountain View based start-up Phantom Auto. “An autonomous vehicle company might have a system that works 95 or even 99 percent of the time, but that last 1 percent is a very difficult piece of the puzzle to solve,” Magzimof told IEEE Spectrum in January. However, Magzimof believes that his company has found the solution: teleoperation.
During autonomous test rides, there are still many tricky situations like construction sites, gravel roads or the double lane merge. They confuse machines, so a human driver needs to be kept in the loop to master those tasks. The premise behind teleoperation is that yes, we do need a human driver, but he doesn’t necessarily need to be in the car. If a self-driving vehicle encounters a situation it can’t handle on its own, it enables a remote human operator to take over. The operator sits in a control center in front of several screens, providing a 360-degree view of what surrounds the vehicle. His console is fitted with a steering wheel, brake and accelerator pedals. Phantom Auto just presented its system to the public (see video below) and also announced three deals with Chinese firm NEVS (formerly SAAB Automobile) and European companies Einride and Transdev who will use the technology for their testing.
Teleoperated driving is not an idea only pursued by Phantom Auto. Daimler and Bosch are using it for their automated valet parking project and Nissan announced its Seamless Autonomous Mobility system (SAM) a while ago. The technology does have its pitfalls: for instance, communication with the control center requires extremely low latency. Any delay due to weak data connections could provide fatal when driving at highway speeds. But it has the potential to allow for much safer testing. And not only that: many customers still hesitate to hand over control to a machine. Wouldn’t it be reassuring to know that when the going gets tough, there is a human operator who has your back?
No verdict: GM and the crash
It’s impossible to say whether teleoperated assistance might have prevented the accident that occurred last December. Back then, an autonomous Chevy Bolt collided with a motorbike on a Californian highway after an unsuccessful lane change. What’s clear is that the crash had the potential to have a far-reaching impact. The motorcycle rider suffered minor injuries and subsequently sued General Motors. It was shaping up to be one of the first court trials involving an accident with an autonomous car.
Except there won’t be a trial. Last week, GM and the motorcyclist’s attorney announced that they have settled the lawsuit out of court. GM acknowledged the aborted lane change in a report filed with the California Department of Motor Vehicles, but claims that the rider had been driving between two lanes when he “moved into the center lane, glanced the side of the [car]… wobbled, and fell over.”
The settlement doesn’t necessarily mean that GM was at fault. In any case, it might have been the most pragmatic move for the carmaker to put this matter to rest as quickly and quietly as possible in the interest of its automated driving program. Trials including driverless cars are still uncharted waters, so companies have a lot at stake. The Waymo lawsuit against Uber was a cautionary example: it tied up the ride-hailing firm’s resources for months, blocking development. That’s why we can expect more out-of-court settlements in similar cases to come.
Softbank: Discreetly pulling the strings
GM can now focus on further speeding up the development of their driverless car unit Cruise Automation. Funding should not be an issue as Japanese conglomerate SoftBank’s Vision Fund just invested 2.25 billion U.S. dollars in Cruise. SoftBank now owns 20 percent of the company and has been on a massive shopping spree lately – splashing out on autonomous driving. Other investments include big shots like Nvidia (5 percent), Uber (15 percent) and Didi Chuxing (20 percent), among many others. SoftBank also fully acquired Arm Holdings for an eye-popping 31 billion U.S. dollars back in 2017. The U.K. chipmaker produces all kinds of automotive sensors, safety features and automotive and autonomous products.
SoftBank clearly puts high hopes in autonomous driving eventually yielding massive profits. There are many players vying for leadership in the realm of automated driving: be it vehicle manufacturers, software providers, mobility providers or chipmakers. As it looks, SoftBank aims to be the number one investor behind those players.
So long, drive safely (until cars are driverless),