Uber self-driving crash has made manufacturers ‘cautious’ on autonomy

There are a lot of carmakers making some bold claims about what their autonomous cars will be able to do.

We’ve seen some bold claims from car manufacturers about their autonomous programs. General Motors wants to deliver a car without pedals or a steering wheel in 2019, while BMW and Ford are targeting 2021 with their self-driving programs.

But the push to automation hasn’t been smooth – far from it, in fact. Earlier this year, an Uber test vehicle struck and killed a pedestrian in Tempe, Arizona, garnering the attention of media around the world.

Tesla has also had trouble with drivers crashing while Autopilot is engaged, most notably in an accident that killed an Apple software engineer.

Speaking with CarAdvice at the third International Driverless Vehicle Summit in Adelaide last week, Adwait Kale, business development and project development manager for EasyMile Asia Pacific, suggested the reaction to the Uber self-driving death will make manufacturers more cautious than they otherwise would’ve been in rolling out self-driving vehicles.

“Certain manufacturers are quite aggressive in terms of their strategy,” Kale explained. “Certain brands, like Lexus, are very safety-focused.”

“I strongly believe that everyone, especially after the Volvo/Uber incident, everyone is going to be very, very careful in rolling these out,” he later argued. “They might offer a solution, but there will be a lot of caveats behind it.”

“From my perspective, it would be very hard to believe they’d have a fully-autonomous self-driving car by 2022,” he said, referencing both BMW and Ford’s stated goals for automation.

EasyMile is a company specialising in autonomous software for last-mile projects. It’s the supplier for the just-announced shuttle trial in rural South Australia, and is involved in similar projects around the world.

As for the race to deliver a fully self-driving car? Kale said it’s more of a question of definitions than anything else, and argued true level five automation could be eight years away.

“The key difference is, we are at level four automation,” he said, later adding, “a self-driving vehicle available for the consumer, which can just run by itself, I think it’s still seven years in the making. Seven to eight years.”

“The reason being there are so many different aspects you need to take into consideration… lane-changes, traffic, environmental changes. To be able to cater for that, the technology, our AI, needs to evolve.”