Developing self-driving car technology is an expensive endeavor. Just ask General Motors, Google, or Uber, all of which have dedicated millions toward creating autonomous driving systems. But does taking driving duties from humans and handing it over to robots have to be so spendy? One British tech startup doesn’t think so.
It’s called Wayve. And, in a recent company blog, the company boasted that is developing a self-driving system that relies on just a camera and GPS-based navigation.
What’s Wayve’s trick, and the advantage it has over other companies? Machine learning, which the company says, “is surpassing hand-engineered systems everywhere.”
Outfitted with only a forward-facing camera and navigation software, Wayve’s test car was able to traverse roads in Cambridge, England that it’d never traveled before. You can see snippets from this test run in the company-provided video below.
If Wayve’s reasoning sounds specious to you, it should. That’s because the more big brands like GM or Apple dig into the tasks of automated driving, the more sensors and systems they throw at their cars.
Just this week we learned that Apple has patented sensors — likely intended for self-driving cars — that scan the road beneath the vehicle. These are designed to measure whether the vehicle is turning as it intended or whether it’s, for example, slipping on ice. There’s little chance Wayve’s single-camera setup can measure that sort of data, let alone correct for it.
It’s nice that there are more than just legacy automakers or tech giants in the automated driving system development market — especially ones keen to bring down the cost of self-driving tech. However, they should rely less on hubris and err more to the side of caution when it comes to roadway safety.