x
AI/Automation

Tesla's 5G stop sign: Self-driving needs 'no connectivity whatsoever'

The global wireless industry continues to salivate at the prospect of selling high-speed, low-latency and revenue-generating services to automotive companies and motorists in support of the much-ballyhooed autonomous driving services of the future.

But Tesla – the world's most valuable car company by a shockingly wide margin – is boasting of a new self-driving technology that does not need a cellular connection at all.

"We put it out there last night, and then we'll see how it goes and then probably release it to more people this weekend or early next week," said Elon Musk, Tesla's billionaire CEO, of the company's new Full Self-Driving (FSD) service. "And then just gradually step it up until we have, hopefully, a wide release by the end of this year."

And Musk argued that one of the best things about FSD is that it does not require a cellular connection – 5G, 4G or even 3G.

"There is no need for high-definition maps or a cellphone connection. So the car – the system is designed such that even if you have no connectivity whatsoever and you're in a place that you have never been to before and no Tesla has ever been there, the car should still be able to drive, just like a person," Musk said this week during Tesla's quarterly conference call, according to a Seeking Alpha transcript of the event. "That is the system that we are developing and aiming to release this year."

Musk's announcement coincided with the release of what he described as Tesla's "best quarter in history." The Wall Street Journal noted Tesla reported a net profit of $331 million for the third quarter, putting the company on track to make 2020 its first year in the black.

Investors cheered, to put it mildly. Tesla's market value is now approaching $400 billion – roughly five times the combined value of Ford and General Motors.

High tech or hyperbole?

However, there are plenty of caveats to Tesla's performance and Musk's Full Self-Driving promises. Indeed, the WSJ pointed out that Musk is prone to hyperbole: For example, he unveiled a Tesla semi truck and second-generation Roadster sports car in 2017, and neither has yet made an appearance.

FSD, a technology Tesla first announced in 2016, might be different. "FSD beta rollout happening tonight. Will be extremely slow & cautious, as it should," Musk tweeted this week.

(It's worth noting that Twitter also remains one of the few sources of new information about Tesla and FSD; Cnet reported Tesla has completely eliminated its public relations department.)

According to Tesla, FSD can perform a number of advanced maneuvers, including navigating interstate onramps and lane changes (dubbed "Navigate on Autopilot" by Tesla), driving through parking lots (Smart Summon) and slowing for stop signs (Traffic and Stop Sign Control). Such capabilities are astounding to those currently driving 1998 Toyotas with more than 300,000 miles, like this author, but they fall far short of full-blown "level 5" autonomous driving, which doesn't require a driver at all.

Moreover, Tesla's commercial introduction of FSD is raising some concerns. After all, Consumer Reports ran detailed tests on the service and found that some functions don't perform anywhere near Tesla's promises.

"Even when Smart Summon was able to arrive at the specified location, it would spend time in the wrong lane of the parking lot, not pause for parking-lot stop signs, or take wide turns and head toward parked cars and then drive in Reverse to avoid a collision," the publication wrote.

Yet FSD apparently is hitting the road in real-world situations right now,

"Public road testing is a serious responsibility and using untrained consumers to validate beta-level software on public roads is dangerous and inconsistent with existing guidance and industry norms," argued the Partners for Automated Vehicle Education (PAVE) association, a coalition of nonprofits, to the Washington Post. "Moreover, it is extremely important to clarify the line between driver assistance and autonomy. Systems requiring human driver oversight are not self-driving and should not be called self-driving."

Learning to avoid 5G

Interestingly, Musk explained that FSD will work like Google's search: It will get better the more it's used.

"Having on the order of 1 million cars that are providing feedback and specifically feedback on strange corner case situations that you just can't even come up with in simulation" will help to improve the system, Musk said on Tesla's earnings call. (Tesla inked a deal with AT&T in 2014 to connect its cars to the Internet.) "This is the thing that is really valuable."

Musk is pinning most of his self-driving hopes on the ability of FSD to learn to drive properly. The technology basically watches the road via a handful of cameras festooning the company's vehicles, routing that optical information to a "neural network" that can then react to obstacles.

As the Washington Post pointed out, Tesla's FSD doesn't even use Lidar like many other autonomous driving systems. Lidar, which typically requires expensive equipment, uses light from a laser to create a type of radar readout of an environment.

Musk is not a fan of Lidar. He said Tesla wouldn't use the technology even if it were free.

In discussing the technologies involved in autonomous driving, Musk said Tesla is focusing on optical solutions because once you design a system that works via optical inputs, "why bother with anything else?"

It's that statement, from the CEO of the world's most valuable automobile company, that ought to give the mobile industry pause. After all, autonomous vehicles remain one of the top 5G use cases for a wide range of vendors.

Tesla is set to charge $10,000 for FSD. 5G, meanwhile, remains free to users on most top-tier mobile plans.

Related posts:

Mike Dano, Editorial Director, 5G & Mobile Strategies, Light Reading | @mikeddano

Be the first to post a comment regarding this story.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE