Tag: Autopilot

A fatal crash, production problems and now a recall.

Tesla is starting the second quarter in a defensive crouch:

• Last week, the company revealed that Autopilot, its semi autonomous feature, was engaged during a recent fatal crash in California — its second confirmed Autopilot-related fatality in the U.S.
• Tesla is struggling to meet its production goals for the Model 3, its first-ever mass-market car. Today, CEO Elon Musk reportedly said the company is producing 2,000 Model 3s a week — 500 short of his goal, which has been adjusted twice.
• Last week, Tesla voluntarily recalled 123,000 of its Model S luxury sedans to fix a power-steering issue. That is a lot of cars — close to half of all the vehicles the company has produced.
• Tesla stock is down about 36 percent since its September 2017 peak.

By the company’s own admission, this is a critical time for Tesla. The electric vehicle movement the company arguably popularized is seeing momentum from new and existing players, while self-driving competitors like Alphabet’s Waymo strike deals with automakers to develop vehicles that could rival Tesla’s own offerings. As both an automaker and a self-driving tech company, Tesla still has a lot to prove.

The crash

It’s not yet known whether Autopilot was at fault for 38-year-old Tesla driver Walter Huang’s death, but the simple fact that it was involved has put Tesla’s already fraught future — as well as the self-driving industry — at risk.

On March 23, Huang crashed his Model X into a median on a California highway while the SUV was operating in Autopilot mode. Tesla recovered the logs from the vehicle, and upon analyzing them said that the driver had received “several visual and one audible” cue to take back control of the car.

“The driver had about five seconds and 150 meters of unobstructed view of the concrete divider with the crushed crash attenuator, but the vehicle logs show that no action was taken,” the company wrote in a blog post.

This is the second U.S. crash of a Tesla confirmed to be operating Autopilot that has led to a fatality. The first was in Williston, Fla., in May 2016.

The National Transportation Safety Board, which is also investigating the March 23 crash, found that the first Autopilot-related fatality in 2016 was in part a result of the driver overrelying on Tesla’s semiautonomous software, but that Autopilot operated the way it was supposed to.

The NTSB’s investigation into this crash is ongoing, but the agency said that it was “unhappy” that Tesla revealed the details of the investigation to the public. The NTSB is also looking into reports that the driver previously complained about the performance of the Autopilot software.

Relatives of Huang said that he took his Tesla to the dealership because the software caused the car to swerve toward the highway barrier that his vehicle ultimately crashed into.

A Tesla spokesperson declined to comment on the NTSB’s comments but said they found no record of Huang bringing the vehicle into a dealership to service its Autopilot software.

“We’ve been doing a thorough search of our service records and we cannot find anything suggesting that the customer ever complained to Tesla about the performance of Autopilot,” a Tesla spokesperson said in a statement. “There was a concern raised once about navigation not working correctly, but Autopilot’s performance is unrelated to navigation.”

The fallout

The tragic death comes as both the industry and Tesla brace for the fallout from a recent fatality that involved an Uber-operated semi-autonomous vehicle in Tempe, Ariz.

The NTSB, along with local police and the National Highway Traffic Safety Administration, is also investigating the Uber crash, which resulted in the death of 49-year-old Elaine Herzberg.

Both crashes hit at a larger question many in the industry have: Is semi-autonomous technology safe?

With Uber and Tesla being two of the most prominent brands in the auto and tech industry working on some version of self-driving, consumer trust in the new technology could take a hit.

When it launched Autopilot, Tesla set the benchmark for the most advanced adaptive cruise control available in consumer vehicles. That technology has received multiple updates, and Musk has said he expects the second generation of the software to be capable of a high level of self-driving in about two years.

However, as it exists today, Autopilot is not intended to operate in all circumstances, and in fact is limited to highway driving. In other words, drivers need to be alert and ready to take over at all times — which creates an odd situation that is now clearly prone to failure.

That was also the case in Uber’s crash: The system relies on a trained operator to take over when the technology doesn’t work, though there are some important distinctions that need to be made between the two. For instance, Uber’s technology, which is still in development, is intended to operate on local roads with variables including pedestrians. Tesla’s Autopilot is only supposed to ease the highway-driving task.

Uber’s vehicles, however, are not available to the wider public, and are not being sold direct to consumers. Tesla, which says its technology is also still in beta, is putting its technology in the hands of consumers. Still, if either of the companies’ semiautonomous software is found to be at fault, there could be a resounding impact on consumer trust around self-driving.

“The consequences of the public not using Autopilot, because of an inaccurate belief that it is less safe, would be extremely severe,” Tesla wrote in a blog post. “There are about 1.25 million automotive deaths worldwide. If the current safety level of a Tesla vehicle were to be applied, it would mean about 900,000 lives saved per year.”

Production woes

Tesla’s voluntary recall of 123,000 Model S cars punctuated its ongoing struggles with meeting production goals of its mass-market vehicle, the Model 3.

The Model 3 is a significant barometer by which investors and the industry are measuring Tesla’s capability as an automaker. Can Tesla make the shift away from being just a luxury player to a mass-market carmaker at scale?

By Musk’s own admission, the early years of Tesla — from the Roadster to the Model X — were in service of laying the groundwork for building and selling a mass-market electric vehicle.

But the company has gotten off to a rough start in meeting the many ambitious goals Musk has set for the production of the vehicle.

In July 2017, Musk said that he aimed to produce 5,000 Model 3 vehicles per week by the end of 2017. The company then shifted that rate goal to 5,000 cars per week by the end of March 2018. But then in January, Musk lowered that goal to 2,500.

Today, Tesla is producing 2,000 Model 3s a week, according to emails obtained by Jalopnik.

“If things go as planned today, we will comfortably exceed that number over a seven-day period!” Musk wrote, referring to the current rate of production.

The company’s head of engineering also tried to rally the troops last week, saying the company needed to prove the “haters” wrong, as Bloomberg first reported.

“The world is watching us very closely, to understand one thing: How many Model 3s can Tesla build in a week?” Doug Field wrote. “This is a critical moment in Tesla’s history, and there are a number of reasons it’s so important. You should pick the one that hits you in the gut and makes you want to win.”

Recode – All