Uber’s semi-autonomous car detected the pedestrian six seconds before the fatal crash, a federal agency says

A federal agency has released its preliminary report after investigating a fatal crash involving an Uber semi-autonomous Volvo in Tempe, Ariz.

After analyzing data obtained from the Uber vehicle, the National Transportation Safety Board found that the Uber system detected the pedestrian, Elaine Herzberg, six seconds before the crash. The report also reveals that Uber relies heavily on its vehicle operators to brake in the case of emergencies.

The preliminary report does not yet indicate probable cause or fault. The NTSB is still working with Uber, Volvo and the Arizona Department of Transportation to complete its investigation into the crash so that the agency can eventually issue safety recommendations.

Uber’s self-driving tests continue to be on hold. But the report brings an ongoing debate about the safety of semi-autonomous systems to the forefront. Some industry experts say it’s actually less safe to introduce technology that only takes over some of the driving task, since it relies on two imperfect systems: Technology that is inherently limited in its capabilities, and humans.

In the case of Uber’s system, the NTSB wrote that the company’s automatic emergency braking system does not work when the semi-autonomous system is operating the vehicle “to reduce the potential for erratic vehicle behavior,” the company told the agency. Instead, the system depends on the safety driver to intervene.

But, as the report goes on to say, the system has no means of alerting the driver.

The other facts the NTSB is looking at to determine the cause of the crash in addition to how the system operated include how the vehicle operator handled the crash and whether the vehicle operator was impaired.

The NTSB said that the system itself initially had difficulty categorizing Herzberg. Initially the system categorized Herzberg, who was walking her bike across the road, as an unknown object, then a vehicle and then as a bike. About 1.3 seconds before the crash, the NTSB wrote, the Uber system determined that it needed to enable its emergency braking system.

“The self-driving system data showed that the vehicle operator intervened less than a second before impact by engaging the steering wheel,” the report states. “The vehicle speed at impact was 39 mph. The operator began braking less than a second after the impact. The data also showed that all aspects of the self-driving system were operating normally at the time of the crash, and that there were no faults or diagnostic messages.”

Uber’s own internal investigation into the crash also shows that the system detected Herzberg and did not react in time. But the company had tuned its system to not abruptly brake for “false positives” or objects that the vehicle can drive through.

“Over the course of the last two months, we’ve worked closely with the NTSB,” an Uber spokesperson said. “As their investigation continues, we’ve initiated our own safety review of our self-driving vehicles program. We’ve also brought on former NTSB Chair Christopher Hart to advise us on our overall safety culture, and we look forward to sharing more on the changes we’ll make in the coming weeks.”

The report also says Herzberg was wearing dark clothes and her bike didn’t have any reflectors, making it potentially difficult for a vision system to detect her. The NTSB is continuing to look at the drivers’ business and personal phones which she kept with her.

Video footage of the operator during the crash shows her looking down several times before the crash. The vehicle operator told the agency in an interview that she was merely monitoring the self-driving interface.

The agency is still looking for information about Herzberg and said toxicology reports show that she tested positive for methamphetamine and marijuana.

This article originally appeared on Recode.net.

Will you support Vox’s explanatory journalism?

Most news outlets make their money through advertising or subscriptions. But when it comes to what we’re trying to do at Vox, there are a couple reasons that we can't rely only on ads and subscriptions to keep the lights on.

First, advertising dollars go up and down with the economy. We often only know a few months out what our advertising revenue will be, which makes it hard to plan ahead.

Second, we’re not in the subscriptions business. Vox is here to help everyone understand the complex issues shaping the world — not just the people who can afford to pay for a subscription. We believe that’s an important part of building a more equal society. We can’t do that if we have a paywall.

That’s why we also turn to you, our readers, to help us keep Vox free. If you also believe that everyone deserves access to trusted high-quality information, will you make a gift to Vox today?

One-Time Monthly Annual

$5/month

$10/month

$25/month

$50/month

Other

$

Yes, I'll give $5/month

Yes, I'll give $5/month

We accept credit card, Apple Pay, and Google Pay. You can also contribute via