Tesla's Autopilot engaged during Utah crash

May 15, 2018 by Julian Hattem
In this Friday, May 11, 2018, photo released by the South Jordan Police Department shows a traffic collision involving a Tesla Model S sedan with a Fire Department mechanic truck stopped at a red light in South Jordan, Utah. Witnesses indicated the Tesla Model S did not brake prior to impact. Police Sgt. Samuel Winkler said the car's air bags were activated and that the Tesla's 28-year-old driver suffered a broken right ankle, while the driver of the mechanic truck didn't require treatment. Police in a Salt Lake City suburb say it's not immediately known whether a Tesla Model S sedan's semi-autonomous Autopilot driving system was in use when it rear-ended a truck apparently without braking before impact at approximately 60 mph. (South Jordan Police Department via AP)

The driver of a Tesla electric car had the vehicle's semi-autonomous Autopilot mode engaged when she slammed into the back of a Utah fire truck over the weekend, in the latest crash involving a car with self-driving features.

The 28-year-old driver of the car told in suburban Salt Lake City that the system was switched on and that she had been looking at her phone before the Friday evening crash.

Tesla's Autopilot system uses radar, cameras with 360-degree visibility and sensors to detect nearby cars and objects. It's built so cars can automatically change lanes, steer, park and brake to help avoid collisions.

The auto company markets the system as the "future of driving" but warns to remain alert while using Autopilot and not to rely on it to entirely avoid accidents. Police reiterated that warning Monday.

A Tesla spokesperson did not comment following the disclosure about the use of the feature.

On Twitter, co-founder Elon Musk said it was "super messed up" that the incident was garnering public attention, while thousands of accidents involving traditional automobiles "get almost no coverage."

South Jordan police said the Tesla Model S was going 60 mph (97 kph) when it slammed into the back of a fire truck stopped at a red light. The car appeared not to brake before impact, police said.

In this Friday, May 11, 2018, photo released by the South Jordan Police Department shows a traffic collision involving a Tesla Model S sedan with a Fire Department mechanic truck stopped at a red light in South Jordan, Utah. Witnesses indicated the Tesla Model S did not brake prior to impact. Police Sgt. Samuel Winkler said the car's air bags were activated and that the Tesla's 28-year-old driver suffered a broken right ankle, while the driver of the mechanic truck didn't require treatment. Police in a Salt Lake City suburb say it's not immediately known whether a Tesla Model S sedan's semi-autonomous Autopilot driving system was in use when it rear-ended a truck apparently without braking before impact at approximately 60 mph. (South Jordan Police Department via AP)

The driver, whom police have not named, was taken to a hospital with a broken foot. The driver of the fire truck suffered whiplash and was not taken to a hospital.

"What's actually amazing about this accident is that a Model S hit a fire truck at 60 mph and the driver only broke an ankle," Musk tweeted. "An impact at that speed usually results in severe injury or death."

The National Transportation Safety Board has not opened an investigation into the crash, spokesman Keith Holloway said, though it could decide to do so.

Over the past two months, federal officials have opened investigations into at least two other crashes involving Tesla vehicles.

In this Friday, May 11, 2018, photo released by the South Jordan Police Department shows a traffic collision involving a Tesla Model S sedan with a Fire Department mechanic truck stopped at a red light in South Jordan, Utah. Witnesses indicated the Tesla Model S did not brake prior to impact. Police Sgt. Samuel Winkler said the car's air bags were activated and that the Tesla's 28-year-old driver suffered a broken right ankle, while the driver of the mechanic truck didn't require treatment. Police in a Salt Lake City suburb say it's not immediately known whether a Tesla Model S sedan's semi-autonomous Autopilot driving system was in use when it rear-ended a truck apparently without braking before impact at approximately 60 mph. (South Jordan Police Department via AP)

Last week, the NTSB opened a probe into an incident in which a Model S caught fire after crashing into a wall in Florida.

Two 18-year-olds were trapped in the vehicle and killed in the flames. The agency has said it does not expect the semi-autonomous system to be a focus of that investigation.

The NTSB and the National Highway Traffic Safety Administration are also looking into the performance of the company's Autopilot system in the March crash of a Tesla Model X on a California highway. The driver in that incident died.

In March, an Arizona pedestrian was killed by a self-driving Uber car, in the first death of its kind. A driver was behind the wheel of the test vehicle in that case but failed to halt in time.

In this Friday, May 11, 2018, photo released by the South Jordan Police Department shows a traffic collision involving a Tesla Model S sedan with a Fire Department mechanic truck stopped at a red light in South Jordan, Utah. Witnesses indicated the Tesla Model S did not brake prior to impact. Police Sgt. Samuel Winkler said the car's air bags were activated and that the Tesla's 28-year-old driver suffered a broken right ankle, while the driver of the mechanic truck didn't require treatment. Police in a Salt Lake City suburb say it's not immediately known whether a Tesla Model S sedan's semi-autonomous Autopilot driving system was in use when it rear-ended a truck apparently without braking before impact at approximately 60 mph. (South Jordan Police Department via AP)

The investigation into the in Utah is ongoing, police said.

The driver of the Tesla may face charges for failing to maintain the safety of her vehicle, which would be a traffic infraction, according to police spokesman Sgt. Samuel Winkler.

In this Friday, May 11, 2018, photo released by the South Jordan Police Department shows a traffic collision involving a Tesla Model S sedan with a Fire Department mechanic truck stopped at a red light in South Jordan, Utah. Witnesses indicated the Tesla Model S did not brake prior to impact. Police Sgt. Samuel Winkler said the car's air bags were activated and that the Tesla's 28-year-old driver suffered a broken right ankle, while the driver of the mechanic truck didn't require treatment. Police in a Salt Lake City suburb say it's not immediately known whether a Tesla Model S sedan's semi-autonomous Autopilot driving system was in use when it rear-ended a truck apparently without braking before impact at approximately 60 mph. (South Jordan Police Department via AP)
In this Friday, May 11, 2018, photo released by the South Jordan Police Department shows a traffic collision involving a Tesla Model S sedan with a Fire Department mechanic truck stopped at a red light in South Jordan, Utah. Witnesses indicated the Tesla Model S did not brake prior to impact. Police Sgt. Samuel Winkler said the car's air bags were activated and that the Tesla's 28-year-old driver suffered a broken right ankle, while the driver of the mechanic truck didn't require treatment. Police in a Salt Lake City suburb say it's not immediately known whether a Tesla Model S sedan's semi-autonomous Autopilot driving system was in use when it rear-ended a truck apparently without braking before impact at approximately 60 mph. (South Jordan Police Department via AP)

Explore further: Tesla feuds with one federal agency, cooperates with another

Related Stories

Recommended for you

Permanent, wireless self-charging system using NIR band

October 8, 2018

As wearable devices are emerging, there are numerous studies on wireless charging systems. Here, a KAIST research team has developed a permanent, wireless self-charging platform for low-power wearable electronics by converting ...

Facebook launches AI video-calling device 'Portal'

October 8, 2018

Facebook on Monday launched a range of AI-powered video-calling devices, a strategic revolution for the social network giant which is aiming for a slice of the smart speaker market that is currently dominated by Amazon and ...

Artificial enzymes convert solar energy into hydrogen gas

October 4, 2018

In a new scientific article, researchers at Uppsala University describe how, using a completely new method, they have synthesised an artificial enzyme that functions in the metabolism of living cells. These enzymes can utilize ...

6 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

EyeNStein
not rated yet May 15, 2018
It will be interesting to hear if a fire truck getting closer to the front of the car at 60mph could be read as a 'false positive'?? Unless the system saw the issue and yielded authority back but the driver was too slow taking over.

The coding that arbitrates and compares the Inputs from the multiple sensor systems may need attention. One sub-systems misread could have de-stabilised the whole system.
antialias_physorg
3 / 5 (2) May 15, 2018
The coding...may need attention.

The coding works like this: You have a number of training samples that are represented by a vector in n dimensional feature-space (where the n dimensions can be stuff like color, speed, noise, acceleration...whatever you measure). Then you let an algorithm run on this (usually a neural network of some kind, or other clustering algorithms) that sorts all the vector endpoints into feature clusters.

When you then encounter an object during real driving you check which cluster it best fits into.

There's 2 issues with this:
1) Clusters can overlap - so a classification into a unique class may not be possible (e.g. cancer cells vs. non-cancer cells show morphologies that aren't totally distinct)
2) A radical deviation can prevent correct classification (e.g. cars in training images are nonreflective and you meet a mirrored car)

Tweaking these is not really easy because making one cluster more distinct may make another one less so.
EyeNStein
5 / 5 (1) May 15, 2018
The investigations of these high profile AI failures will hopefully open up whole new avenues in general AI research.
A way has to be developed to recognise and process these anomalies and indistinct case overlaps before this field can move forward safely.
BobSage
not rated yet May 15, 2018
I test drove a Tesla. I used the turn indicator and let the car decide to change lanes by itself. And then I thought, "Are you nuts?".

This must be the age of trading self-reliance for being taken care of.
DonGateley
not rated yet May 15, 2018
Elon's complaint about coverage of the accident shows a remarkable tunnel vision of its own. He just fell more than a few notches in my estimation. As did the credibility of self driving cars.

OTOH, his noting the lack of major injury from a 60 mph crash does shine a justifiably favorable light on the crashworthiness of the car itself.
leetennant
5 / 5 (2) May 15, 2018
Elon's complaint about coverage of the accident shows a remarkable tunnel vision of its own. He just fell more than a few notches in my estimation. As did the credibility of self driving cars.

OTOH, his noting the lack of major injury from a 60 mph crash does shine a justifiably favorable light on the crashworthiness of the car itself.


His position is justified from a purely statistical perspective.

But everyone knows that humans are flawed and when car crashes happen there's a sense of rightness that an individual was responsible for their own decisions leading up to it. As soon as we delegate this responsibility we're going to have an existential fear about it. This car didn't crash because of speeding, drunk driving, fatigue, youth, inexperience or rain. It crashed because of an inadequate algorithm.

Statistics aren't the only issue here.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.