Investigating Autopilot

I agree on a universal standard. Currently KIA uses a warning alarm and the system that is malfunctioning flashes red on the drivers Instrument Cluster - so a visual and audible alarm.

I find the Rear Cross Traffic and Blind Spot Monitors to help me the most. Our vehicle is a 2019 and KIA doesn't let you take your hands off the wheel, if you do the car sounds an alarm and will begin to slow down - the detection interval seems to be about 10 seconds with no hands on the wheel.

The vehicle also has an alarm for a reckless driver approaching at high speed and warns you about it ( approaching from the rear or side)- this alarm took some getting used to.

 

Some bits of the NHTSA of the report were leaked.
Data likely shows Teslas on Autopilot crash more than rivals | AP News

 

Compared to other automakers, Tesla has over ¾ million Individual Car's data. How many cars do other manufacturers have, that are running different forms of driver-assist? The NTSA has to Calc rates per 100,000 if there's to be any meaningful interpretation.
.

 

They've started doing that.

 

How many other car companies even HAVE 100,000 vehicles w/ serious driver assist?
.

 

For the math impaired, that AP article points to crashes per just 1,000 vehicles. All the big car companies appear to have that many.

============

Since Trollbait's AP news link back at #23, the AP appears to have received crash figures from more car companies:
US report: nearly 400 crashes of automated tech vehicles | AP News

 

You have to define serious.

Subaru introduced Eyesight with the 2013MY. IIRC, that's the same year the Model S came out. Subaru sold over 424k cars that year in the US, getting to 700k before dropping for the pandemic years. Eyesight wasn't standard then, but it is now on cars with a CVT or automatic. I think it's a good bet they have over 100k vehicles on the road with 'serious' driver assist.

Volvo started introducing automatic braking back around 2009.

In the above article, Honda reports 90 crashes, and states they have 6 million cars on US roads with advanced driver assist systems.

What I want to know is how many crashes have been reported for Open Pilot?

PS: Subaru reported 10 crashes within the time frame.

 

One has to wonder how pathetically small is a 'per-1,000' accident sample is considering there are 17,000 accidents on average per day in the United States.
How Many Car Accidents Are There in the USA per Day?.
.

 

A comparison of accidents per vehicle mile traveled per model year per car maker and including so-called self-driving and person driven categories would be illuminating. Cheap headlines with Tesla in them just don't help me.

 

Isn't the open pilot kit strictly DIY? Seems that subset of drivers would naturally be more in tune to using driver assist.
.

 

First off, "accident" is a euphemism, to avoid the reality that the overwhelming majority of vehicle crashes are caused by human misconduct or negligence. Many entities have retired that word, now calling them crashes or collisions.

For the math impaired again, "per-1,000" or "per-100,000" doesn't mean they look at just a sample of 1,000 or 100,000 vehicles. They are looking at the entire population, then expressing the rate on a comparable basis.

==================
Separately, I hope readers also see this paragraph: "[2019] An additional 2.74 million people suffered injuries that same year, according to ... NHTSA ... There are 7,507 people injured in the United States in car accidents every day." I leave out the article's listed death toll because it has risen sharply since that year, and NHTSA's count excludes a lot of MV-associated deaths that I prefer to include. CDC has better numbers for MV deaths.

 

The NHTSA didn't just request reports from automakers, but also tech companies working in the advanced driver assistant systems field. Don't know if the request went out to the Open Pilot company, but the article says it was a hundred companies.

 

WaPo: 17 fatalities, 736 crashes: The shocking toll of Tesla’s Autopilot
Tesla’s driver-assistance system, known as Autopilot, has been involved in far more crashes than previously reported

"SAN FRANCISCO — The school bus was displaying its stop sign and flashing red warning lights, a police report said, when Tillman Mitchell, 17, stepped off one afternoon in March. Then a Tesla Model Y approached on North Carolina Highway 561.

The car — allegedly in Autopilot mode — never slowed down. ..."

"The number of deaths and serious injuries associated with Autopilot also has grown significantly, the data shows. When authorities first released a partial accounting of accidents involving Autopilot in June 2022, they counted only three deaths definitively linked to the technology. The most recent data includes at least 17 fatal incidents, 11 of them since last May, and five serious injuries."

"Tesla’s 17 fatal crashes reveal distinct patterns, The Post found: Four involved a motorcycle. Another involved an emergency vehicle. Meanwhile, some of Musk’s decisions — such as widely expanding the availability of the features and stripping the vehicles of radar sensors — appear to have contributed to the reported uptick in incidents, according to experts who spoke with The Post."

"Since the reporting requirements were introduced, the vast majority of the 807 automation-related crashes have involved Tesla, the data show. Tesla — which has experimented more aggressively with automation than other automakers — also is linked to almost all of the deaths."

"The uptick in crashes coincides with Tesla’s aggressive rollout of Full Self-Driving, which has expanded from around 12,000 users to nearly 400,000 in a little more than a year. Nearly two-thirds of all driver-assistance crashes that Tesla has reported to NHTSA occurred in the past year."

"While Tesla constantly tweaked its driver-assistance software, it also took the unprecedented step of eliminating its radar sensors from new cars and disabling them from vehicles already on the road — depriving them of a critical sensor as Musk pushed a simpler hardware set amid the global computer chip shortage. ... It has recently taken steps to reintroduce radar sensors, according to government filings first reported by Electrek."

"In a March presentation, Tesla claimed Full Self-Driving crashes at a rate at least five times lower than vehicles in normal driving, in a comparison of miles driven per collision. That claim, and Musk’s characterization of Autopilot as “unequivocally safer,” is impossible to test without access to the detailed data that Tesla possesses.

Autopilot, largely a highway system, operates in a less complex environment than the range of situations experienced by a typical road user."


The headline and some of the article's tone may be a bit excessive. And I'm still very reluctant to make comparisons between Tesla and other makers, and also between AP/FSD and regular humans, due to various potentially large selection biases. But the recent sharp increase in Tesla's driver-assistance crash and fatality figures seems concerning.

 

The article seems to leave out basic information. I don’t know if this is because they don’t have it or chose not to share it.
The question I have is, what is the number of vehicles equipped with the level of automation being tracked, for each manufacturer?
If Tesla has 90% of the vehicles in the sample size, I would expect them to have 90% of the reported accidents.

As others have said, at this point, any accident is because the driver didn’t avoid it. The system still warns drivers to be aware and responsive.

I would like to see more, and better data from all parties.

 

Was there ever an explanation for this?

NHTSA Finds Teslas Deactivated Autopilot Seconds Before Crashes