Model X hits semi.

[stopcrazypp]

I guess my biggest question is why didn't the car AEB? Not every accident is preventable, but I would hope the computer would detect the imminent collison and apply the brakes faster than I could in a situation like this.

Emergency braking systems don't initiate braking unless there is no other option present (e.g. in order to not sabotage a driver's ability to swerve around an accident).

Plus, witness testimony tends to be extremely unreliable in terms of timelines and whether or not they felt the car brake during the moments leading up to a severe accident like this.

It would be nice to have some dashcam video from this. It looks like a high speed collision, it really depends on exactly how and when the truck cut him off whether or not this is a freak accident or a distracted driver scenario.

@chillaban pretty much explained why. There are confidence levels related to false positives and Tesla's system errs on avoiding false positives.

Fact of the matter is no auto braking system guarantees that it will prevent all accidents (even preventable ones), so even if it doesn't activate, the automaker is not liable. However, if the system caused an accident (either by interfering with the driver's ability to swerve, causing a spinout, or causing car to be rear ended by unnecessary hard braking), then the automaker is potentially liable.

 

[larmor]

From my experience AP1 has a sensory blind spot, using a clock face, from b/w 1-2 O'clock to about 4 O'clock and simlarly on the other side 10-11 O'clock from 8 O'clock. While AP1 can't be upgraded, it would be an easy retrofit to put a camera, and radar in the mirrors with wireless connectivity.
An aside, not sure why those mirrors are so huge and could certainly benefit from some slimming...

 

[earlyretirement]

I think Teslas are far from the safest cars out there. I really hope they eventually add some basic things that are already in many vehicles.

The Cadillac Escalades are really fabulous with the 360 degree view. Makes parking in any parking spot super simple with the 360 degree camera which the Model S and Model X would benefit from. Also, the side detection where any car coming from either side gives you a warning signal. And not to mention the most basic of safety functions like blind spot detection with the light underneath the mirrors. Many, many cars have those these days. Tesla should implement these safety features ASAP.

 

[voltaren]

I think Teslas are far from the safest cars out there. I really hope they eventually add some basic things that are already in many vehicles.

Thank you for your opinion; the NHTSA seems to disagree. I think I'll side with the crash tests and the countless stories of people surviving severe accidents.

 

[Az_Rael]

Thank you for your opinion; the NHTSA seems to disagree. I think I'll side with the crash tests and the countless stories of people surviving severe accidents.

Depends on which crash test results you use, I suppose.
Two electric cars miss IIHS awards

 

[Krypto Kat]

like lots of traffic with slow speeds) or no one around at all

Me too, those are the only situations I use autopilot.

As far a safety if there was a front passenger they likely would not have walked away from that crash.

I am also in disagreement with claims that Tesla is safer than other cars that have high crash test scores.

 

[apacheguy]

Emergency braking systems don't initiate braking unless there is no other option present (e.g. in order to not sabotage a driver's ability to swerve around an accident).

Plus, witness testimony tends to be extremely unreliable in terms of timelines and whether or not they felt the car brake during the moments leading up to a severe accident like this.

It would be nice to have some dashcam video from this. It looks like a high speed collision, it really depends on exactly how and when the truck cut him off whether or not this is a freak accident or a distracted driver scenario.

How on Earth can a computer determine if there are no other options available? I mean, based on your description it sounds like AEB will never activate because there will always be a possibility (however small) of another outcome.

 

[xkwizit]

Just couple of weeks back in SF on 101N, I had a similar situation. A car two lanes left of me suddenly jumped a lane. The BMW in that lane jumped lane into my lane to avoid impact. The first car jumped another lane so now suddenly I had two additional cars in front of me (the BMW and the other errant driver) traveling 60mph and into sharply slowing traffic ahead.

My AP1 X appropriately detected first the BMW and started to slow down and then the other car that was sharply braking and the X braked hard and avoided the impact. While I was ready to take over and attentive to what was happening ahead and behind me, I let the car take appropriate actions and I was pleased with the results.

But yes, I know AP1 is not anywhere L3 or L4 autonomy so if needed I am always completely attentive and ready to avoid a collision. What I still worry many times are the other vehicles rear ending me.

 

[EarlyAdopter]

If you want to blame anything here, blame weak underride guards on semi trailers. Rear ending a semi shouldn't be an immediate death sentence (or near-death sentence only because there was no passenger as in this case).

At least some progress is being made with newer semi trailers:
New award spotlights underride guards

 

[Canuck]

I am also in disagreement with claims that Tesla is safer than other cars that have high crash test scores.

You'll never convince me of that until you can you show me a crumple zone equal to the size of my classic S's frunk (before they halved its size), plus a sheet of solid batteries encased in aluminum at me feet. It took years until there was the first fatality in a Tesla, and it was to someone who stole one and took it for a crazy joy ride, splitting it in half. Then AP came out and it could be argued there's been an increase in fatalities but that's not really fair since there's also a lot more vehicles out there now so more fatalities are inevitable.

 

[Krypto Kat]

plus a sheet of solid batteries encased in aluminum at me feet

You mean under your feet.... I can see how this should help if you hit something that is planted in the ground like a tree, but doesn't a big engine block provide good energy absorption too?

more vehicles out there now so more fatalities are inevitable

Ahh..... there vehicles that sell at 100 times the volume and have no fatalities, of course that may have to do with the type of driver that gravitates to those vehicles too.

 

[vandacca]

You mean under your feet.... I can see how this should help if you hit something that is planted in the ground like a tree, but doesn't a big engine block provide good energy absorption too?

Ahh..... there vehicles that sell at 100 times the volume and have no fatalities, of course that may have to do with the type of driver that gravitates to those vehicles too.

Nope, engine block will most likely get pushed into the cabin. It's not nearly as well-integrated and connected to the main structure as the batteries.

Also, I very much doubt there are any vehicles out there (selling 100 times a Tesla) without fatalities.

 

[Az_Rael]

Nope, engine block will most likely get pushed into the cabin. It's not nearly as well-integrated and connected to the main structure as the batteries.

Also, I very much doubt there are any vehicles out there (selling 100 times a Tesla) without fatalities.

Study: 9 car models had zero traffic deaths

Toyota Highlander is on the list. There are waaaay more of those out there than Teslas

 

[vandacca]

Study: 9 car models had zero traffic deaths

Toyota Highlander is on the list. There are waaaay more of those out there than Teslas

Yeah, the study only focused on model years between 2009 - 2012, but sounds like they focused primarily on 2011 model vehicles. Vaguely (poorly) written article, since it doesn't even mention their sample space. Was it world-wide statistics or was it restricted to only the US or even to a particular state?

The article title should have been "Nine 2011 car models with zero US traffic deaths".

 

[Az_Rael]

Yeah, the study only focused on model years between 2009 - 2012, but sounds like they focused primarily on 2011 model vehicles. Vaguely (poorly) written article, since it doesn't even mention their sample space. Was it world-wide statistics or was it restricted to only the US or even to a particular state?

The article title should have been "Nine 2011 car models with zero US traffic deaths".

Here is the source for the data: Driver death rates

They only compile the information several years after the fact, so it will be a while before we see updated info that will include Tesla.

 

[bak_phy]

The Tesla number would be something like 7-10.
There have been around 5 deaths (I think) and there have been a bit under a million (700Kish) vehicle-years.

 

[voltaren]

Here is the source for the data: Driver death rates

They only compile the information several years after the fact, so it will be a while before we see updated info that will include Tesla.

Sourcing is everything and that source uses flawed and incomplete information.

IIHS gets their data from the NHTSA. While I trust NHTSA crash tests, their database of accidents is incomplete. When I say incomplete, I mean they only get an incredibly small fraction of details on accidents. Their intent is to study accidents and trends - NOT to create an all-inclusive database.

Case in point, none of my accidents are in the database. 3 coworkers of mine were killed in an accident in 2008 - not in there. See for yourself: FARS Encyclopedia

It relies upon accurate reporting. Many law enforcement departments don't report because they use(d) paper systems incompatible with databasing. Additionally, vague descriptors such as "2010 Toyota sedan" gives no indication of the model. Times are changing with the ongoing shift to computers (slow process in law enforcement).

And SearchForm

 

[earlyretirement]

Thank you for your opinion; the NHTSA seems to disagree. I think I'll side with the crash tests and the countless stories of people surviving severe accidents.

Well, I'm not just talking about surviving a crash and living. I'm talking about general safety. I could make the argument that many Teslas are involved in many accidents because of a lack of basic safety features that are in other cars. I'm not just talking about surviving a horrific accident because the vast majority of people are never in a horrific accident.

However, just go to any Tesla authorized body shop and you will see TONS and TONS of cars in the shop and an argument can be made that probably many less would be there with the features that some of us are talking about.

 

[bushburner]

Well, I'm not just talking about surviving a crash and living. I'm talking about general safety. I could make the argument that many Teslas are involved in many accidents because of a lack of basic safety features that are in other cars. I'm not just talking about surviving a horrific accident because the vast majority of people are never in a horrific accident.

However, just go to any Tesla authorized body shop and you will see TONS and TONS of cars in the shop and an argument can be made that probably many less would be there with the features that some of us are talking about.

or is it just that Tesla owners are crappy drivers?

 

[chillaban]

or is it just that Tesla owners are crappy drivers?

Quite frankly if every accident by a car had been scrutinized as much as Tesla's, then you'll find that crappy drivers exist with every car model…