[UPDATED] 2 die in Tesla crash - NHTSA reports driver seat occupied

[ev_go123]

In my opinion, whether or not AP can be activated on a street without lane markings is not as interesting as the other cirumstances. That is, I find it hard to believe you can accelerate quickly enough within 300 feet to reach an excessive speed, engage AP, and also move to the back seat. Those circumstances alone make it very unlikely AP was engaged with the driver in the back seat. You'd have about 5 seconds to mash the throttle, gain speed, engage AP, and get from the driver's seat to the back seat. Seems like the most unlikely of all the scenarios given that very short distance they had to accelerate before the turn.

Mike

I definitely agree it seems highly unlikely AP could have gotten to a high enough speed in 300-400 ft to cause that crash.

That being said, the driver might have been able to set up AP from a near standstill, switch seats, then started the car. This video demonstrates that:

But AP would then have had to accelerate very very hard on its own to end up at those speeds in 400ft or so.

 

[Tam]

If they tried to turn on AP and the car wouldn’t accept the command to do so, would that also show up in the logs?

I think so. It would log many activities and whether the responses to those activities were successful or not. In real life, a driver can hear the successful or unsuccessful tones/chimes.

 

[SMAlset]

Is there anything out there confirming the Dr. was the owner of the car and how long he had been driving it?

 

[SMAlset]

I think so. It would log many activities and whether the responses to those activities were successful or not. In real life, a driver can hear the successful or unsuccessful tones/chimes.

that should settle then if the car recognized the curb or street paving as an area AP could be turned on then. Doesn’t matter about other stretches of road, only matters where the accident happened.

 

[Needsdecaf]

Do you have access to that place? Probably a gated community? Can you please do that? It will clear up a lot of confusion.

It is a gated community. I do not have ready access. However, the streets in my community are identical in construction.

 

[Electroman]

It is a gated community. I do not have ready access. However, the streets in my community are identical in construction.

Any testing other than on that street, starting from that same location is of no value and does not prove or disprove anything. If it proved that AP can be engaged, it again means nothing unless the 5 others things I mentioned earlier can be proved.

 

[Dan D.]

I definitely agree it seems highly unlikely AP could have gotten to a high enough speed in 300-400 ft to cause that crash.

That being said, the driver might have been able to set up AP from a near standstill, switch seats, then started the car. This video demonstrates that:

But AP would then have had to accelerate very very hard on its own to end up at those speeds in 400ft or so.

Yeesh, I never saw that video before. AP working without anyone in the car at all, someone controlling the speed wheel from outside.
It needed lines to first enable AP, and the seat belt to be clicked, but then stayed in AP even after getting out of the seat and climbing out of the car. So much for the seat sensor (though if there was a seat sensor the video guy would just have put a weight on the seat and disabled that too). Obviously the video shows multiple ways are needed to defeat the safety mechanisms, but it is possible to drive AP while not being in the car at all.

Does not prove anything definite about the Spring, Texas crash. The video may be model, software version, and feature-specific.

 

[ev_go123]

One has to prove that:
1. AP can be enabled in that street starting from that same exact point where they started driving
2. AP can be enabled without a weight on the driver seat (100% not possible)
3. AP can be enabled without seat belt being buckled (100% not possible)
5. AP had wrong speed data for that road
6. AP will can race to 50 mph or above on that street

All of these has to be proved. Even of one of them is cannot be proved, then it completely absolves AP

1. Might be pretty hard to prove either way, behavior might be different car to car. day to day, shadows, sensor dirt, etc.
2. The video I linked seems to show AP doesn't care about seat weight sensor.
3. Buckling seat-belt behind you isn't too hard.
4. You skipped 4 haha. I'll add my own. AP is disabled if you open the door, so he would've had to climb over the seats to get into the back seat to keep it enabled, assuming he had set it up with a set point of 0 mph. If we agree he could climb over the seat in this AP senario, couldn't we believe he climbed over the seat after the crash in the hold-my-beer,-watch-this senario?
5. Possible, probably pretty easy to check though.
6. This is the one that gets me. Accelerating to 50+ in 400ft seems too fast for AP.

 

[Needsdecaf]

I definitely agree it seems highly unlikely AP could have gotten to a high enough speed in 300-400 ft to cause that crash.

That being said, the driver might have been able to set up AP from a near standstill, switch seats, then started the car. This video demonstrates that:

But AP would then have had to accelerate very very hard on its own to end up at those speeds in 400ft or so.

So a 59 year old doctor and a 69 year old engineer, both professional men and of some means judging by where they live, on a Saturday night after dinner with their wives, drop their wives off at home in their neighborhood, and then go through all the trouble to try to trick AutoPilot on a residential street where it has a close to zero chance of engaging?

This is a theory people are really getting behind?

 

[EVNow]

Buckling seat-belt behind you isn't too hard.

If the car was hacked in anyway to enable AP - then this crash is not because of AP.

The media seems to think / insinuate that AP has zero checks. Extensive hacking to enable AP disproves that assertion.

 

[Dan D.]

1. Might be pretty hard to prove either way, behavior might be different car to car. day to day, shadows, sensor dirt, etc.
2. The video I linked seems to show AP doesn't care about seat weight sensor.
3. Buckling seat-belt behind you isn't too hard.
4. You skipped 4 haha. I'll add my own. AP is disabled if you open the door, so he would've had to climb over the seats to get into the back seat to keep it enabled, assuming he had set it up with a set point of 0 mph. If we agree he could climb over the seat in this AP senario, couldn't we believe he climbed over the seat after the crash in the hold-my-beer,-watch-this senario?
5. Possible, probably pretty easy to check though.
6. This is the one that gets me. Accelerating to 50+ in 400ft seems too fast for AP.

6. We don't know definitively how fast, just that it hopped the curb at the curve and crashed. Damage to front end looks severe and the responders did say high-speed impact. How fast does it get up to speed when you increase the set speed (passenger could have done it)? Can it get to a high enough speed to cause the crash damage - yes seems doubtful. Should be easy enough for someone with AP to test though. How fast does AP accelerate from stop to high speed in 400-500ft? What if you go around the cul-de-sac a couple of times first?

 

[Iain]

Lots of speculation going on. Bottom line is we don’t know but we can all agree it sounds fishy as hell. Hold my beer while I click the seat belt, climb out the window get a brick on the accelerator and pretend I have FSD beta is about as accurate as the articles I’ve read from supposed professional journalists. My experience with FSD tells me it had NADA to do with the deaths of these two, but let’s see if they ever produce a real report .

 

[mongo]

There are already youtube videos showing engaging AP with no markings. So, what does a video showing you can't engage AP on a road without markings prove ?

View attachment 655486

What are you trying to indicate with that screen capture?
They (driver in the video) did not engage AP without lane markings previous to your screenshot.
AP was unavailable until there were lines, at which point they activated AP and it stayed on until the frame you posted.
This starts just before that section:

No AP option due to no line before this section:

Grey Wheel:

Blue Wheel:

Still blue wheel, AP stayed on.

 

[Dan D.]

So a 59 year old doctor and a 69 year old engineer, both professional men and of some means judging by where they live, on a Saturday night after dinner with their wives, drop their wives off at home in their neighborhood, and then go through all the trouble to try to trick AutoPilot on a residential street where it has a close to zero chance of engaging?

This is a theory people are really getting behind?

I'm gonna say yes. Argument. Bet. Dare. Boredom. Bragging. Foolishness. Curiosity.
Theory A being they hadn't enabled AP and crashed after an unfortunate launch, then crawling to the back seat.
Theory B being this one. Still I think it has a chance of happening that way.

 

[qdeathstar]

Autopilot CANNOT be enabled on a street like this one. You're welcome to drive down here and test it out.

Should I go make a video to counter that other idiot in the video?

Yes, yes you should.

 

[Tam]

...So much for the seat sensor (though if there was a seat sensor the video guy would just have put a weight on the seat and disabled that too)....

Yes. There's a weight sensor but people have found out a long time ago since about 2015 till now that it's for seatbelt alert and airbag function but not as safety measures against the driverless procedure. That's for all AP1, AP2, and now AP3.0. That's not news.

What news are people now say that cannot be done!

Since 2015, if you follow the youtube procedure, the car can continue to drive without a driver in the driver's seat.

 

[EVNow]

What are you trying to indicate with that screen capture?

I had watched this video a long time back. Just clicked in the middle to capture the screenshot. I forget whether in the video he could engage AP when there were no markings ...

 

[EVNow]

My experience with FSD tells me it had NADA to do with the deaths of these two, but let’s see if they ever produce a real report .

NHTSA will produce a real report. Just too late to make any difference in the public mind.

 

[Silicon Desert]

The media seems to think / insinuate that AP has zero checks.

yes and while on that subject, I noted that one financial network I was watching yesterday and today are not attempting to correct any misguided information they gave. Yesterday they were all over this story and blaming AP all day long. Today, not a single mention at all. They are never going to say otherwise.

 

[Iain]

6. We don't know definitively how fast, just that it hopped the curb at the curve and crashed. Damage to front end looks severe and the responders did say high-speed impact. How fast does it get up to speed when you increase the set speed (passenger could have done it)? Can it get to a high enough speed to cause the crash damage - yes seems doubtful. Should be easy enough for someone with AP to test though. How fast does AP accelerate from stop to high speed in 400-500ft?

Mine is way too slow. I normally use the accelerator so it can keep up with traffic ( i own a a DM3 and not a PS, hopefully someone who own the same car can answer)