Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

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

This site may earn commission on affiliate links.
Here's a video of a Tesla owner Sergio Rodriguez activating AP on a nothing street with no lines. AP sets itself to 45mph which he says is too high. AP proceeds to accelerate and fails to react to the curve except he stops it. Whether this video has any bearing on the crash in Texas it is interesting allowable behavior in the car. It seems to do what we are told Tesla's on AP cannot do.

Let's say the car "backed from the driveway" and started driving as we were told in witness statements. One guy in the back, one guy in the front. Some form of disable devices in the seatbelt and/or seat weight. Whatever was required. Let's suppose that starting position is halfway up the driveway at #2 Hammock Dunes Pl for the sake of argument, where there is a reddish car on Google Maps. Not sure if that is the correct house, but let's start there.
They back from the parking spot and drive down the driveway, through the cul de sac and down the street. Total distance is 450-550 ft. Somehow they have enabled Autopilot - as Sergio Rodriguez was able to do in his video. The Tesla "should" not allow this, but does anyway - as Sergio's car did. Let's say it accelerates to 45mph either given a nudge from the passenger's foot or by itself. Perhaps they entered a higher speed or resumed a previous set speed. Again, the Tesla "should" not do that, but perhaps they discovered that it actually does - as Sergio Rodriquez did. Perhaps they get it up much higher than 45mph.

Next, either AP fails to make the turn - like Sergio's car, or they nudged the wheel by mistake, disabling AP and it doesn't even try to turn. If they "nudged it off" then AP would show as not engaged, as Elon said.

45mph is not very fast, but perhaps fast enough to crash into the tree injuring the occupants. Perhaps they were not belted. Perhaps running over the curb/drain and hitting the tree was enough to cause an instant fire. Perhaps the fire was so sudden that though just injured they were incapable of escaping. Perhaps the passenger's seat was wedged against a tree, the battery had failed and the rear passenger didn't know the escape procedure. Perhaps 45mph was fast enough. Lots of "perhaps" I know, but it fits the scenario.

Just reconciling the witness statements with the list of possible reasons they might have tested the car in this manner. They wanted to see the car navigate on AP on their little street (maybe the owner had previously found that it would do it, even though it's not supposed to be able to do it). It seems AP can activate on streets with no lines. It seems the car can accelerate to an unsuitable 45mph on a small street. It seems the car on AP can fail to turn whatsoever on those streets (thanks to Sergio Rodriguez's video).

Would not be good for Tesla if AP could/did actually do all these dangerous things that Tesla claims it would never do.
 
Last edited:
Red calipers, so it was a performance model, which very likely meant.....speeding. Heavy braking at the last minute. Car hit hard enough to catch on fire. Just need to know if AP was on and if they were really being idiots using it without supervision. If it turns out they were being reckless, I have no sympathy since they could have killed someone else.
 
Here's a video of a Tesla owner Sergio Rodriguez activating AP on a nothing street with no lines. AP sets itself to 45mph which he says is too high. AP proceeds to accelerate and fails to react to the curve except he stops it. Whether this video has any bearing on the crash in Texas it is interesting allowable behavior in the car. It seems to do what we are told Tesla's on AP cannot do.

I'm not sure his example is relevant:


Notice that Elon said standard AP requires lane lines. That person paid for FSD so it has different features/requirements.
 
I'm not sure his example is relevant:


Notice that Elon said basic AP requires lane lines. That person paid for FSD so it has different features/requirements.
Why do you think it's not relevant? Looks entirely relevant, same type of street, fails to turn, and would have crashed. A different car with different software maybe. Maybe the code base used is the same. It's a workable theory anyway.
 
Why do you think it's not relevant? Looks entirely relevant, same type of street, fails to turn, and would have crashed. Just a different car, probably the same software.

Because that person has paid for FSD while the person that crashed didn't. (Elon said standard AP requires lane lines. It seems that one of the features of paying for FSD is the ability to enable Autosteer in more places.)
 
  • Like
  • Helpful
Reactions: Mr X and scottf200
autosteer works the same on fsd and ap. I cannot activate auto steer without lane markings.

However, if you watch the video, auto steer is engaged at 11mph, the he increases the speed to 55 using the scroll wheel... it fails to turn because he turned off auto steer at the end.
Do you think that supports the type of crash seen in Spring, Texas? Whatever happened in that crash, their drive certainly did not go the way they intended. Can you theorize a set of conditions with AP, auto steer, etc that could lead to the crash?
 
Two things I find interesting about the new info. One is that the article that discusses how the fire didn't take 4 hours to put out said that the car was "wedged between two trees". That might explain the inability to get out and why a driver might try to get to the back to get out the back doors (which could have also been wedged). The second, on Sergio's video, I find interesting because when he first activates AP, there is a clearly visible "seam" in the center of the road that I could see how AP might interpret as a line.

Mike
 
  • Like
Reactions: Dan D.
Do you think that supports the type of crash seen in Spring, Texas? Whatever happened in that crash, their drive certainly did not go the way they intended. Can you theorize a set of conditions with AP, auto steer, etc that could lead to the crash?
No. Therefore I think it is unlikely that AP, auto steer, etc could lead to the crash. All of them require positive input from a driver to achieve excessive speed. According to Chief Doof there was no driver.
 
No. Therefore I think it is unlikely that AP, auto steer, etc could lead to the crash. All of them require positive input from a driver to achieve excessive speed.
But there was a driver/enabler, just not in the driver's seat. Can not the front passenger provide the positive input as required? If the crash happened with no driver - (if) - then they were doing something to enable their unfortunate demise.
 
Because that person has paid for FSD while the person that crashed didn't. (Elon said standard AP requires lane lines. It seems that one of the features of paying for FSD is the ability to enable Autosteer in more places.)
As someone else suggested, is it possible they were testing FSD in "try before you buy" mode? Elon said the "car had not purchased FSD", I assume he would know if they activated a trial FSD session, which could be the "Test" the two men were intending to perform.

But even if so, I think it would constitute an Autopilot activation, and Elon said AP was not engaged.
 
  • Disagree
Reactions: Needsdecaf
As someone else suggested, is it possible they were testing FSD in "try before you buy" mode? Elon said the "car had not purchased FSD", I assume he would know if they activated a trial FSD session, which could be the "Test" the two men were intending to perform.

But even if so, I think it would constitute an Autopilot activation, and Elon said was AP was not engaged.

the car was bought in 2019 so I doubt it. I also don’t think it makes much of a difference here.
 
the chief said the car was driverless.



Just about anyone can crash into a tree I suppose....
Yes, I get that, but "driverless" means nobody in the driver's seat. I could drive my current car from the passenger seat just it would be rather stupid to do that. My question was, can someone control AP, auto steer, set speeds, whatever, from the passenger seat (or by leaning over from the back seat), and create the conditions that crashed them. The front passenger can press the accelerator from the front passenger seat.

Again, that's all very stupid to do. But, if that's what happened (not saying that it did), then it was rather stupid.

I'm trying to determine if the car CAN crash in the manner the witnesses say, despite what Tesla says about the car. Creating other situations like 3rd driver, climbing into the back while driving, thrown into the back after crashing, even climbing into the back after crashing - all seem to go against witness statements. SOMETHING happened. I'm suggesting they created the situation through careful and unfortunate manipulation of behaviors in the Tesla's systems. Behaviors that other Tesla drivers have suggested do indeed exist.
 
Two things I find interesting about the new info. One is that the article that discusses how the fire didn't take 4 hours to put out said that the car was "wedged between two trees". That might explain the inability to get out and why a driver might try to get to the back to get out the back doors (which could have also been wedged). The second, on Sergio's video, I find interesting because when he first activates AP, there is a clearly visible "seam" in the center of the road that I could see how AP might interpret as a line.

Mike
They did say the fire kept igniting. 4 hours total. They also said ~30,000 gallons of water, so a lot of water, at 250gpm that's 120 minutes total. Agree that if the car was wedged it might explain the inability to self-extricate. As to enabling AP, there may be seams or other markings on the crash road that fooled AP. Again, it's just a theory that AP could somehow be activated. The wives of the deceased said the guys were talking about Autopilot, thus if that is true, they were likely trying to activate it. Perhaps they did.

Elon said that "Data logs recovered so far show Autopilot was not enabled". That is not the same as proof, he could change that statement later with further information. Also there is wiggle room over time. E.g. Autopilot was enabled for 50 seconds but was not enabled for the final 5 seconds. Autopilot could have been enabled, got them into the dire predicament, then been disabled 5 seconds before the crash. Also he did not say Autopilot was never enabled. They may also have been trying to enable Autopilot but failed, would logs show attempts to enable Autopilot?
 
Something else I thought of in relation to this snap:


What if those are not skid marks but acceleration marks? Someone not familiar with regen might be lulled into thinking their foot is on the brake if they are applying light pressure to a pedal and the car is slowing down: so they press the wrong pedal? The photo isn't HD enough to determine which way the grass is laying (forward for braking or back for accel) but in looking at the right track, it looks like more dirt is kicked up off the right side of the track indicating a sharp left turn - trying to stay on the road. The marks are also grooved a bit sharply. That could indicate that the front wheels were already turned significantly to the left when the car entered the grass and reduced traction caused the turned wheels to plow straight ahead while the wheels were turned left. At that point, the driver may have tried to avoid a direct collision with a tree and hence ended up wedged between two trees instead.

Whatever happened, it has signs of excessive speed and loss of control being factors. I don't think any scenario with AP would have lead to those marks/ruts. Those look like driver panic.

Mike
 
  • Like
Reactions: alexgr