Sudden Unintended Acceleration

[Chaserr]

When someone wants something, but denies the thing when offered, they don't want what they say they want.

Have you ever seen a homeless person asking for money to eat? Offer to buy him all the food he wants. If he's hungry he'll be thankful, if he lied he will tell you no to the thing he was only pretending to want. I've experienced both kinds of people, helping those in need feels good and dishonest people refuse help.

 

[CalBlue 85D]

Are you sure you weren't driving an Audi? Call 60 Minutes quick! Time for a follow up news report 30 years later.....

 

[Hota]

+1 on the creep idea.

Creep should maybe default as "on" when cars are first delivered and stays on for first 500 or so, then becomes an option to slide off.

Double-down: maybe chill mode should also default on... same idea becomes available option to turn off after 500. Or maybe the initial delivered chill "thaws" slow and automatically over time... eventually uncorking what the car is meant to be, so when you reach 500 on the odo, you're good to go.

Tesla can "save the customer's face" by claiming it to be a "manufacturer's break-in period" instead of a consumer protection / nanny thing. And a little pop-up reminder if you tromp on the go pedal before 500 is reached, to explain why your car isn't going fast yet.

The car that I drove for the past 12 years before my Model S (M3 with sequential manual gearbox - which I still have for fun) doesn't creep. I don't want creep. I don't want to replicate a slushbox torque converter.

 

[Hota]

It can, albeit unintentionally. It depends on how Tesla has it set up. To accept Tesla's logs as truth would require 100% trust and faith that Tesla coded things properly. Hypothetically... What if one software or hardware "module" triggered the acceleration by bug or otherwise, causing the pedal to depress and a physical pedal sensor reading the pedal going down. The logging software reads this pedal press data, sees that all driver assistance features are off, and by process of elimination determines it was the driver who did it. But it wasn't.

I'm not saying this is what happened, but you certainly have to allow for the possibility. HAL can very easily open the pod bay doors, but show in the logs that they were never opened. There's no way to know for sure unless you do a deeper exploration of Tesla's process, which would require legal discovery.

My advice to the OP: It was probably user error, which is unfortunate, but if you are convinced it wasn't you then I would probably contact an attorney.

For sure. This is like Elon stating that AP data shows that it has 90% less lane excursions than human drivers based on it's own data. Of course that's based on AP itself determining where the lanes are and if it can keep the car within those hypothetical lanes it drew for itself.

 

[brkaus]

The problem with the quote in the manual is that the non-warranty repair does not require any data. It’s bodywork.

If your looking for impartial, wk057 is your person. Look up his history. He has dissected these things inside and out, sometimes to the dismay of Tesla.

 

[Ying]

I interpret this as: If you don't intend to claim this accident under the warranty then it is under "non-warranty repair service" (your insurance pays, not Tesla pays) so you as an owner should be qualified for the car log.

The problem with the quote in the manual is that the non-warranty repair does not require any data. It’s bodywork.

If your looking for impartial, wk057 is your person. Look up his history. He has dissected these things inside and out, sometimes to the dismay of Tesla.

I am definitely going to do so. Thank you for the tip.

 

[Hota]

I appreciate the offer and I will certainly do so after Tesla gives me their log. SO there is a point to map it.

The electricity footprint is a result of design failure or defect. I am not saying there is one, but we have to review before we make any conclusion. At this point, Tesla rejected to look at my car and provide the log.

I don't think you understand. You don't need Tesla to pull the log. wk057 has offered to do that for you.

 

[Shock-On-T]

1/18/2018, morning 9:05 AM.

I was pulling in to my office parking spot where I parked for 6 years, I let the car slow to roll closer in front of the curb. All of a sudden the car accelerated, got on the curb, hit the office building. The car was still going until I applied the brake. The acceleration point was right before the curb, measuring from the point to the wall is only 7 feet. Tesla said I was on the pedal for a second long, I applied pedal from 0-18%, and quickly applied the brake. If I applied the pedal, I would not be able to apply my brake in a split second. Tesla refused to provide the force given to the pedal and any telegraphy. Tesla said they will not be responsible for any damage and said it is driver's error. I told my car was only 6 week new, I no longer feel safe to drive this vehicle, I request to return the car. Tesla refused. Tesla claimed there is no parts failure and refused to do any failure analysis without even looking at the car and accident pictures.

Tow truck driver told me I was the second brand new Model S he towed the same week. The other lady had the same problem SUA and the car ran into her laundry room, broke the drywall.

What can I do at this point?

It's most likely that you pushed the accerator pedal thinking it was the brake.

These incidents almost always happen when the driver is doing a manoeuvre that requires braking.

It's worse in Teslas, because the driver doesn't get any cues (revving engine, laggy increase in power) to warn them they've hit the wrong pedal. It's just an instant, silent lurch forward into the wall.

 

[SucreTease]

I would argue that it should be a configurable setting to ignore or greatly reduce high acceleration from a standstill (or nearly so) when sensors (radar and/or ultrasonic) detect a wall or large object. For a car as advanced as the Tesla is, this should really be available to drivers to protect people from their own mistakes (though I will not argue that it should be the default).

 

[Shock-On-T]

I would argue that it should be a configurable setting to ignore or greatly reduce high acceleration when sensors (radar and/or ultrasonic) detect a wall or large object. For a car as advanced as the Tesla is, this should really be available to drivers to protect people from their own mistakes (though I will not argue that it should be the default).

Problem is that that would make normal driving intolerable.
The car would be constantly baulking at objects you were planning to avoid.

 

[SucreTease]

Problem is that that would make normal driving intolerable.
The car would be constantly baulking at objects you were planning to avoid.

I modified my post to read "from a standstill (or nearly so)". One could put enough circumstances on it to avoid any but the intended constraint. Can you think of a scenario where one would want rapid acceleration from a standstill (or nearly so) when sensors detect a large object less than 20 feet away and directly in front of the vehicle?

 

[Tam]

...as advanced as the Tesla is...

People might overestimate current Tesla technology and ignore the fact that there's a 2016 fatal Autopilot accident in Florida.

People might be so impressed with all the radar, sonars and up to 8 external cameras and ignoring the fact that it can still crash into garage with summoning as well.

There is no doubt that some day Tesla will do what we now overestimate of what it can do today but that possibility is in the future and we still have to deal with the reality of today.

 

[Mark Z]

Vehicle manufacturers are demonstrating guardian technology that prevents driver error by checking the camera and sensors in advance of driver action. Preventing acceleration from a parallel parking spot when a bicycler approaches or non-acceleration at a signal when a camera detects a car entering the intersection from the side are examples. The guardian technology should be active at all times. In a few years, we may be able to purchase vehicles that will never crash during parking, even with driver error.

 

[Tiger]

I appreciate the offer and I will certainly do so after Tesla gives me their log. SO there is a point to map it.

The electricity footprint is a result of design failure or defect. I am not saying there is one, but we have to review before we make any conclusion. At this point, Tesla rejected to look at my car and provide the log.

Maybe I misunderstood, but to my understanding, @wk057 can pull the logs for you without Tesla's help. Isn't that true?

 

[cwerdna]

I think at this point Tesla needs to put a camera in the footwell of the car. People always claim they did nothing wrong, Tesla can show that they did from the pedal sensors, but the driver isn't going to believe that, they'll just claim the sensors were wrong (despite there being 2 that read in opposite directions). Seeing footage that shows that they stomped on the wrong pedal might convince them.

The funny part is during the whole Toyota SUA fiasco/PR disaster, there was one guy on Priuschat who kept suggesting Toyota needed to install footwell cameras.

BTW, back to the whole brakes are powerful enough to overcome acceleration and pressing on 2 pedals, I've pointed to this a number of times: Smart pedals won’t put the brakes on driver error.

Excerpt:

In the cases that went to court, jurors naturally asked, why would a driver with decades of driving experience suddenly mistake the accelerator for the brake? And why would the episode last so long — often 6 to 10 seconds or more? Wouldn’t that be ample time to shut off the ignition, shift to neutral or engage the parking brake?

First, in these situations, the driver does not really confuse the accelerator and the brake. Rather, the limbs do not do exactly what the brain tells them to. Noisy neuromuscular processes intervene to make the action slightly different from the one intended. The driver intends to press the brake, but once in a while these neuromuscular processes cause the foot to deviate from the intended trajectory — just as a basketball player who makes 90 percent of his free throws sometimes misses the hoop. This effect would be enhanced by the driver being slightly misaligned in the seat when he first gets in the car.

The answer to the second question is that, when a car accelerates unexpectedly, the driver often panics, and just presses the brake harder and harder. Drivers typically do not shut off the ignition, shift to neutral or apply the parking brake.

This is why i use creep. I want to be feathering the brake, not gas, when I'm parking.

Agreed 100%.

It's probably related to why some people choose to turn on creep and some don't. If your driving experience is with auto transmissions, the lack of creep might lead to a confusion of where your right foot is at right now.

Yep. Creep also provides directional cueing. See Federal Register, Volume 68 Issue 94 (Thursday, May 15, 2003)

B. Safety Importance of Creep Force in Cueing the Driver--Creep Force
in Reverse
...
When the current wording of FMVSS No. 102 was adopted in 1968,
vehicles were equipped mostly with ICEs and mechanical/hydraulic
automatic transmissions, which have always provided creep force. Today,
drivers can rely on creep force to avoid crashes that would result from
shifting errors. When a driver places the automatic transmission shift
lever in a drive position and reduces service brake pressure slowly by
easing up on the service brake pedal, the vehicle begins to move slowly
in the direction that has been selected by the transmission shift
lever. This creep force in the correct direction cues the driver that
when the accelerator is depressed, the vehicle will move in the
anticipated direction.
It is important for creep force to initiate motion of the vehicle
before the driver's foot leaves the service brake pedal and before the
service brakes are completely disengaged. Then, if a shifting error has
occurred, the driver's foot is still on the brake pedal and the error
can be safely and quickly corrected. For example, if there is no creep
force associated with an automatic transmission equipped vehicle and
the driver thought he had selected Drive but instead had selected
Reverse, when he removes his foot from the brake and depresses the
accelerator, the vehicle would unexpectedly move rearward instead of
forward. The unexpected movement of the vehicle rearward may cause the
driver to further depress the accelerator. By the time the driver
realizes his mistake and applies the brake again, the vehicle may have
moved rearward a considerable distance and possibly struck a pedestrian
or an object, causing injury and/or property damage.
From years of driving ICE-powered automatic transmission vehicles,
drivers are familiar with cues in the direction of travel indicated by
creep force. Since it is not inherent in hybrid vehicles, it is NHTSA's
view that there is a safety need to at least design creep force into
the vehicles when in Reverse.

 

[schonelucht]

Problem is that that would make normal driving intolerable.
The car would be constantly baulking at objects you were planning to avoid.

Enhanced Autopilot, once it implements the powerful summon should be able to feed precise data about which objects are avoidable and which are not. The wall in front in OP's case is certainly unavoidable and the car should simply refuse to run into it under any circumstances.

 

[mattack4000]

Is this James sikes???

 

[AnxietyRanger]

You have pointed out a blatant fact- there are multiple cases hanging there not only during parking, it happened when making turns as well. Whenever you car slows down, the SUA might happen.

Making a turn may be different - it has been speculated it is about confusing the cruise control stalk with the blinker stalk. Especially when turning right, one's hand could move on the steering wheel in a manner that pushes the cruise stalk upwards instead of the blinker stalk, thus enabling cruise at the current speed and surprising the turning driver with a steady speed coming out the corner, instead of a steadily slowing car... This, here, is something I think Tesla would be wise to change so that this wouldn't happen so easily.

But given that a stalk press is unlikely in your case (and your speed insufficient for it)? I have to side with @bonnie and @wk057 on the likely reason in your case being pedal confusion - and join the generic recommendation on this thread to use Creep mode. Not using it can be confusing to some, including me who still uses it three years into Tesla ownership. I would recommend Tesla making Creep mode the standard mode for new users (I think it isn't? Or is it?). The silent and immediate torque response in a Tesla makes using accelerator precarious in tight spaces anyway, where an ICE would be more easily reigned in before anything happens...

I do wish Tesla would have promptly, as a standard matter, furnished you with the logs, though. Their insistence not to is one thing that causes the conspiracy theories to swirl in these cases. Again a bit of transparency could have stopped things from escalating in the first place...

 

[AnxietyRanger]

Thank you for the informative message @wk057.

It's also always pedal misapplication, too, where the driver presses the accelerator when they should be braking.

I have to disagree with your certainty on that, though. The cruise control stalk misapplication (confusion with blinker stalk) would seem to be the likely reason in at least one or two cases posted on TMC, where a car has exited a turn at speed and the driver swore no pedal was pressed. They reported it as a SUA. So pedal and cruise control stalk misapplication would seem to be possible reasons IMO, when speaking generally of the SUA report phenomenon regarding Teslas (and perhaps Mercedes). Both user error of course, but the latter also a usability design decision that could be further improved upon through software.

Anyone can try it themselves. Approach a right-hand corner, taking off feet from pedals and the car starts to slow down in preparation for the turn. Then put on the "blinker", but instead of flicking the blinker stalk up, flick the cruise up (thinking it is the blinker) and complete the turn without any pedals pressed... It will feel like SUA coming out the corner, if you don't expect it, as the car will no longer slow down but maintain speed - steadniness of which is momentarily masked by your attention in making a turn. The sudden lack of expected deceleration (no pedals pressed "engine" - regen - braking into the turn) feeling like acceleration.

The later you "blink", the worse it is - those who put on the blinker as they turn the wheel, bad driving manners of course, being arguably the most vulnerable. One, because they may be focused on turning the wheel and flicking up the blinker at the same time as the wheel turns (cruise stalk being the first in their line of hand movement in that case, on right-hand turns). And two, because the stop in deceleration comes just in-turn or even out the corner, where it may be the hardest to notice and instinctively push brake. (This I believe why someone hit a parked car on TMC.)

The latter phenomenon is limited in scope by the cruise enabling speed-limit, so one would have to "blink" at sufficient speed, but it does not seem to be a non-existent one.

 

[jelloslug]

I modified my post to read "from a standstill (or nearly so)". One could put enough circumstances on it to avoid any but the intended constraint. Can you think of a scenario where one would want rapid acceleration from a standstill (or nearly so) when sensors detect a large object less than 20 feet away and directly in front of the vehicle?

To prevent an accident, to prevent a car jacking, or anytime you need to get out of the way in a hurry.