I just did a quick scrape of the NHTSA consumer-reported complaints for all models of Teslas for the month of May. 78 reports. 2 concern unintended acceleration.
Mind you, 'unintended acceleration' reports come from all manufacturers and all models of cars -- so a fair starting position is that the driver mistook the accelerator for the brake. Nevertheless, it seems that giving the crash data promptly to the consumer would clarify to the consumer that they have 'fat feet', and settle the matter, at least in the consumer's mind.
Incidentally, I narrowly missed getting hit by a lady who drove her car (non-Tesla) into a local Fedex-Kinkos, having left the store 5 minutes earlier, and returned 5 minutes after her car 'parked' inside the store.
In my Tesla I've had the experience a couple times of trying to hit the break, but accidentally pressing both petals at once. When this happens, a notice appears on the dashboard saying both petals are being pressed and so the car has decided you intended to brake.
I don't think I've ever had this problem in another car (though admittedly, probably no other car I've driven would alert me if I did). It seems like the petals are placed a little differently from most cars and this makes it easier to accidentally hit both petals. I wonder if the computer's detection of this condition and decision to prioritize braking is a new feature since 2018...
The first time I wore heavy steel-toed boots while driving, I had a hard time "finding" the accelerator pedal and kept fat-footing them both. But it didn't take me long before I was completely used to it.
Do you think it is because you rarely need to use the brake in a Tesla? My limited experiencing driving one was that simply removing your foot from the accelerator often provides sufficient braking power. Maybe you lose some muscle memory.
I once overheard a conversation between two elderly ladies in a café. They were talking about how to drive a car without a clutch, and one was telling the other how difficult it was for her to drive with one foot on the brake and the other on the gas pedal.
I'm not sure driver mistaking the pedals is a fair starting point with the amount of software involved now a days. Telsas push updates over the air that affect these sorts of things. In this article[0] for example, it's suggested a once optional feature that is now mandatory can be used for one pedal driving.
Of the cases where 'pedal mistake' doesn't explain it, self-interested, logical game theory suggests that a large portion are people who KNOW they accelerated manually and are seeking to avoid liability. Of those even not deliberately seeking innocence, there is a portion still whose human memory has post-rationalized that the car was in fact responsible, because they would never..
One pedal driving is for braking by recharging your battery, I've never heard of it actually using the physical breaks. It wouldn't make sense since when there's no emergency, you want to minimize the use of physical breaks since the turn the mechanical energy into thermal energy, essentially wasting it.
Remember the issue of Toyota’s with unintended acceleration? That was 100% due to fat feet so this is indeed a very fair starting point. This has happened so so so many times without any technology involved. The tech may have changed but the human behind the wheel hasn’t.
The first major cause of unintended acceleration was found in March 2007, when an engineering analysis showed that unsecured all-weather mats had led to pedal entrapment and drivers accelerating up to 90 mph with decreased braking power...
...Early on, Toyota suggested that driver error was to blame, saying that some people may have hit the gas when they meant to hit the brake...
...This led to Toyota sending a letter to the owners of the affected car, a 2007 Lexus ES350, asking that they bring their cars in to switch out the all-weather mats...
...After this recall, Toyota decided to revise the internal design of their cars to ensure that there was "10 millimetres (0.39 in) between a fully depressed gas pedal and the floor," but decided to only implement the new designs upon the next "full model redesign", which wouldn't take place until 2010...
...In an attempt to hide these defects from investigators, Toyota switched to verbal communication on the defect rather than traceable forms of communication. As a result, many new cars were knowingly produced with the same floor mat issues that had been identified as being having the potential to cause SUA problems in association with the defective pedal design..."
That famous quote from Fight Club seems especially relevant here…
Reproduced below, for those who haven’t seen the film:
Narrator : [20:35] A new car built by my company leaves somewhere traveling at 60 mph. The rear differential locks up. The car crashes and burns with everyone trapped inside. Now, should we initiate a recall? Take the number of vehicles in the field, A, multiply by the probable rate of failure, B, multiply by the average out-of-court settlement, C. A times B times C equals X. If X is less than the cost of a recall, we don't do one.
Woman on Plane : Are there a lot of these kinds of accidents?
Narrator : You wouldn't believe.
Woman on Plane : Which car company do you work for?
> The first major cause of unintended acceleration was found in March 2007, when an engineering analysis showed that unsecured all-weather mats had led to pedal entrapment and drivers accelerating up to 90 mph with decreased braking power...
That doesn't doesn't really align with all the people who were 100% sure they were stomping on the brake and the car was accelerating.
I think a very large portion of their "guilt" was due to being a non-Detroit/American company at a time when those companies where hurting. Everyone just loved to pile on with no evidence at all.
The exact same thing played out with the Audi 5000 in the 80's -- people hit the wrong pedal and lie. Until this issue is fixed, I would continue to suspect that first.
> Toyota Motor Corporation Admits to Misleading Consumers and U.S. Regulator About Safety Issues Related to Unintended Acceleration in Its Cars ... TOYOTA to pay a $1.2 billion financial penalty – the largest penalty of its kind ever imposed on an automotive company
Probably not with the Toyota given the model, but this can often be both a human error and bad design.
In some cases, "human error" is at least partly attributable to poor human factors engineering. It sometimes happens with SUVs and large trucks with large transmission wells that effectively move the pedals towards the driver side door. When in a panic situation, people will sometimes mistake the location of accelerator for the brake out of sheer muscle memory.
The problem with that is that I wouldn't put it past Tesla to update that software and/or the data even after a crash if the car can be reached to avoid liability.
You'll probably find similar numbers of false-braking reports for most other luxury brands, because radar autobraking is dangerously shitty.
I turned mine off the first time it slammed on the brakes in traffic for no reason. Still get false !!!COLLISION WARNING!!! chimes about once every thousand miles or so.
This tech isn't ready for prime time, and it's about to become mandatory.
"As of September 2019, the Tesla Model X, Model S and Model 3 have had more than 108 sudden unintended acceleration reports, as per NHTSA's Early Warning Page...
...A petition was issued for further investigation from the NHTSA, claiming Tesla's vehicles may have a structural flaw that can endanger public safety, and that the number of sudden unintended acceleration reports filed for the Tesla models were as much as 6000% higher as other brands' cars on similar class or otherwise (pages 63 through 66 on the petition report)...
...Furthermore, it also alleges Tesla knows about the "unintended acceleration" and is unresponsive to driver engagement following the incidents. It also asks for the recall of all Model 3, S and X vehicles produced by Tesla since 2013.The NHTSA has issued a statement saying they are in the process of reviewing the petition issued..."
True, but it can be indicative of a bad safety/design culture when multiple instances start surfacing. That's a much bigger deal than explaining a one-off
This is a bit tangential, but this is far from just a Tesla problem. All cars since 2011, but many cars back to I think 2005 have Crash Data Recorders. I had to jump through hoops to get the data from the CDR in my Toyota that sound almost exactly like what's described here.
My insurance company refused to listen to me without it because the other driver was insured by the same company and they had NO incentive to correctly assess fault, never mind that they do have the tool to pull the data and the expensive proprietary software to read it.
I had to contact the manufacturer of my car, who sent me to a website with a list of independent people who own the $40,000 tool that can pull the CDR data. Fortunately one of them (and ONLY one of them) was within a two hour drive from me.
And then the insurance company refused to review it anyway until my state insurance commissioner got involved and told them they can't refuse to review the facts of the case just because they're on the hook for all the damage no matter who caused it.
And that's how they ended up on the list of companies I'll never do business with again.
Did you go to all that trouble to lower your premium?
Edit: I was on mobile and wanted to ask why they went to that trouble if the car insurance company was the same and it didn't matter who's fault it was. Didn't mean to imply it wasn't worth it.
I didn't want an at-fault accident on my record when I wasn't at fault. That'd screw me for years, even though all I did was save a guy's life when he pulled out immediately in front of me (and hit a post instead).
The problem was partially the fault of the police officer who wrote the report. She told me at the scene that she had never written a report for an accident where the at-fault vehicle had no damage before.
She noted that there were no skid marks at the scene, so concluded my speed may have been a factor, totally ignoring the fact that the vehicle has ABS and it had been raining lightly all day.
She also said I told her I knew I wasn't speeding, but I didn't know what the speed limit was. What I actually said was, "There are 3 speed limits on this stretch of road (I live on that road--I drive down it every day), 30, 35, and 40, and I don't know whether I was just before a change or just after one, but I'm certain I wasn't going more than 30."
Turns out the pole I hit was 50 feet after the sign that changed from 30 to 35, and the CDR agreed I hadn't been going more than 30.
And the driver of the car the at-fault driver was pulling around to get in front of me actually called the police station the next day to try to make absolutely sure they knew it was 100% the other driver's fault, because he had the perfect view on it and heard the guy say, out loud, "I can't see, so I'm just going to go."
And even after the insurance commissioner got involved, GEICO wouldn't go below holding me 10% responsible for the accident, because when you're the judge of your own accident, apparently facts don't have to matter at all...except who has the higher deductible, so who you can get more from.
It took more than 6 months to get through this garbage. I think they saw someone driving a 15yo car and assumed they had my back against a wall and I couldn't do anything about it.
I happen just to be a superfan of the 2nd gen Prius for its climate controlled car camping potential (and generally just being one of the most well-made cars in history)! Finding them is my challenge, not paying for them!
This is why I covered the internal cam on my model y. Basically, I can't see the video, only Tesla can, and the video would never help me in an accident, only hurt me.
It's interesting that all the accounts reported by Tesla indicate that ..."the accelerator pedal was abruptly increased to 100%.” How likely is it that, even if by accident, the driver would fully depress the accelerator pedal? How many people fully depress the brake pedal when stopping? Usually while parking, there is no need to fully depress the brake because sufficient braking action is achieved with less pressure, and the car will come to a stop gradually instead of abruptly.
Maybe they have a CAN bus problem, or a software problem that they do not want to reveal.
The brake pedal is physically connected to a hydraulic cylinder. You need to be very strong to fully engage the brake pedal, such that many vehicles, when presented with a sharp brake pedal stab, will electronically increase the amount of brake pressure given (because the system assumes you want to stop, but the driver may not be strong enough/have the right feeling for where maximum pedal application is). If you did this at speed, you'd probably engage ABS which tends to mess with the force-feedback though.
On the other hand, the accelerator pedal is just a sensor, it doesn't require as much force to fully depress. I'd estimate the force required to fully depress the accelerator pedal to be 1/3 of the brake pedal.
But if it's just a sensor, couldn't it be possible that sensor generates an erroneous signal such that it tells the car to accelerate when nobody actually pressed it?
Acceleration pedal sensors are ASIL D and by ISO 26262 should have a probability of failure per operational hour of <10^-8. So yes, but it should be incredibly unlikely for sensor failures to occur and there's a very thick document somewhere inside Tesla detailing all the considered failure modes and how they've addressed them.
I believe the pedal sensor has two totally redundant position sensors. If one fails or if they disagree, the car will go into limp mode and turn on the MIL. Maybe not all of them are like that, but I remember AvE (the YouTube channel) having to replace the pedal sensor on one of his cars a number of years ago and he talked about how they worked.
Yes and no. A lot of EVs can do two different modes, where it "coasts" and mostly only starts the regen through the brake pedal. You're right though, from what I understand Tesla kind of forces the single-pedal driving mode.
Eh, from what I understand Tesla has made a number of changes to the configurability to regen. I don't drive a Tesla so I'm not entirely sure what that means in terms of single pedal and the overall driving feel.
To most people, brake-by-wire means totally electronic control. Tesla's current models are only electronically assisted (in that there is an electrically powered hydraulic pump). There is still a physical connection between the brake pedal and the master cylinder. This means you are still available to provide some braking force if you loose 12v entirely.
I was worried about the idea of a fully electric braking system because of the failure mode you mentioned. It looks like hydraulic brakes are not ubiquitous.
"Some Tesla vehicles use hydraulic brakes, but not all. Recently, Tesla has equipped its cars with a full self-driving system. Tesla’s latest car comes with full-electric Brembo brakes. These brakes use electric actuators instead of hydraulic fluid pressure to actuate."[1]
I'd be curious what mitigations they put in place regarding an electrical system failure.
The mech4cars.com article is conflating the service brakes and the parking brake. The service brakes (the primary, foot-pedal actuated brakes) in a Tesla are hydraulic. The mech4cars.com article uses a link to [1] as evidence for the "electric brake system", but this is the _parking_ brake (which is indeed an electric motor, but is an entirely different system, separate caliper, etc). The parking brake will remain engaged even if you lose 12V, so you won't roll away if power cuts out (which you'll know if you've ever tried to tow a Tesla that went totally dead while parked). An electronic park brake is a setup seen in many modern cars.
Probably unrelated but if you have a manual gearbox, with a clutch pedal, when you are about to stop, you disengage the clutch to prevent the engine from stalling.
Clutch pedals are operated with the left foot and meant to be fully depressed. If you are used to drive stick and switch to auto, you may accidentally hit the break hard with your left foot, as if it was the clutch.
So if you see someone breaking hard at the red light for no reason, especially if it is a rental car just out of the airport, it is probably the reason.
I did it, Most of my (European) friends who rented a car in the US did it too.
I'd say it's quite likely that people would fully depress the accelerator when they meant to fully depress the brake. Especially with cars with cruise control where your foot might not already be on a pedal.
I don't think you need to reach for any more conspiratorial explanations when there's a blindingly obvious one available.
It would be interesting to see the data just before - were any pedals pressed?
Fully depressing the brake pedal would involve some fairly serious reason - you only ever do it in an emergency situation as it produces a very sudden stop. It's therefore highly unlikely then that someone would accidentally fully depress the accelerator because they mistook it for the brake pedal.
FWIW, I've done this once in my Tesla. Fortunately nothing was immediately in front of me at the time, and I was able to correct quick enough. Here's how the logic goes:
Or maybe someone is intending to put a boot full of force onto the brake pedal, hits the gas pedal with that same amount of force, which mats it to 100%?
If they don't have anything to hide, why not allow access to the logs? I highly doubt the 0-100 claim. Raw sensor data of the accelerator is likely to be sampled at a high enough rate that should make something like that impossible. In fact, the car should have at minimum a low pass filter and reject sudden accelerations like that, which would indicate sensor or signal failure. If it's a digital sensor, and sampled high enough, even if someone did "floor it", you'd still be able to see a gradual increase, just at a faster rate. When a sensor all of a sudden reports a huge, non-linear jump, it generally means there is some sort of failure. And if that's within normal operating conditions, the sample rate of the sensor isn't enough to tell the difference between failures and intentional (real) readings, and reject a failed sensor. And that's a problem too and 101 in any sensors and signals implementation. If it's a digital sensor on something as important as braking and acceleration, there had better be redundant sensors, a mechanism to fuse the data together, and reject a bad sensor from the state estimate of the pedal.
Either way, just based on the statements they made anyone who has worked with anything robotics or autonomous can tell you that first instinct would be sensor failure. Of course, all of this is impossible to tell without the raw log data. Companies who are unwilling to provide such data or make it difficult to do so generally have something to hide in either their implementation, data quality, or the actual findings from the logs themselves.
It would also be interesting to see what legal basis (https://gdpr-info.eu/art-6-gdpr/) under GDPR they use to collect the data in the first place, or whether they simply don't collect that data from the EU.
I suspect that this policy violates some privacy provisions in some states and countries. Don't California (and probably other) laws guarantee you the right to obtain any data a company has on you?
The CA state Web site says:
"You may request that businesses disclose to you what personal information they have collected, used, shared, or sold about you, and why they collected, used, shared, or sold that information. Specifically, you may request that businesses disclose:
-The categories of personal information collected
-Specific pieces of personal information collected
-The categories of sources from which the business collected personal information
-The purposes for which the business uses the personal information
-The categories of third parties with whom the business shares the personal information
-The categories of information that the business sells or discloses to third parties
Businesses must provide you this information for the 12-month period preceding your request. They must provide this information to you free of charge."
The wording is unfortunately vague, though. Disclosing WHAT information they've collected may be interpreted as different from disclosing THE information they've collected. However, I wonder if the "specific pieces" clause might be a basis for demanding "all data specific to a crash on such-and-such a date."
From that page: "Pursuant to Cal. Civ. Code Sections 1798,100(a), and 1798.110(a) and (b), a consumer has a right to request, and a business that 'collects personal information about a consumer' has an obligation to disclose and deliver upon a verifiable request, 'the specific pieces of personal information the business has collected.'"
It seems to me that in California, Tesla has to cough it up.
It's only if they link the data to an identity. If I were Tesla I would collect the data in an anonymous manner, meaning the data isn't YOURs i.e. there is no way to identify you in the data. Likely they do this via the VIN, it is not private, attaching data to VIN is vehicle data not personal data.
Pretty sure that would be compliant with California/European privacy law.
That’s not how it works. If the data is connected to the VIN it is likely considered Teslas data, not yours. (Just like googles logs are theirs, even though your IP/fingerprint is in there). The data they store explicitly against your identity is yours by right in those states that have decent privacy law. It doesn’t include secondary/tertiary/derived data.
If you request your data from Google - you don’t get any web logs… but they definitely have them.
In this case it looks more like loss aversion, protecting themselves from customer lawsuits. I'm sure there are other use cases for the data. Self-driving improvements have been mentioned.
They're not selling it as far as I know, but you're essentially providing them with free training data for their autopilot/self-driving AI which they do/plan to make money off of. In that respect it seems odd that they would be so stingy in handing data over to the person that generated it though you can apparently purchase a connection kit to access the event recorder data for about $3k.
Funny how they are so "protective of the customer's privacy" that they have their CEO tweeting about the data but can't share it privately with the affected customer.
Tesla doesn't have a PR department but they're full of PR bullshit anyway, starting at the top.
He does seem to know a bit about rockets though. At least enough not to sound stupid. I am not talking about his Mars colony ambitions but practical engineering problems like steering a falling rocket.
I still agree that he is "full of flim-flam and bluster" in most of his communication.
I wouldn't call him an engineer, he is not, he is a businessman, but one with enough engineering skills to understand the engineers he hired. He probably would be a decent engineer if it was his primary job.
Tesla deserves credit for popularizing electric cars, but the technology is hardly all of their making. Indeed, one could say that the Tesla is the iPhone of cars - integrating existing technologies better than their competitors and showing what's possible. A lot of the other novel aspects of Tesla's cars suck though - the "giant iPad" interface, the "yoke" steering wheel, the electric door handles. Meanwhile the one thing that Musk has been talking about since 2015 at least (FSD) still doesn't exist at anywhere near the level he promised.
One could even argue that the electric car was inevitable, hybrids were already pushing electric drivetrains forward, and Musk was just in the right place at the right time.
If any of that were true it wouldn't be taking this long for auto companies with billions of cash to replicate and surpass Tesla, or for competitors to show with no ("giant iPad" interface, the "yoke" steering wheel, the electric door handles).
Instead they are still far behind and it's looking like some will probably literally die trying to compete. The battery technology at the very least has plenty of innovation. Teslas have some of the highest margins in cars and very less competition.
I mean, I think you may be suffering from US bias here (I’m assuming you’re in the US, or at least in a country with a similar auto market). The industry has, in general, not been prioritising the US market for electric cars, though this is starting to change (this shouldn’t be a big surprise; fuel is generally cheaper in the US than in Europe and the wealthy Asian economies, and range anxiety is a larger problem. Europe buys about four times as many electric cars as the US; China even more).
In the EU last year, Tesla was third in sales, after VW Group and Stellantis. It will likely be overtaken by Hyundai-Kia this year, and maybe by Renault. This also… shouldn’t be super surprising? The ‘catch-up’ argument is a little overblown; Renault’s Z-E platform and the Nissan Leaf are contemporaries of theTesla model S, and VW wasn’t far behind. Electric cars are likely having a moment more because battery prices started getting into the realms of the possible 15 years ago more than anything else.
In particular, it’s notable that VW group’s electric car market dominance is similar to its overall car market dominance.
For some it is, as they've hyped Tesla to the point they believe it will come to dominate even Ford. They think that just because Tesla has a higher market cap than Ford and GM that somehow will translate to Tesla dominating Ford and GM. And the reason they believe this is because they think TSLA is properly valued, and actually probably undervalued. They don't see the hype because they are the hype.
Why do people post "news" from 4 years ago? News mostly isn't relevant a few weeks later; it would make sense if their was some follow up story or something about this. Even weirder is how people comment on this story like it just happened, did anyone actually check to see if this is still the case 4 years later?
> Why do people post "news" from 4 years ago? News mostly isn't relevant a few weeks later
This is just a report on something they do, not a specific event or a revelation. It's not the kind of "news" that stops being relevant. It's equally relevant every year unless something changes.
You can look at any of these threads and it's basically the same stuff. It's not because it's not interesting but because it's pretty easy to exhaust most of the interesting things to say about the same topic.
there are also those folks who don't know large pieces of current events and keep surfacing 10 year ago news as actual news (and treat it as occurring now).
It would be less popular it Tesla had leadership that lead by example instead of dictate and if they stopped making promises they can't execute on year after year.
that's not why a lot of the attacks on tesla are happening - the twitter acquisition and elon being political on twitter have opened him up to lots of people attacking his companies.
Tesla is very divisive, a lot of people get a sense of power and superiority by posting anything that makes them look bad. The subtext is "See? I'm smarter than you by hating Tesla."
I don’t doubt that in the future, Teslas and Apple cars won’t have a steering wheel. Level 3 autonomy is already being tested by some manufacturers. However I can’t imagine not being able to buy a car with a steering wheel. Even if the car was super good at self driving, better than most humans, there will always be a market for low tech drive it yourself cars.
Even if what you say is true, I predict that there will be some municipalities which will ban human driven cars, if for no other reason, than for the control of crime. (There are already shenanigans played with accessibility by public transport, motivated this way.)
The 2 year contracts are really just payment plans in disguise (sometimes there isn't even interest!); I'm not sure there is any major provider that wants your phone back after your contract is up. (I have personally not had a cell phone contract in over a decade.) This is a relatively recent change, within the last 5 years or so. Back in the day, yeah, you were just renting your phone. (And the same might be true of your wired ISP. You definitely don't own your CPE unless you made some special deal.)
Netflix vs. buying physical media is not clearly an ownership thing. A year's worth of Netflix's middle-tier plan buys you 9 DVDs. (Actually, I don't think they make DVDs anymore, so BluRay or something.) So if you were going to watch more than one film a month, you come out ahead there.
I thought about amending my earlier comment to mention apples iPhone subscription service. It's still a rent to own thing, but the trend here is clear: people don't really care about owning, they care about being able to afford the benefit of using the device at the most affordable upfront cost.
Anecdotally, I don't know a single person who has bought a DVD in years. And when people do buy media, they don't buy a disk but a digital license to the media on a store.
Another example? Gabe Newell famously said piracy is a convenience issue. People will pay for convenience. It's also why furniture rental companies (like Feather) are proliferating versus closing up shop. Flexibility is king, and ownership is the opposite of flexibility, it is liability.
Well if you buy Tesla insurance, it seems like you won't need the data, and in the cases where you really do it would be extremely straightforward to compell it legally due to their obvious natural bias.
Mind you, 'unintended acceleration' reports come from all manufacturers and all models of cars -- so a fair starting position is that the driver mistook the accelerator for the brake. Nevertheless, it seems that giving the crash data promptly to the consumer would clarify to the consumer that they have 'fat feet', and settle the matter, at least in the consumer's mind.
Incidentally, I narrowly missed getting hit by a lady who drove her car (non-Tesla) into a local Fedex-Kinkos, having left the store 5 minutes earlier, and returned 5 minutes after her car 'parked' inside the store.