Phantom Braking

[OncomingStorm]

How many people on this thread don’t own a tesla? How many The tens to hundreds of phantom braking threads on TMC were started by people who don’t own a Tesla? And what’s the point of your post?

i'm not saying YOU dont own a tesla, but read through enough of these and its very clear, there is handful of people griping and dozens chiming in, with a good chunk of them having very clearly no or extremely limited experience with the car and its functionality and are making their arguments based on other people's, sometimes very suspect, claims. generally people bitching on the internet is NOT good indicator of anything!

 

[AndreP]

i'm not saying YOU dont own a tesla, but read through enough of these and its very clear, there is handful of people griping and dozens chiming in, with a good chunk of them having very clearly no or extremely limited experience with the car and its functionality and are making their arguments based on other people's, sometimes very suspect, claims. generally people bitching on the internet is NOT good indicator of anything!

Relying on your own personal experience with the car could be problematic in the opposite direction with the hugely variable performance of these systems depending on location and a range of other factors, not the least of which is how each individual perceives the experience.

If I was Tesla, I'd be much more concerned about that perception. Tesla is playing a high-risk game and IMO is demonstrating what could be some significant hubris.

 

[Big Dog]

...and here's a case where phantom braking actually caused an accident:

Phantom braking from FSD

I just received FSD for a week and was really exciting to enjoy it. Unfortunately, yesterday when I was using the FSD to turn left at an intersection, a phantom brake appeared suddenly and the car behind me failed to react and crashed into each other, causing a rear-end collision. I don't know...

teslamotorsclub.com

Correlation <> causation.

Hint: Rear-enders at street corners occur everyday in many communities in cars without any self-driving features. So just claiming it happened bcos of Phantom Breaking doesn't make it so.

That said, I have experienced PB numerous times and do not downplay it. It's so bad that can't use it when driving with the wife as she goes crazy. (Yeah, its her issue, but that makes it my issue when I'm driving.)

 

[Big Dog]

Here is another hypothetical situation. You are approaching a traffic light and there is no one ahead of you. The traffic light is green for you.
Will you slow down as you approach and pass through or will you maintain or even speed up going through it?

Not doc, but I'll bite.

To me, it depends on the street/situation. If the street I'm on is residential and the speed limit is 25 mph, I will slow a tad as I go thru a green light to double check corners bcos you never know if a kid or dog may run out. If the street is 2-3 lanes in both directions and the speed limit is 45+, no way I'm slowing down.

 

[OncomingStorm]

Relying on your own personal experience with the car could be problematic in the opposite direction with the hugely variable performance of these systems depending on location and a range of other factors, not the least of which is how each individual perceives the experience.

If I was Tesla, I'd be much more concerned about that perception. Tesla is playing a high-risk game and IMO is demonstrating some significant hubris.

i'm not sure i fully understand your comment so if i'm off base forgive me.... to my comment about how many ppl on here dont own Teslas, thats pretty evident not just from me making assumptions but also from many users full on admissions. and thats fine, i'm not saying they are lying, they are just chiming in because they think something to be the case and making the overall argument appear more credible than it actually is when their facts are simply wrong (not a judgment just a statement). I very much agree that those who experience lots of problems with the system are in a niche, their specific situation is something the car can't handle very well, and i'm sympathetic i'd be pissed too, but me not being happy with a product doesn't really give me right to demand they change their product just to suit me.

i do agree Tesla is doing themselves damage by not addressing these kinds of problems at all, even if at the end of the day it was simple PR efforts that would be better than radio silence imho (tho a good argument can be made for the silence being more effective). i mean the bottom line for them is they simply can't produce cars fast enough to sell them so why spend 1 cent trying to appease a very small number of people when the vast majority are perfectly happy (remember happy customers rarely go online because the thing the bought did what they expected it to do lol).

 

[Sporty]

To my knowledge, despite many many many years of the occasional person here screaming OMG THE PHANTOM BRAKING, there's been zero accidents caused by it.

If it was an actual significant widespread problem …

Regardless. It’s poor quality is enough where I no longer use it in numerous cases where I would like to. If at all. That is enough to post & say “hey this feature isn’t usable / has a serious (usability) problem”. It leaves a bad taste in the users month. And if it happens enough we will see many complaints, people posting & blogging about it. And maybe news articles.

All that is happening. Why it is a problem for me, I’m just as sure there are a lot of instances of Tesla’s where it is not a problem. Eg it’s a fine statement to say “works for me” just like it’s fine to say “it doesn’t work for me”.

But to say it’s not a problem for anyone doesn’t fly. Multiple people are reporting and stating that it has a problem (for them).

 

[sleepydoc]

Not doc, but I'll bite.

To me, it depends on the street/situation. If the street I'm on is residential and the speed limit is 25 mph, I will slow a tad as I go thru a green light to double check corners bcos you never know if a kid or dog may run out. If the street is 2-3 lanes in both directions and the speed limit is 45+, no way I'm slowing down.

True but it’s really irrelevant to the phantom braking issue.

 

[AndreP]

If you want to see the actual complaints/reports, the submissions appear to be public information although I'm guessing none of them have been confirmed by the NHTSA so we don't know if any are bogus with fake VINs etc

But you can see complaints by model and issue, for example

2021 TESLA MODEL Y 5-SEAT SUV RWD | NHTSA

www.nhtsa.gov

Scroll down to the Recalls & Safety Issues section and you can filter by category, it's pretty user-friendly and you can read all the details

These reports are exactly what people should be submitting, light a fire under the regulator to light a fire under Tesla to get these issues rectified.

 

[Knightshade]

Those reports often (and do in at least one you list) cite using AP on 2-lane non-divided roads.

That's user error.

Not that it never does it on just TACC or anything, but one needs to separate out actual system issues with system being used where not intended.


Note too all the cases of "Well, nobody hit anybody, but they COULD HAVE"

If it was actually dangerous instead of just surprising you'd see actual accidents.


One confusing one is where the guy says a motorcycle cut in front of him and the car braked.... it's [B}supposed to[/B] when that happens.

 

[AndreP]

Those reports often (and do in at least one you list) cite using AP on 2-lane non-divided roads.

That's user error.

Not that it never does it on just TACC or anything, but one needs to separate out actual system issues with system being used where not intended.


Note too all the cases of "Well, nobody hit anybody, but they COULD HAVE"

If it was actually dangerous instead of just surprising you'd see actual accidents.


One confusing one is where the guy says a motorcycle cut in front of him and the car braked.... it's [B}supposed to[/B] when that happens.

I think if I was working with the regulator and saw numerous reports of people being worried about their vehicle behaving in ways that could cause accidents in scenarios and particularly scenarios where the system is being used where it shouldn't, I'd be looking at ways to either better educate the drivers or wonder why the system is allowed to operate under those circumstances along with some risk vs reward consideration.

I don't think the NHTSA would be looking to point fingers or curb innovation, they just want the roads to be as safe as possible for users. But something wrong is happening here and there must be ways to mitigate it before these leading indicators become real accidents. The NHTSA wants to prevent accidents, not let accidents occur and then say "well you should have read Page 382 in the manual where it says to not use this on 2-lane roads".

If 2-lane roads are such an issue, maybe there needs to be a pop-up warning when AP is activated.

 

[Knightshade]

NHTSAs job is actual safety though.

So if accidents were being reported, that would certainly prompt an actual open investigation (indeed, the Nissan and other unexpected baking incidents I cited earlier involved numerous actual accidents)

PERCEPTION of safety isn't really their job though.

"Owner didn't read the manual" or "Owner was surprised by behavior that ultimately didn't endanger anyone" is pretty low on their list of stuff to address.

 

[Sporty]

Those reports often (and do in at least one you list) cite using AP on 2-lane non-divided roads.

That's user error.

.

See “causes” under https://en.m.wikipedia.org/wiki/User_error

Too many user errors is a symptom of a bad design. Hearing “it’s user error” frequently is a red flag that perhaps there is something more subtle going on.

Or one could ignore their users. Which I suppose if you have back orders on all you can manufacture who cares?

But these are different departments. The sw team could be paying attention to this. And clearly with fsd attach rate going lower maybe they should start working on some qa issues.

 

[AndreP]

When real accidents occur, it's already too late.

We would call this type of stuff "leading indicators" in engineering/construction and there are industry-standard ratios for probability of injury/fatality. A severe phantom braking event without an accident could be considered a near miss. Just to throw made-up numbers out there to illustrate, you could determine about 1,000 near misses would statistically result in an incident. For every 500 incidents, you could expect 20 injuries, and for every 100 injuries you could expect 1 fatality. And then you can calculate how many near misses would statistically result in a fatality.

The NHTSA needs to watch their backside as well, because leaders and politicians will hold them accountable for deficiencies in regulation when accidents do occur.

 

[Knightshade]

This isn't bad design though- and it's not new. Folks have been ignoring the manual on where/when to use AP for years and years.

When real accidents occur, it's already too late.

We would call this type of stuff "leading indicators" in engineering/construction and there are industry-standard ratios for probability of injury/fatality.

Except folks have been complaining about TACC phantom braking for 7-8 years now going back to AP1.

And same on every other car makers TACC-equivalent system.


Some are so bad there ARE accidents (I cited several) and NHTSA takes action.

The rest of the companies it's just annoying, and does not cause accidents (Tesla and many others too) and NHTSA doesn't take action because it's not an actual safety defect

 

[particleman]

Those reports often (and do in at least one you list) cite using AP on 2-lane non-divided roads.

That's user error.

Not that it never does it on just TACC or anything, but one needs to separate out actual system issues with system being used where not intended.


Note too all the cases of "Well, nobody hit anybody, but they COULD HAVE"

If it was actually dangerous instead of just surprising you'd see actual accidents.


One confusing one is where the guy says a motorcycle cut in front of him and the car braked.... it's [B}supposed to[/B] when that happens.

From the Model Y User manual under TACC

Model Y Owner's Manual | Tesla

Learn about Traffic-Aware Cruise Control and its specifications.

www.tesla.com

"Traffic-Aware Cruise Control is primarily intended for driving on dry, straight roads, such as highways."

I haven't been able to find where you can't use Autosteer or TACC on a two lane highway. If that is the case Tesla should not allow it to be enabled.


This Verge article was a direct result of the Washington Post article. The authors of the Post reached out to me directly regarding my NHTSA complaint to verify it was accurate and if I had anything to add. I can't speak to anyone else that was quoted, but I'd assume they did the same. I'm quoted in the article and I stand behind the things I said in the complaint and the article. My wife was nearly rear ended as a result of the phantom braking on 1 of 2 occasions. We were absolutely not willing to continue risking an accident. We exclusively used TACC NOT autosteer. We dealt with it daily on our 2021 Y, I filmed mild cases of it before we traded the car in November.

https://www.washingtonpost.com/technology/2022/02/02/tesla-phantom-braking/

 

[AndreP]

This isn't bad design though- and it's not new. Folks have been ignoring the manual on where/when to use AP for years and years.





Except folks have been complaining about TACC phantom braking for 7-8 years now going back to AP1.

And same on every other car makers TACC-equivalent system.


Some are so bad there ARE accidents (I cited several) and NHTSA takes action.

The rest of the companies it's just annoying, and does not cause accidents (Tesla and many others too) and NHTSA doesn't take action because it's not an actual safety defect

I think you're thinking the NHTSA is far more hands off and draconian than they actually are, they are in constant communication with Tesla regarding this stuff and lots of changes are happening behind the scenes that they are influencing. Like this uptick in phantom braking reports, boil it down and there's likely just someone at the NHTSA who emails someone at Tesla asking why they think this is happening and what can be done about it.

There doesn't need to be a true safety defect and the immediate solution doesn't need to be magically fixing phantom braking on 2-lane roads, there are other subtle tweaks that can have a huge impact with minimal effort/cost.

We can look at something like this NHTSA report regarding Autopilot accidents from back in 2016

https://static.nhtsa.gov/odi/inv/2016/INCLA-PE16007-7876.PDF

No defect was identified

A safety-related defect trend has not been identified at this time and further examination of this issue does not appear to be warranted.

But the Human-machine interface gained some appreciation

Tesla appears to have recognized HMI factors, such as the potential for driver distraction, in its design process for the Autopilot system. Tesla's design included a hands-on the steering wheel system for monitoring driver engagement. That system has been updated to further reinforce the need for driver engagement through a "strike out" strategy. Drivers that do not respond to visual cues in the driver monitoring system alerts may "strike out" and lose Autopilot function for the remainder of the drive cycle.

And this is the origin of the three-strike and lockout aspect of AP. It happened after accidents occurred, but obviously it would be preferable to avoid any accidents at all.

I imagine that Tesla finally using the cabin camera for driver monitoring also came about in a similar way, and who knows what else that we aren't aware of.

 

[drtimhill]

To my knowledge, despite many many many years of the occasional person here screaming OMG THE PHANTOM BRAKING, there's been zero accidents caused by it.

If it was an actual significant widespread problem that actually caused accidents there'd be actual evidence of it by now.

Instead it's just a bunch of random anecdotes.... which are themselves self selecting because the vast majority for whom everything is fine are much less inclined to randomly make videos saying "Here's a drive where nothing happened" or even "Here's where the car slowed down slightly- it was surprising, but watching the video in retrospect, not actually dangerous to anyone" which is what is usually actually happening.

+1

 

[drtimhill]

All that is happening. Why it is a problem for me, I’m just as sure there are a lot of instances of Tesla’s where it is not a problem. Eg it’s a fine statement to say “works for me” just like it’s fine to say “it doesn’t work for me”.

But as has been noted, Tesla are working on a fix .. its called FSD, which, as a side-effect of having to be much more sophisticated in understanding vehicles and oncoming trajectories, also handles phantom braking much more effectively. I strongly suspect their view is that this is the long-term solution (since even simple TACC will eventually use the same NN stack, even if you dont have the full FSD option).

What this means, imho, is that if enough people complain to NHTSA to force them to intervene, they wont get a quick fix at all .. they will get Tesla disabling TACC entirely until FSD is complete. So, "be careful what you ask for" guys.

 

[DayTrippin]

Fine with me, disable TACC and implement a dumb cruise control in the interim. I am good with that option.

 

[drtimhill]

Fine with me, disable TACC and implement a dumb cruise control in the interim. I am good with that option.

What makes you think they would do that? And are you saying that if its fine with you everyone else has to go along with it? It's "fine with me" if they leave TACC/AP as-is since I dont get any, so therefore I declare phantom braking isnt an issue! Are you ok with that? If not, why?