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

Phantom Braking

This site may earn commission on affiliate links.
Here is what people know. THE 5HIT aint working. No one cares about "well, its due to the SQL redundancy of the database being misaligned with the DOJO and the cloud agnostic elaboration of the RGB camera version 2.1, and none of that is due to engineers lack of leaning in"

ITS BRAKING ON AN EMPTY 6 LANE INTERSTATE IN BROAD DAYLIGHT. FOR $12K. CUSTOMERS ARENT HAPPY
no, no, no! It's working perfectly! It doesn't matter if it's using bad data, bad assumptions, bad sensors or anything else. The only thing that matters is that the algorithm is working perfectly as designed. If it's not doing it's job then that's your problem for not liking the supremely perfect algorithm that Tesla has written! /s

But yes, you are exactly correct. As I've said before, what matters is results. People expound on vision vs radar and GPS map data, etc but in the end what matters is what works. If the have a tribe of monkeys making the decisions and they get it right I'd be perfectly happy!
 
Good that it it getting some attention in the news I guess. I just hope they don’t “fix” it by removing the feature. I do like AP/TACC even if is does have phantom events.
Ultimately, that's what may matter most. The negative press will start to hurt sales and also get the attention of investors and board members. I imagine they won't be too impressed by Sonic or fart noises.
 
It's important to note that as of yesterday's AP press report, not a single accident or injury has yet to occur with regards to 'phantom braking' in this gov't investigation. It seems that the department is reacting to the number of complains rather than the severity.

I wonder how many lives the system has saved thus far...

View attachment 770788
It does say none reported, which means there could have been accidents that haven't been reported. That said, I haven't heard anyone here complain of anything but near misses (maybe I haven't read every account). With as many accounts as I have read, it is actually very surprising that there haven't been any reported accidents, and it really, really makes me wonder why.

I certainly hope that means that far more lives are being saved (and accidents avoided) with this highly sensitive, low specificity system. I also hope the specificity can be improved upon as well.
 
... what matters is results. People expound on vision vs radar and GPS map data, etc but in the end what matters is what works. If the have a tribe of monkeys making the decisions and they get it right I'd be perfectly happy!
THIS. RESULTS. If you are going to talk the talk, walk the walk. No one cares about all the various excuses. Less time tweeting Hitler references and more time sleeping in Fremont working on code or hardware or whatever it is, to fix the problems.
 
Here is a comment on Reddit. I wonder if this is why I rarely ever see phantom braking and some see it all the time ? I’ve the following distance set to max and FCW to Early.

A friend of mine has 84000 miles in, mostly Autopilot with 5 months of FSD Beta. He says with the following distance set to max (7) and Forward Collision Warning to Early, he's never had phantom braking.​
Out of curiosity, why are you using "Early" and not "Late"? Wouldn't Late cut down on PBs more?
 
That setting is what the safety score uses … so, no thought behind it.

But, early could mean slowdown earlier but more slowly ? This would reduce the deceleration and severity of PB.
My understanding is that FCW is the red wheel warning you get when it thinks there could be a collision. If you don't do anything when the red warning comes up, then AEB engages to stop you.

So setting it to Early would mean that if a car slows down ahead of you, the red warning will happen quickly. Late will let the car slow down more and the distance reduce more before the red warning comes up. I haven't tested it, so I can't be 100% sure on this.
 
  • Like
Reactions: SpaceKat
Ultimately, that's what may matter most. The negative press will start to hurt sales and also get the attention of investors and board members. I imagine they won't be too impressed by Sonic or fart noises.

They have some margin right now where they could fix some of these issues. Ie there’s a backlog on sales. Now is the time to do it thou.
 
  • Like
Reactions: Xdama and sleepydoc
They have some margin right now where they could fix some of these issues. Ie there’s a backlog on sales. Now is the time to do it thou.


There's been a backlog on sales for years.

There'll continue to be one for years going forward.

Demand for good EVs greatly exceeds supply and that won't be changing for a long time.

As to improving FSD, they're obviously working on that- and have been on an ongoing basis.

Meanwhile the "competition" is garbage like the Mach E system that just gives up during gentle highway curves
 
I've noted this before but I will again: Tesla HAS implemented dumb cruise control and exists in the software that ships today, because early Model 3s were available without TACC. I think precious few buyers actually opted for that configuration, but it was out there. So, I can't speak for the Y, but I suspect it would be trivial to offer it as a stop-gap on the 3 until this all blows over. Of course, the optics of "admitting defeat" and rolling out dumb cruise control wouldn't be great for them, but... not my problem.
 
  • Like
Reactions: gforce2002
Another thing I'll add is that several BMW Group products over the past several years have used a single-camera adaptive cruise system without any radar, and it doesn't suffer from any phantom braking in my experience - BUT it also disengages itself for all sorts of reasons. It hates bright direct sunlight, for instance. I don't know that any shipping vehicle has figured out a flawless vision-only adaptive cruise system.
 
Another thing I'll add is that several BMW Group products over the past several years have used a single-camera adaptive cruise system without any radar, and it doesn't suffer from any phantom braking in my experience

*sigh*

MEANWHILE ON THE BMW FORUMS...


BMW owner said:
I was driving along a dual carriageway yesterday afternoon (in London) at approx. 50mph when my X3 suddenly slammed the brakes on

Different BMW owner said:
driving along and there was an undulating road (almost like a bump) and the car started to brake (was on active/adaptive cruise control).

Different BMW owner though he's talking about his VW here said:
we have a Golf Alltrack which has a lot of the same type of systems and we've had it mis-read situations and emergency brake or do various other odd things when in traffic, or suddenly slow down while in adaptive cruise when there's nothing going on

Yet another BMW owner said:
Ironically mine (M40i) just did this yesterday too.

Yet another BMW owner said:
I joined this group just for this issue!
It has happened twice in two weeks to my girlfriend while driving our new X3. Both times straight road and (thank god) no other cars around.

Yet another BMW owner said:
I had an identical situation where there was no one in front of me, and I was traveling about 50 mp


Or from a different BMW forum


BMW owner said:
Auto-Cruise: Driving along, the car will suddenly brake as though something is there but there isn't

Another BMW owner in the thread said:
Yes it has happened to me more than once when the cruise suddenly stopped. Glare, shade and trees suppose to do that i heard

Yet another BMW owner said:
Yes, I've had the cruise control suddenly shut off, or brake, because it thinks it seems something when there is nothing there.

Yet another BMW owner said:
my wife was driving on an empty expressway with cruise control on and set at 67, there were no other cars on the road, as she approached a "hill" overpass the car braked hard





Every adaptive cruise system has these issues with false positives causing braking.

Every. One.


But BMW doesn't have a rocket company competing with the owner of the Washington Post, and mentioning BMW doesn't attract clicks anyway, so it's not a headline for anybody.
 
  • Informative
Reactions: philw1776
Have you seen the earnings reports and wait times, lately?
Yes, but we're also in a dynamic and artificial situation. Tesla was an (extremely) early leader in EV development and earned a justifiable reputation for being the leader. They also have that image so when they introduced their more mainstream models 3 and Y people naturally gravitated towards them.

Tesla still has one of the best EV platforms on the market, but they are having issues with quality control and service, and many people don't care about FSD; they want a car comparable to the ICE car they're replacing. Many other car companies are entering the market with strong offerings. Tesla already has some competition and it will get stronger.

The supply chain issues that the industry is facing also adds a different facet to the situation. Tesla and all the other manufacturers are supply-constrained so Tesla's waiting list is artificially long and the competition is artificially suppressed.

All this means that Tesla is fine for the time being, but they also can't afford to sit on their laurels and ignore issues, either. Once the market corrects in a year or so things could well be quite different.
That setting is what the safety score uses … so, no thought behind it.

But, early could mean slowdown earlier but more slowly ? This would reduce the deceleration and severity of PB.
I think a lot of the PB complaints would be reduced it the slowing/baking wasn't as dramatic. Like I've said before, we really don't know what's going on. It may just be that the algorithm is taking its electronic foot off the accelerator for a few seconds. In an ice car, this would just be coasting and only result in a drop a few MPH but with regenerative braking it's much more dramatic and noticeable. This wouldn't change the AEB and other incidents where the car appears to be actively braking, though.
 
That’s what I thought! He seemed to be referring to me! 😏

I am not planning to buy the EQS. It’s the EQE on top of my list. I love the interior, the rear wheel steering, MBZ build quality..

I bet I can get $30K for my Model S. It’s in top condition!

So about $40K or so outlay to get the EQE. I’ll take it any day over the current Model S LR. I’ll be getting out of the Tesla jail! 😅
Yeah, I just glitched and didn't realize there was another post between what I was responding to.

Oops.
 
  • Funny
Reactions: DrGriz
Another thing I'll add is that several BMW Group products over the past several years have used a single-camera adaptive cruise system without any radar, and it doesn't suffer from any phantom braking in my experience - BUT it also disengages itself for all sorts of reasons. It hates bright direct sunlight, for instance. I don't know that any shipping vehicle has figured out a flawless vision-only adaptive cruise system.
no, but many have implemented systems that are much closer to flawless. I've repeatedly mentioned my experience with our 2020 Subaru Forester. In a little over a year of ownership we've experienced zero phantom braking events and zero false AEB events. I'm not saying the system is perfect but at this point I can say both the AEB and adaptive cruise control systems are dramatically better.
 
*sigh*

MEANWHILE ON THE BMW FORUMS...















Or from a different BMW forum














Every adaptive cruise system has these issues with false positives causing braking.

Every. One.


But BMW doesn't have a rocket company competing with the owner of the Washington Post, and mentioning BMW doesn't attract clicks anyway, so it's not a headline for anybody.

Fair enough! I appreciate the thorough debunking - and apologies, didn't mean to cause any consternation. I was mainly pointing out that I didn't experience "phantom braking" (at least as Tesla owners, myself included, have been experiencing it) when I owned one of these camera-only BMW systems, but I also recognize that my own experience is a sample size of one and really not helpful to the conversation.
 
Phantom breaking is real. If you want proof, try driving fm2657 in central Texas. It is an approx. 20 miles road. It is curvy but the PD happens mostly on dry straight sections with no on coming traffic on a clear day. It happens about 30 times on that road every time I try to use FSD or ACC. It gets so bad that I can not use it if there is a car close behind me. I have notified Tesla numerous times.
 
Something else to remember in all these discussions - some members persistently post stories/reports of Issues with AEB activations. Those are a separate issues from TACC phantom braking and should be considered separately.

The other point is that it’s a matter of degrees. It’s easy to make a black/white comparison, but unless a system is perfect it will have false activations. Simply posting a random report of an AEB activation is both meaningless and irrelevant to the discussion about TACC phantom braking. (It does serve to confuse the picture which is the likely goal.)

Tesla, unfortunately has both an AEB system and a TACC system that are strikingly below average for the industry.