Firmware 8.0

[sillydriver]

After reading various articles and the blog post, one question comes to mind: under 8.0, will my car activate AEB when there is a coke can rolling laterally across the highway, with the bottom facing the radar? All the conditions would be met, no?

- large object (at least to the radar)
- object is moving
- object is not whitelisted

This kind of scenario definitely concerns me and I'm sure we can invent others. The problem is that emergencies often involve odd situations that are hard to anticipate (otherwise they wouldn't cause emergencies). For example, here is an situation where I already did get a false positive for emergency braking under 7.1: I was heading into a roundabout at about 25 mph where a car in the circle was passing directly in front of me from left to right at about the same speed. My intent was to tuck right in behind him without reducing speed (and no, there wasn't another car in the roundabout, I wasn't trying to cut in front of anyone -- I believe my maneuver was perfectly legitimate). Anyway, my MS interpreted the situation as heading into the side of another car at ramming speed and slammed on the brakes. Fortunately there was no car following right behind me.

My fear is that until there is a awful lot of fleet experience and algorithm adjustments to allow for unanticipated situations where there car actually shouldn't slam on the brakes, there will be more exposure to being rear-ended and safety will be reduced rather than enhanced. I suppose the car could be programmed to check it's rear ultrasonic sensors and not apply emergency braking if closely followed.

 

[Vitold]

Because they had to due to involvement of regulators looking into AP after the fatal crash. It's likely they tried "mild tweaks" but they didn't solve the problem.

You should read Elons blog again. It's all about detecting objects that are moving and overhead objects plus keeping the drivers hands on the wheel and focused on what's going on. All of those things are what Elon described at the time as being factors in the crash. If you don't think the NHTSA is all over this I have a bridge to sell you in Brooklyn.

Yes I doubt that Tesla is doing this because of NHTSA investigation. I expect NHTSA conclusion to be that system is not perfect and cannot and it was drivers responsibility.

Tesla did not need to improve the Autopilot because of that crash. Tesla improved it because they want to be in the lead.

 

[Vitold]

This kind of scenario definitely concerns me and I'm sure we can invent others. The problem is that emergencies often involve odd situations that are hard to anticipate (otherwise they wouldn't cause emergencies). For example, here is an situation where I already did get a false positive for emergency braking under 7.1: I was heading into a roundabout at about 25 mph where a car in the circle was passing directly in front of me from left to right at about the same speed. My intent was to tuck right in behind him without reducing speed (and no, there wasn't another car in the roundabout, I wasn't trying to cut in front of anyone -- I believe my maneuver was perfectly legitimate). Anyway, my MS interpreted the situation as heading into the side of another car at ramming speed and slammed on the brakes. Fortunately there was no car following right behind me.

My fear is that until there is a awful lot of fleet experience and algorithm adjustments to allow for unanticipated situations where there car actually shouldn't slam on the brakes, there will be more exposure to being rear-ended and safety will be reduced rather than enhanced. I suppose the car could be programmed to check it's rear ultrasonic sensors and not apply emergency braking if closely followed.

Did you have AP engaged at that time or was that AEB?

One of the articles about v8.0 said that false positive (braking) is expected once every 5 years. I don't know how they figured that out but that to me is pretty strong confidence in the system.

 

[sillydriver]

Did you have AP engaged at that time or was that AEB?

One of the articles about v8.0 said that false positive (braking) is expected once every 5 years. I don't know how they figured that out but that to me is pretty strong confidence in the system.

I certainly didn't. This was emergency braking independent of AP. My fear is that more emergency braking whether in AP or not would lead to more false positives.

 

[nico180]

Neen, niet meer dan dat het de grootste update wordt die ze ooit gedaan hebben.

Vanwege de "belangstelling" voor AP heeft Elon de aanpassingen die daarop betrekking hebben apart naar voren gehaald en daar duidelijkheid over gegeven.

De rest komst straks met de update. We zullen nog even geduld moeten hebben.

en dit: Approximately 200 small enhancements that aren't worth a bullet point heeft nog steeds betrekking op Autopilot.

wauw, dat is dan mooi

 

[Vitold]

I certainly didn't. This was emergency braking independent of AP. My fear is that more emergency braking whether in AP or not would lead to more false positives.

That's definitive something they need to tweak than. Having more redundancy, as in camera + radar, could prove to be better than camera alone.

 

[travwill]

On the what other features are part of 8.0, I think they purposely held off on those for now until its release in 1-2 weeks. They craft this is advance.

This was about AP functionality, safety, for a TSLA stock pop potentially.

1-2 weeks 8.0 release will be about interface, look/feel, functionality, enough to be a post also and make news, for a TSLA stock pop potentially.

 

[AmpedRealtor]

What's in store for us pre-AP owners?

We are probably one of the 200 bullet points that weren't worth mentioning.

 

[AmpedRealtor]

You should read Elons blog again. It's all about detecting objects that are moving and overhead objects plus keeping the drivers hands on the wheel and focused on what's going on. All of those things are what Elon described at the time as being factors in the crash. If you don't think the NHTSA is all over this I have a bridge to sell you in Brooklyn.

I very much agree, I smell the NHTSA in Tesla's new nanny features and wouldn't doubt that such features were implemented in consultation with the NHTSA. Much in the same way disabling lowering back in 2013 was also done in consultation with the NHTSA.

 

[Panu]

Elon:

"The reason is that it would see a large metal object across the road and knowing that there’s no road sign, there’s no overhead road sign, in that position, therefore it would not be a whitelisted situation and probability would be assessed as high and so it would brake."

So a new overhead road sign will cause emergency braking? A road never driven with an autopilot Tesla will be emergency brakings all the time?

 

[tomas]

We are probably one of the 200 bullet points that weren't worth mentioning.

The blog was clearly about AP and only AP. I took the 200 points as minor AP fixes/revisions. It is already known via other sources that there are UI and media player enhancements in 8. Blog not about 8 except as it pertains to AP.

Regarding NHTSA, isn't it possible that one of Tesla's chief goals is safety? And that they are not simply kow towing, but doing their best to respond to issues as they arise out of desire for safety? If their intent is to address causes of FL crash, including driver in attention, what's wrong with that? As I recall, they were proactive with the lowering. Weren't told to do it. Could be same here. Why do so many on this forum jump to negative conclusions?

 

[aesculus]

So a new overhead road sign will cause emergency braking? A road never driven with an autopilot Tesla will be emergency brakings all the time?

I agree, this sounds conflicting. I guess we will have to see how it evolves. I came away thinking nothing would happen. So if a truck that is moving is detected then it will react. If a truck was parked blocking the road, then the AP would not react at all, unless three or more cars saw it first?

 

[Saghost]

So a new overhead road sign will cause emergency braking? A road never driven with an autopilot Tesla will be emergency brakings all the time?

These are the interesting questions of the day - and they're being talked about in three or four threads right now.

Taking the second half first, my guess is that the phasing in Elon described isn't a matter of time since the firmware release, but rather of number of trips recorded over the road.

Thus, the first couple times an AP equipped car drives the road, it'll be in the recording mode - noting radar returns, but not taking any action on them unless the camera or driver does.

Hopefully there'll be some sort of indication of whether you're on a trusted route or breaking new trail (or working without data, because you couldn't download it for whatever reason) on the instrument panel, but I wouldn't count on it.

The new sign is a tough situation. It's a false positive, but it looks exactly like a real problem to the system described. My expectation is that the car will start to brake, and the driver will need to tell the car not to worry (press the accelerator? a steering wheel button? the CC stalk?) until they have enough passes to be confident and whitelist it.

 

[thimel]

Elon:

"The reason is that it would see a large metal object across the road and knowing that there’s no road sign, there’s no overhead road sign, in that position, therefore it would not be a whitelisted situation and probability would be assessed as high and so it would brake."

So a new overhead road sign will cause emergency braking? A road never driven with an autopilot Tesla will be emergency brakings all the time?

For the new overhead sign I share your concern. For the road never driven, I would expect the info from fleet learning to know it had never been driven and tell AP not to use radar for AEB on that road.

 

[Lex]

I am anxious about this update now as I've heard contradictory info on the changes to the nags.

Is it 3 ignored audible warnings that will suspend it ? (this is the blog's verbiage, suggesting so -- "Car will not allow reengagement of Autosteer until parked if user ignores repeated warnings")
Or is it 3 "tugs" at the wheel to clear audible warnings that will deactivate it ? (as seems to be suggested by some of the news articles)

This is the difference between useful (former) or useless (latter) to me. And in fact if it's the latter I can already foresee "expert users" needing to make emergency stops to reset the Autopilot. Is this how we enhance safety ?

 

[Saghost]

I am anxious about this update now as I've heard contradictory info on the changes to the nags.

Is it 3 ignored audible warnings that will suspend it ? (this is the blog's verbiage, suggesting so"Car will not allow reengagement of Autosteer until parked if user ignores repeated warnings"
Or is it 3 "tugs" at the wheel to clear audible warnings that will deactivate it ?

This is the difference between useful (former) or useless (latter) to me. And in fact if it's the latter I can already foresee "expert users" needing to make emergency stops to reset the Autopilot. Is this how we enhance safety ?

I sure hope it isn't the latter. Maybe it's different for other cars, but I routinely get the warning when I'm actually holding the wheel lightly, and I actually have to turn it a fair bit against AP to get it to clear. So even when I'm in the seat, paying full attention, and holding the wheel, I'm tugging at the wheel to clear the warning (usually the first one, not the audible one.)

 

[Panu]

Is it 3 ignored audible warnings that will suspend it ? (this is the blog's verbiage, suggesting so -- "Car will not allow reengagement of Autosteer until parked if user ignores repeated warnings")
Or is it 3 "tugs" at the wheel to clear audible warnings that will deactivate it ? (as seems to be suggested by some of the news articles)

What Elon said is quite clear:

"if it ignores the audible alarm more than 3 times in an hour then the driver will have to park and restart in order to enable Autosteer."

"It" refers to "the user" here. I hope this statement holds.

 

[Az_Rael]

What Elon said is quite clear:

"if it ignores the audible alarm more than 3 times in an hour then the driver will have to park and restart in order to enable Autosteer."

"It" refers to "the user" here. I hope this statement holds.

Elons statements were a bit contradictory on this point. He then goes on to say:

They get very comfortable with it and repeatedly ignore the car’s warnings. It’s like a reflex. The car will beep at them, they tug the wheel, the car will beep at them, they tug the wheel, and it becomes an unconscious reflex action. So we will see half a dozen or more, sometimes as many as 10 warning in one hour continuously ignored by the driver. We really want to avoid that situation.

Which implies "ignoring" might mean having the audio cue come up three times (I.e. "Ignoring" the visual cue). It's hard to say. How else can you "ignore" 10 warnings now and not have autopilot go into emergency stop mode?

 

[LastGas]

Elon:

"The reason is that it would see a large metal object across the road and knowing that there’s no road sign, there’s no overhead road sign, in that position, therefore it would not be a whitelisted situation and probability would be assessed as high and so it would brake."

So a new overhead road sign will cause emergency braking? A road never driven with an autopilot Tesla will be emergency brakings all the time?

First, that road sign is only in play if you're cresting a hill and visually the sign is directly in front of you, not above. Second, what Musk said is that the car for now will NOT brake for the sign. Here's the relevant portion:

This is where fleet learning comes in handy. Initially, the vehicle fleet will take no action except to note the position of road signs, bridges and other stationary objects, mapping the world according to radar. ...When the data shows that false braking events would be rare, the car will begin mild braking using radar, even if the camera doesn't notice the object ahead. As the system confidence level rises, the braking force will gradually increase to full strength when it is approximately 99.99% certain of a collision. This may not always prevent a collision entirely, but the impact speed will be dramatically reduced to the point where there are unlikely to be serious injuries to the vehicle occupants.

 

[Saghost]

First, that road sign is only in play if you're cresting a hill and visually the sign is directly in front of you, not above. Second, what Musk said is that the car for now will NOT brake for the sign. Here's the relevant portion:

I think Panu is proposing a new overhead sign on a well established route the car is confident of already - so the fleet learning has already happened.

Granting the point about the sign needing to appear to be in line for a crash (not sure how much vertical resolution the radar has,) it seems like a valid case and one the system can't address without help from the driver.