Something interesting happened with my Service Loaner and Autopilot

[⚡️ELECTROMAN⚡️]

It absolutely does

He meant it doesn't do lane changes without the driver initiating the lane change.

 

[Ostrichsak]

Yes it does.

That’s not true. Mine does... requires a user input to decide when, though, but then smoothly moves over a lane.

Au contraire. My Mobile Eye AP1 in a 2016 90D does just fine with lane changes, thank you

I have AP1 and mine does both lane changes and auto lane changes.

It absolutely does

Yeah, I think we've got that sorted out now. Half the posts in this thread are correcting the same thing.

 

[GlmnAlyAirCar]

Yeah, I think we've got that sorted out now. Half the posts in this thread are correcting the same thing.

But it's the internet. I need to reply NOW before I read everything else!

 

[Electroman]

In fact I corrected my incorrect statement in two posts later on that same page

 

[David.85D]

You know that feeling when you go through an “orange” light and look up to see someone well behind you went through after you? I was feeling a little guilty about piling on until...

 

[SeminoleFSU]

Just for the record, I drive both an AP1 Dec.2014 build MS and an AP3 Mar. 2019 build MS.

IMO, AP3 is capable of many more things than AP1,certainly. But, AP1 still does a better, less stressful job for straight highway cruising for 1 simple reason:

NO PHANTOM BRAKING!!!! Never, not once, ever, in 140,000 + miles.

AP3? Well, not very often but often enough that I hesitate to use it. I REALLY REALLY wish they would fix that.

On the other hand though a lot of the AP accidents where it failed to brake and the Tesla ran into another vehicle on the freeway have been AP1. So the phantom braking yes is annoying but I think with that comes the lower likelihood it will run into a stationary object in the road in front of you

 

[Ostrichsak]

On the other hand though a lot of the AP accidents where it failed to brake and the Tesla ran into another vehicle on the freeway have been AP1. So the phantom braking yes is annoying but I think with that comes the lower likelihood it will run into a stationary object in the road in front of you

You got a link to the stats for this claim? Most of the high profile examples I can think of were AP2+ but that is purely anecdotal... much like your claim.

 

[dannycamps]

You got a link to the stats for this claim? Most of the high profile examples I can think of were AP2+ but that is purely anecdotal... much like your claim.

I doubt there will be actual stats for this kind of thing and even if there were, the sheer difference in the number of cars on the road with AP1 vs. AP2 would completely skew the data. For example, AP1 would be 2014 though mid-2016 S and X cars. AP2+ would be 2016+ S, 3, X, and Y with significant increase in the number of cars on the road due to the 3.

The first publicized AP death was AP1 for sure when it failed to stop for a truck crossing the highway in Florida.

 

[Ostrichsak]

I doubt there will be actual stats for this kind of thing and even if there were, the sheer difference in the number of cars on the road with AP1 vs. AP2 would completely skew the data. For example, AP1 would be 2014 though mid-2016 S and X cars. AP2+ would be 2016+ S, 3, X, and Y with significant increase in the number of cars on the road due to the 3.

The first publicized AP death was AP1 for sure when it failed to stop for a truck crossing the highway in Florida.

I think you missed my point: his post was anecdotal (at best) and yet was presented as if it was a fact.

 

[bmah]

This is an interesting but confusing discussion in several ways, not the least of which is...

Mobile Eye's AP1 ...

He meant ...

...it took an extra reading when I realized there's actually two of you in this thread.

Still rocking AP1 on my S85D. While not as featureful as the latest AP hardware/software, it's rock-solid at doing driver assistance for freeway driving. No ping-ponging, and maybe two instances of phantom braking in 5 years. I love it. It'd be fun to try out the newfangled goodies in an AP3 car sometime.

Bruce.

 

[⚡️ELECTROMAN⚡️]

This is an interesting but confusing discussion in several ways, not the least of which is...





...it took an extra reading when I realized there's actually two of you in this thread.

Still rocking AP1 on my S85D. While not as featureful as the latest AP hardware/software, it's rock-solid at doing driver assistance for freeway driving. No ping-ponging, and maybe two instances of phantom braking in 5 years. I love it. It'd be fun to try out the newfangled goodies in an AP3 car sometime.

Bruce.

I’m the one with the Tesla Tequila bottles in his name. I’m both an alcoholic and a Teslaholic.

 

[SeminoleFSU]

You got a link to the stats for this claim? Most of the high profile examples I can think of were AP2+ but that is purely anecdotal... much like your claim.

Nope. I just remember every time a Tesla slammed into a stationary object I would look at the year and it would've been an AP1 car.
I also remember:
1. AP2 phantom braking like crazy under every bridge early on. Bridges being a stationary object.
2. AP2 has more powerful compute. More cameras. Color vs B&W cameras (at least on the sides). Longer throw radar. Higher res cameras. Longer throw ultrasonic sensors.
3. MobileEye making comments at the time they split with Tesla that their hardware/tech was being used by Tesla past the point of it's intended use case. I personally think they meant lane changing but who knows maybe they were unhappy about other things like that potential for the car to have trouble with stationary objects? Particularly white objects seemed to give it trouble IIRC

 

[Ostrichsak]

Nope. I just remember every time a Tesla slammed into a stationary object I would look at the year and it would've been an AP1 car.
I also remember:
1. AP2 phantom braking like crazy under every bridge early on. Bridges being a stationary object.
2. AP2 has more powerful compute. More cameras. Color vs B&W cameras (at least on the sides). Longer throw radar. Higher res cameras. Longer throw ultrasonic sensors.
3. MobileEye making comments at the time they split with Tesla that their hardware/tech was being used by Tesla past the point of it's intended use case. I personally think they meant lane changing but who knows maybe they were unhappy about other things like that potential for the car to have trouble with stationary objects? Particularly white objects seemed to give it trouble IIRC

So now you're taking anecdotal information and then layering opinionated guesses on top of that to form a conclusion that you present as undeniable fact. Seems highly scientific.

 

[SeminoleFSU]

So now you're taking anecdotal information and then layering opinionated guesses on top of that to form a conclusion that you present as undeniable fact. Seems highly scientific.

It's a fact the following hardware is better in HW2, HW2.5, and HW3 over HW1:

• The radar is improved. (Longer throw)
• There are more cameras
  
• The cameras even improved from H2.0 to HW2.5
• The ultrasonic sensors improved (longer throw)
• The CPU/GPUs are faster. There's also 2 chips on the boards now that work together to check each other before the car makes a decision (they have to agree)
• The software improved. Far greater number of things being tracked and the ability to forecast the movement of objects over time

These are facts.

However, it is definitely my opinion that with all of the added complexity above, the car is more likely to brake (for better or worse) because it has more data coming in and is measuring more of its surroundings. Therefore, it is more likely to brake (yes, this includes phantom braking). However, that phantom braking as much as it scares the beejeebus out of our spouses (and sometimes us) is in my OPINION more likely to keep us from running into something if we aren't paying attention. It is also my perception (from my memory) of investigating many of the worse (and highly publicized) accidents where AP was engaged that they were AP1 cars. Could my memory be wrong? Yes.

Hopefully that clears it up my position for you.

 

[Ostrichsak]

It's a fact the following hardware is better in HW2, HW2.5, and HW3 over HW1:

• The radar is improved. (Longer throw)
• There are more cameras
  
• The cameras even improved from H2.0 to HW2.5
• The ultrasonic sensors improved (longer throw)
• The CPU/GPUs are faster. There's also 2 chips on the boards now that work together to check each other before the car makes a decision (they have to agree)
• The software improved. Far greater number of things being tracked and the ability to forecast the movement of objects over time

These are facts.

However, it is definitely my opinion that with all of the added complexity above, the car is more likely to brake (for better or worse) because it has more data coming in and is measuring more of its surroundings. Therefore, it is more likely to brake (yes, this includes phantom braking). However, that phantom braking as much as it scares the beejeebus out of our spouses (and sometimes us) is in my OPINION more likely to keep us from running into something if we aren't paying attention. It is also my perception (from my memory) of investigating many of the worse (and highly publicized) accidents where AP was engaged that they were AP1 cars. Could my memory be wrong? Yes.

Hopefully that clears it up my position for you.

Your hardware difference facts aren't what's being questioned here so I'm really not sure why you keep going there.

You can post all the facts you want but it doesn't account for software that makes all of that work and it sure as heck doesn't connect the dots to the claim you made earlier about the AP1 cars being more dangerous as if you had some sort of empirical data to support such a bold claim.

You notice what you want to notice (ever heard of confirmation bias) and acting like the AP1 cars are somehow death traps based entirely on your memory of stuff you once say is a bit of a stretch.

 

[SeminoleFSU]

Your hardware difference facts aren't what's being questioned here so I'm really not sure why you keep going there.

You can post all the facts you want but it doesn't account for software that makes all of that work and it sure as heck doesn't connect the dots to the claim you made earlier about the AP1 cars being more dangerous as if you had some sort of empirical data to support such a bold claim.

You notice what you want to notice (ever heard of confirmation bias) and acting like the AP1 cars are somehow death traps based entirely on your memory of stuff you once say is a bit of a stretch.

LOL. I never said they were death traps. Now who's embellishing? I'm simply stating that the hardware and software has improved- this is a fact that you want to ignore.

I also looked at the model years of nearly all of the highly publicized accidents where AP was engaged and the car ran into a stationary object. From memory, most of them were AP1.. I was particularly interested in this back when these were happening for the very reason we're discussing right now. You can dismiss all those data points if you want. You have your opinion and I have mine. As a previous poster pointed out, I don't have a complete data set on the crashes because I only had the info the news media sensationalized. I'm sure there were many other wrecks with AP engaged they perhaps didn't report... Although the most concerning ones where the cars hit something stationary in front of it they definitely hyped. A few were emergency vehicles so it made it that much more sensational. IIRC, the one where the man was driving on the divided hwy and ran under a tractor trailer decapitating him was AP1 as well as many of the emergency vehicle examples. That system was particularly hit or miss when the car in front was half way out of the lane (AP2 has improved on this dramatically). It also had trouble with stationary objects that were white as they blended with the background. These are all things you can look up for yourself.

 

[Ostrichsak]

LOL. I never said they were death traps. Now who's embellishing? I'm simply stating that the hardware and software has improved- this is a fact that you want to ignore.

I also looked at the model years of nearly all of the highly publicized accidents where AP was engaged and the car ran into a stationary object. From memory, most of them were AP1.. I was particularly interested in this back when these were happening for the very reason we're discussing right now. You can dismiss all those data points if you want. You have your opinion and I have mine. As a previous poster pointed out, I don't have a complete data set on the crashes because I only had the info the news media sensationalized. I'm sure there were many other wrecks with AP engaged they perhaps didn't report... Although the most concerning ones where the cars hit something stationary in front of it they definitely hyped. A few were emergency vehicles so it made it that much more sensational. IIRC, the one where the man was driving on the divided hwy and ran under a tractor trailer decapitating him was AP1 as well as many of the emergency vehicle examples. That system was particularly hit or miss when the car in front was half way out of the lane (AP2 has improved on this dramatically). It also had trouble with stationary objects that were white as they blended with the background. These are all things you can look up for yourself.

I'm using hyperbole because you keep trying to turn this into a hardware debate to deflect from what you originally said:

On the other hand though a lot of the AP accidents where it failed to brake and the Tesla ran into another vehicle on the freeway have been AP1. So the phantom braking yes is annoying but I think with that comes the lower likelihood it will run into a stationary object in the road in front of you

You presented your opinion as fact. If you don't have the data to back up such claims it would be a better approach to not present such things as declarative factual and statements when (by your own admission later on) you really don't know definitively.