Does using AP increase safety?

[green1]

There are also a number of accidents where Tesla drivers have fallen asleep. Maybe its the combination of smoothness plus quietness. Or maybe they're just old.

At least in the Tesla if they fall asleep they have a chance. I've responded to the aftermath of people who have fallen asleep behind the wheel of other vehicles. It's not pretty.

 

[EinSV]

It is only remarkable if we know with equal or greater certainty that 200M miles of equivalent non-AP driving would have had a higher
death rate. Does anyone have data to that effect?

The publicly available data suggest that not a single person has been killed with AP when is used as instructed. I find that a truly remarkable statistic.

It will be interesting to see the data in six months or a year with the improvements in the v8.0 system. If Elon's predictions are correct, the data should show a significant decrease in serious injuries and fatalities (i.e., 2X-3X) when driving with Autopilot compared to vehicles with no Autopilot.

I think it is important to put the incredibly fast developing advances in AP in context. Let's assume for the sake of argument that at this moment AP is only "just as safe," not safer, than a driver without AP. Can there be any doubt that with machine and fleet learning that on average driving with AP will soon be much safer than driving without it?

Autopilot is not just a major advance in automotive safety, but a major advance in public health. The leading cause of death for people aged 1-44 is "unintentional injury," primarily motor vehicle accidents. http://www.cdc.gov/injury/images/lc-charts/leading_causes_of_death_age_group_2014_1050w760h.gif

I expect that within a few short years Autopilot (and comparable technologies, if adopted by other automakers), will save an incredible number of lives.

 

[RogerHScott]

The publicly available data suggest that not a single person has been killed with AP when is used as instructed. I find that a truly remarkable statistic.

If I told you that not a single person has been killed without AP under equivalent circumstances, would you find that equally remarkable?
I don't know that's true, but neither do I know it is not.

 

[S4WRXTTCS]

If however you're unsafe driving while using AP, you would be even MORE unsafe driving without it.

This is part of human nature. The popular example is when a bicyclist isn't wearing a helmet the drivers on the road give them more room than they would if the bicyclist was wearing their helmet. It kinda defeats the purpose of wearing the helmet in a way.

So of course people who text will likely text more with AP than without. Someone like me who absolutely hates people who text and drive would never do so without AP, but on occasion if something pressing is going on I might with AP in specific road situations. My safety margin is pretty high, and still remains high with AP. But, I probably don't gain anything with AP. I think it's pretty much a wash not counting for the safety elements of AP.

With AP I think we have to be realistic when it comes to human nature. That sometimes when trying to make something safe they end up ruining the safety margin we gave them.

 

[RogerHScott]

Pardon me while I jump on my hobby horse here...
This may not apply to the typical Tesla driver today, given that they are in general older and (maybe?) wiser than a lot of the young
people involved in many fatal accidents, but if Tesla wanted to do one thing to reduce driving fatalities it would be to disallow driving
over, say, 20mph when either front passenger is not wearing a seat belt. Nannyism? You betcha. Save countless lives? That, too.
[dismount hobby horse]

 

[S4WRXTTCS]

The publicly available data suggest that not a single person has been killed with AP when is used as instructed. I find that a truly remarkable statistic.

But, does that really mean anything?

It definitely has truck lust and even if you're paying attention I could see it causing an accident due to over-correction. If they don't eliminate this behavior it certainly stands a fair chance someone will die as a result.
It also has other behaviors like suddenly slowing down that could cause an accident.

There have been at least a few accidents where there was a control hand off issue. You can say they weren't using it correctly, but is that their fault or a UI/IX issue?

I think what we're seeing is mixed results, but we're learning a lot.

 

[RogerHScott]

The popular example is when a bicyclist isn't wearing their helmet the drivers on the road give them more room than they would if the bicyclist was wearing their helmet. It kinda defeats the purpose of wearing the helmet in a way.

What???

 

[S4WRXTTCS]

Pardon me while I jump on my hobby horse here...
This may not apply to the typical Tesla driver today, given that they are in general older and (maybe?) wiser than a lot of the young
people involved in many fatal accidents, but if Tesla wanted to do one thing to reduce driving fatalities it would be to disallow driving
over, say, 20mph when either front passenger is not wearing a seat belt. Nannyism? You betcha. Save countless lives? That, too.
[dismount hobby horse]

Much to my dismay it turned off AP when I unbuckled my seatbelt when I tried taking off my coat.

The number one purpose of AP for me was to hold the wheel for me on occasion, and it wouldn't even do that.

 

[S4WRXTTCS]

What???

Strange but True: Helmets Attract Cars to Cyclists

 

[green1]

So the argument then is that we should ban helmets, seatbelts, and all other safety features to make us more safe?

Wow... no wonder we can't have nice things... As I've said before, if this ridiculous society existed in the past, the wheel never would have been invented.

 

[EinSV]

If I told you that not a single person has been killed without AP under equivalent circumstances, would you find that equally remarkable?
I don't know that's true, but neither do I know it is not.

Let me respond this way. I don't understand the animosity that is expressed by many people to what is obviously technology that will save tens of thousands of lives of people in the prime of life and could lead to automobiles no longer being the leading cause of death among kids, teenagers and their parents.

To my knowledge, this technology is too young to have statistically significant data available to prove that lives have already been saved. But having said that I think that the first year of AP in the wild has been an extraordinary success, and will within the next twelve months will result in a dramatic and statistically provable increase in driving safety.

 

[S4WRXTTCS]

So the argument then is that we should ban helmets, seatbelts, and all other safety features to make us more safe?

Wow... no wonder we can't have nice things... As I've said before, if this ridiculous society existed in the past, the wheel never would have been invented.

There is no argument we should ban any of it.

All that's being said is to account for it.

I still wear my helmet when I bike, and I do NOT want to replace my airbag with a dagger even if some studies have shown that driving would be really safe if we all did that.

 

[RogerHScott]

Let me respond this way. I don't understand the animosity that is expressed by many people to what is obviously technology that will save tens of thousands of lives of people in the prime of life and could lead to automobiles no longer being the leading cause of death among kids, teenagers and their parents.

To my knowledge, this technology is too young to have statistically significant data available to prove that lives have already been saved. But having said that I think that the first year of AP in the wild has been an extraordinary success, and will within the next twelve months will result in a dramatic and statistically provable increase in driving safety.

That may be true, but the original question was "Do Tesla's claims about AP safety hold water?". There's no animosity toward the
technology (at least, not on my part), but I'm not a fan of unsubstantiated and quite possibly misleading claims. Especially when
there's no real need for them.

 

[EinSV]

But, does that really mean anything?

It definitely has truck lust and even if you're paying attention I could see it causing an accident due to over-correction. If they don't eliminate this behavior it certainly stands a fair chance someone will die as a result.
It also has other behaviors like suddenly slowing down that could cause an accident.

There have been at least a few accidents where there was a control hand off issue. You can say they weren't using it correctly, but is that their fault or a UI/IX issue?

I think what we're seeing is mixed results, but we're learning a lot.

Let me put it this way. In the very first year this technology is in widespread use, and even viewed in a pessimistic light, driving with AP is at least as safe as driving without AP.

Given how fast the technology is improving, is there any doubt whatsoever that by the end of year 2 driving with AP will be safer than driving without?

I say not a chance.

 

[EinSV]

That may be true, but the original question was "Do Tesla's claims about AP safety hold water?". There's no animosity toward the
technology (at least, not on my part), but I'm not a fan of unsubstantiated and quite possibly misleading claims. Especially when
there's no real need for them.

The thread title is directed to the technology (at least the way I read it):

"Does using AP increase safety."

I think the answer is an unequivocal "yes." And I predict a year from now it will be silly to even ask the question.

 

[S4WRXTTCS]

Let me put it this way. In the very first year this technology is in widespread use, and even viewed in a pessimistic light, driving with AP is at least as safe as driving without AP.

Given how fast the technology is improving, is there any doubt whatsoever that by the end of year 2 driving with AP will be safer than driving without?

I say not a chance.

The enhancements that are being made with V8.0/V8.1 improve passive driving as much as it improves AP.

So I don't think AP driving will be safer than non-AP driving on an identical car.

It's going to be hard for a robot to compete with a human+robot combination. The human can better account for the actions of other humans. The robot is better at paying attention.

 

[Hota]

They also already introduced limits to the speed it will operate at which can cause the car to unexpectedly slow large amounts in the middle of a freeway, and they've introduced nags at meaningless times which condition drivers to ignore the nags.

On the freeway AP is limited to 90mph (at least in the US) regardless of posted speed limits. If you try accelerating past this while AP is engaged, you get huge warning tones and AP cuts off entirely. Also with the latest firmware (2.32.23) there's no nag at specified times. I drove today and spent a stretch of 15 minutes hands off time before I got a warning.

Speak for yourself.
I am LESS distraced with AP than without, I can now focus my FULL attention on what's happening around me, and be in a much better position to make the necessary changes.

Those "1 second at a time" glances at a smartphone? They are ZERO second glances. My phone stays in my pocket, not only would it be illegal for me to pull it out, it's also just plain stupid.

When I use AP (on average 90miles/day), I spend a lot of time looking at other drivers around me observing what they are doing and how they behave. I would say in the few seconds that I spend observing any neighboring driver, roughly 5-10% of them are periodically glancing down in their laps presumably interacting with smart phones. If you give them AP capability, the system will almost certainly give them the false sense of security to be even more distracted even while stressing that it's a driver assistance feature, in beta, etc. That's a lot of people that bad things can happen to.

You can't design systems for the ideal perfect driver - they have to be designed to the way most people will interact with it.

 

[green1]

On the freeway AP is limited to 90mph (at least in the US) regardless of posted speed limits. If you try accelerating past this while AP is engaged, you get huge warning tones and AP cuts off entirely. Also with the latest firmware (2.32.23) there's no nag at specified times. I drove today and spent a stretch of 15 minutes hands off time before I got a warning.

That's based on the class of road. You get different timed nags for different classes of road currently. On 8.0 you will get a nag every 1 minute at freeway speeds.

When I use AP (on average 90miles/day), I spend a lot of time looking at other drivers around me observing what they are doing and how they behave. I would say in the few seconds that I spend observing any neighboring driver, roughly 5-10% of them are periodically glancing down in their laps presumably interacting with smart phones. If you give them AP capability, the system will almost certainly give them the false sense of security to be even more distracted even while stressing that it's a driver assistance feature, in beta, etc. That's a lot of people that bad things can happen to.

Based on what you say, that 5-10% of people are extremely unsafe now, and would likely be much safer if autopilot was there to cover for their inattentiveness. Of course they'd be even safer if they put down the phone.
I see that not as a weakness in AP, but an opportunity for AP to save lives.

You can't design systems for the ideal perfect driver - they have to be designed to the way most people will interact with it.

Which is it, "5-10%" or "most"? I'd argue that you can't put 90-95% of people's lives at risk just to protect the 5-10% of people who are being stupid.

 

[EinSV]

The enhancements that are being made with V8.0/V8.1 improve passive driving as much as it improves AP.

So I don't think AP driving will be safer than non-AP driving on an identical car.

Of course it took an AP accident for Tesla to incorporate the changes that enhances the AEB (that works whether AP is enabled or not).

It's going to be hard for a robot to compete with a human+robot combination.

It 2-5 years these cars are going to be really hard to crash.

I think the final point you make is the key one IMO and I agree with you.

Whether you consider that technology "Autopilot" or enhanced AEB is just semantics as far as I'm concerned.

 

[Hota]

The publicly available data suggest that not a single person has been killed with AP when is used as instructed. I find that a truly remarkable statistic.

That's a faulty way to look at the data. It needs to be looked at under all circumstances in which AP was used because that's the way it WILL be used in reality.

As an extreme example, I could design a car with the brake and gas pedals reversed. Anytime someone got into an accident by hitting the wrong pedal, I could blame the driver for not using the vehicle as instructed and chalk it up to driver error. We know that's not the whole story though.

I love my AP and would be really upset if it were diminished but any discussions on the safety of AP needs to consider the Autopilot + human interface/behavior as a whole and not just look at the computer system.