Driver engagement while in Autopilot?

[johnbro23]

What will be the acceptable level of driver engagement while in Autopilot? I figure these are the scenarios to consider:

Best case - Not necessary to have eyes on the road, only in ideal circumstances (I.e. no traffic on highway, no merging to worry about). So would be free to read a book, etc
Good case - Periferal eye on the road at all times, so things like reading emails / texting while holding the phone at eye level, aware of surroundings
Worst case - eyes on the road at all times, hands on the wheel, no additional freedom vs regular driving

 

[Stoneymonster]

What will be the acceptable level of driver engagement while in Autopilot? I figure these are the scenarios to consider:

Best case - Not necessary to have eyes on the road, only in ideal circumstances (I.e. no traffic on highway, no merging to worry about). So would be free to read a book, etc
Good case - Periferal eye on the road at all times, so things like reading emails / texting while holding the phone at eye level, aware of surroundings
Worst case - eyes on the road at all times, hands on the wheel, no additional freedom vs regular driving

It will be your "worst case" but with no hands and reduced stress level. Your best and good cases are and should remain illegal until we have full autonomous vehicles.

 

[anxman]

It will be your "worst case" but with no hands and reduced stress level. Your best and good cases are and should remain illegal until we have full autonomous vehicles.

Yup, this. It's like driving with TACC -- we still need to be fully aware and capable of taking over, but the mental overhead requirements shift from active involvement to passive.

 

[dirkhh]

What will be the acceptable level of driver engagement while in Autopilot? I figure these are the scenarios to consider:

Good case - Periferal eye on the road at all times, so things like reading emails / texting while holding the phone at eye level, aware of surroundings
Worst case - eyes on the road at all times, hands on the wheel, no additional freedom vs regular driving

Thank you for playing, please hand over your drivers license at the exit and step away from the car...

 

[ItsNotAboutTheMoney]

Studies clearly indicate that drivers will be less attentive with any driver assistance technology.

So with Autopilot (which for distractoon purposes is mainly just TACC plus Lane Keep) drivers will easily become distracted. However, as much as people often insist otherwise, as long as the tech beats the human on average, we're still better off with it.

 

[electrish]

Studies clearly indicate that drivers will be less attentive with any driver assistance technology.

So with Autopilot (which for distractoon purposes is mainly just TACC plus Lane Keep) drivers will easily become distracted. However, as much as people often insist otherwise, as long as the tech beats the human on average, we're still better off with it.

That's the main argument for lane keeping technology: people become distracted (text, etc.) anyways while operating vehicles, with or without autopilot; technology will now hopefully prevent them from crashing into others while distracted.
However, I agree that technological advances will most certainly breed complacency, with people attempting to perform other crazy tasks in the car, while relying on autopilot to steer the vehicle.

 

[tstafford]

People will do all sorts of things while AP/Lanekeeping is active. Otherwise there wouldn't be over 2000 posts in the thread about about how badly everyone wants the feature. If it was simply to reduce "mental overhead" people wouldn't be clamoring for it.

 

[AnOutsider]

Considering the things I see folks doing now (eye makeup, reading a newspaper) while driving to work in the mornings, the AP capabilities are likely to make things safer regardless.

 

[johnbro23]

That's the main argument for lane keeping technology: people become distracted (text, etc.) anyways while operating vehicles, with or without autopilot; technology will now hopefully prevent them from crashing into others while distracted.
However, I agree that technological advances will most certainly breed complacency, with people attempting to perform other crazy tasks in the car, while relying on autopilot to steer the vehicle.

It should always be a judgement call. Like prob not a good idea to take eyes off the road while in heavy traffic around major cities.. But what about highways in middle of America where each exit is 20 miles apart?

 

[invisik]

My thoughts come to avoiding wildlife on the road... Not sure how detailed the autopilot will be. Like debris in the road like shredded semi truck tire pieces and potholes. These areas not typically life threatening but certainly damaging to the vehicle and possibly could lead to loss of control.

No matter the technology level, eyes on the road are probably a good idea.......

-m

 

[sahaight]

Before I got my Tesla I was a somewhat aggressive driver. My first day using TACC though, I was quite happy to let it take over the speed control and, as a result, my driving was much less aggressive. I'm guessing lane keeping and automatic lane changes will only enhance the trend. If that's true of everyone, then the roads will be much safer.

 

[Stoneymonster]

Before I got my Tesla I was a somewhat aggressive driver. My first day using TACC though, I was quite happy to let it take over the speed control and, as a result, my driving was much less aggressive. I'm guessing lane keeping and automatic lane changes will only enhance the trend. If that's true of everyone, then the roads will be much safer.

I found to be that true as well.

 

[Max*]

What will be the acceptable level of driver engagement while in Autopilot? I figure these are the scenarios to consider:

Best case - Not necessary to have eyes on the road, only in ideal circumstances (I.e. no traffic on highway, no merging to worry about). So would be free to read a book, etc
Good case - Periferal eye on the road at all times, so things like reading emails / texting while holding the phone at eye level, aware of surroundings
Worst case - eyes on the road at all times, hands on the wheel, no additional freedom vs regular driving

Autonomous driving != autopilot.

You wont get best case with the current gen Model S. I want to say it's 3-5 years out, but appear MobilEye is way ahead of schedule, so you might see it sooner.

If you crash, it's not the cars fault. It's yours. Also if AP disables (the way TACC randomly disables in certain situations), you better be aware of your surroundings.

 

[Khatsalano]

Autopilot is more like flying a plane with autopilot. The human pilot doesn't get to go read a book or check Facebook or even use the bathroom because autopilot is on. The pilot's own life, the lives of the passengers, and the lives of those below, are all at stake.

This is exactly the same situation as when you're driving with autopilot on in a car.

 

[ggr]

Autopilot is more like flying a plane with autopilot. The human pilot doesn't get to go read a book or check Facebook or even use the bathroom because autopilot is on. The pilot's own life, the lives of the passengers, and the lives of those below, are all at stake.

This is exactly the same situation as when you're driving with autopilot on in a car.

True. But pilots have training to deal with autopilots. Who is going to train all the drivers out there? A paragraph in the user manual saying "Warning: horrible death and dismemberment may be a consequence of not paying attention while autopilot is engaged" is not sufficient.

 

[Max*]

True. But pilots have training to deal with autopilots. Who is going to train all the drivers out there? A paragraph in the user manual saying "Warning: horrible death and dismemberment may be a consequence of not paying attention while autopilot is engaged" is not sufficient.

Adaptive cruise control (Tesla's version: TACC) has been out for over 15 years. There is no training involved, yet people use it all the time, most without dire consequences.

 

[dirkhh]

Adaptive cruise control (Tesla's version: TACC) has been out for over 15 years. There is no training involved, yet people use it all the time, most without dire consequences.

Lane holding and automated lane changes have not been out for over 15 years. There are quite a few cars that offer lane holding according to Lane departure warning system - Wikipedia, the free encyclopedia - the oldest ones are from 2013 (so two years).
And I don't think any of them do the whole lane change thing.

 

[Max*]

Lane holding and automated lane changes have not been out for over 15 years. There are quite a few cars that offer lane holding according to Lane departure warning system - Wikipedia, the free encyclopedia - the oldest ones are from 2013 (so two years).
And I don't think any of them do the whole lane change thing.

Umn... ok?

Maybe my point wasn't clear. When we went form pulsing brakes to ABS, no training was required. When airbags were introduced, no training was required. When we went from no automation (Level 0), to Level 1 (ACC), no training was required. When we're going from Level 1 to Level 2 (ACC+Lane Keep. yes, I'm aware that Tesla isn't first to get there) why would training be required?

 

[dirkhh]

Umn... ok?

Maybe my point wasn't clear. When we went form pulsing brakes to ABS, no training was required. When airbags were introduced, no training was required. When we went from no automation (Level 0), to Level 1 (ACC), no training was required. When we're going from Level 1 to Level 2 (ACC+Lane Keep. yes, I'm aware that Tesla isn't first to get there) why would training be required?

Sorry, I made assumptions about your point from context of other posts. You are of course correct. Adaptive cruise controls have been available for a very long time. No training required (yet my wife refused to use it in our 10 year old Sienna).
My point was slightly different... lane hold and lane change are really brand new. And arguably they might require some training in order to be used safely.
We had the people here in the forums with unrealistic expectations what TACC can do and we had posts about a lot of near misses and about at least one (two?) crashe(s) that I can remember where the poster blamed TACC for not preventing the crash (instead of just driving the car as they are supposed to). And I'm thinking that with full AP there is a risk that we'll see more crashes where people use it with their brains turned off and then blame the technology for the fact that they stopped driving their car and thought they were just a passenger who could be on Facebook or read a book...

 

[Max*]

Sorry, I made assumptions about your point from context of other posts. You are of course correct. Adaptive cruise controls have been available for a very long time. No training required (yet my wife refused to use it in our 10 year old Sienna).
My point was slightly different... lane hold and lane change are really brand new. And arguably they might require some training in order to be used safely.
We had the people here in the forums with unrealistic expectations what TACC can do and we had posts about a lot of near misses and about at least one (two?) crashe(s) that I can remember where the poster blamed TACC for not preventing the crash (instead of just driving the car as they are supposed to). And I'm thinking that with full AP there is a risk that we'll see more crashes where people use it with their brains turned off and then blame the technology for the fact that they stopped driving their car and thought they were just a passenger who could be on Facebook or read a book...

There was a good podcast (I think it was NPR, planet money) and they were making an argument similar to yours, that there's a reason some car manufacturers should skip anything between Person driving and full autonomous. And there's a reason full autonomous should not have steering wheels. ==> People do stupid things.

I agree that AP will lead to more careless driving. The question is, will the reduction in accidents compensate for the small fraction of people who are driving more careless? I would expect it to be so, but time will tell.

WRT unreasonable expectations: There were even some comments about people saying something along the lines of "AP will suck, since it'll follow the navs horrible route guidance". Sigh... I'm sure there will be lots of threads about people misusing AP and either getting into accidents or nearly getting into accidents.

I'm not aware of any training that Mercedes required for AP+Lane keep. So I don't expect Tesla to do it either. Would it be a good thing if they offered OPTIONAL training? Sure. People suck at driving, so any additional training will be beneficial to them. Do I think Tesla will offer anything besides a blurb saying "pay attention, you're responsible, this isn't autonomous driving"? I doubt it.