Self-Driving Car Kills Pedestrian in AZ

[GooooMarquette]

I know they will eventually become a common thing, but this is one example of why they will likely take longer than some think.  And yes - I know that distracted drivers probably killed 50 pedestrians today...but IMHO people will always be more afraid of the new danger than the one they're used to....

http://money.cnn.com/2018/03/19/technology/uber-autonomous-car-fatal-crash/index.html

[Benny B]

I saw this earlier, and all I can think is that there's going to be more to this story.

Even though the car was in "autonomous" mode, there was still a person in the car, and I believe by law, if there's a person in the car, that person must still be able to take full control of the vehicle at any time.

So this wasn't just a runaway vehicle plowing down pedestrians... something was going on that both the vehicle and the driver did not heed any warnings (or were not able to react in time).

[mu_hilltopper]

2018 Pedestrian Deaths caused by humans on their phone while driving: 472
2018 Pedestrian Deaths caused by robocar: 1

I welcome our robot overlords.

[MU82]

2018 Pedestrian Deaths caused by humans on their phone while driving: 472
2018 Pedestrian Deaths caused by robocar: 1

I welcome our robot overlords.

I understand what you're saying, but one has to go by percentages to do an apples/apples comparison, no?

What was the total number of cars driven by humans while they were using their phones? What was the total of robocars on the road?

If there were 500K cars on the road last year in which a human used the phone while driving (and that number is probably low), and 472 pedestrian died in accidents involving humans who were on the phone while driving, that is 1 pedestrian death per 1,059 on-phone-while-driving incidents.

If there were only 200 robocars on the road during the exact same span (and I have no idea if there were 1/10th that many or 10x that many), and 1 pedestrian died in an accident involving a robocar, that is 1 pedestrian death per 200 robocar trips. Or whatever the numbers are.

I am not making any kind of opinion on the goodness or badness of robocars. Just stating that the 472 and 1 figures weren't really an accurate measure.

Need something like ePD%! I'm all about advanced stats!!!

[mu_hilltopper]

You know what makes the 472 figure really screwy?  I made it up.

I have zero doubt that robocars are safer per million miles driven.  That is likely true in 2018, and assuredly in 2028.

I'm not sure a robocar is in my future, but I sure want all you other schlubs to have one.  I want all the 16-22 year old idiots to have one.  And the 75 year olds.  Not to mention Nancy going to work putting makeup on while driving.  And Steven, on his way to work responding to his important emails.

Hey, all these test cars should have cameras rolling constantly, right?  Sure would love to see the footage of this accident.  (Minus the grisly bits.)

[WarriorDad]

2018 Pedestrian Deaths caused by humans on their phone while driving: 472
2018 Pedestrian Deaths caused by robocar: 1

I welcome our robot overlords.

How many robocars out there?  1000?  How many humans driving cars?  250 million?

No doubt robocars will be safer in the long run, but one item I have heard is that the cars / technology have trouble at times anticipating the unpredictable where a human can use human instincts to react.   Maybe a better way to put it, humans will commit more accidents in situations where the auto car will commit very few, the mundane situations and normal rules of the road type situations.  But the non-normal situation which is so difficult to anticipate or program for is the achilles heel at this point.

[Hards Alumni]

Not surprised it was Uber software.  They have spent far less time developing than Google.  They're rushing to market because they're sunk if they don't.

[Spotcheck Billy]

You know what makes the 472 figure really screwy?  I made it up.

I have zero doubt that robocars are safer per million miles driven.  That is likely true in 2018, and assuredly in 2028.

I'm not sure a robocar is in my future, but I sure want all you other schlubs to have one.  I want all the 16-22 year old idiots to have one.  And the 75 year olds.  Not to mention Nancy going to work putting makeup on while driving.  And Steven, on his way to work responding to his important emails.

Hey, all these test cars should have cameras rolling constantly, right?  Sure would love to see the footage of this accident.  (Minus the grisly bits.)

But will the drive-thru at Arby's need redesigning to accomodate robocars?

[ZiggysFryBoy]

But will the drive-thru at Arby's need redesigning to accomodate robocars?

Arby's will have a fleet of self driving BnC delivery cars.  Soon.

[Golden Avalanche]

Jaywalking, no matter the hour of the day, can be deadly.

[GB Warrior]

Arby's will have a fleet of self driving BnC delivery cars.  Soon.

I need another way for Arbys to destroy me.

[MU82]

You know what makes the 472 figure really screwy?  I made it up.

I have zero doubt that robocars are safer per million miles driven.  That is likely true in 2018, and assuredly in 2028.

I'm not sure a robocar is in my future, but I sure want all you other schlubs to have one.  I want all the 16-22 year old idiots to have one.  And the 75 year olds.  Not to mention Nancy going to work putting makeup on while driving.  And Steven, on his way to work responding to his important emails.

Hey, all these test cars should have cameras rolling constantly, right?  Sure would love to see the footage of this accident.  (Minus the grisly bits.)

This schlub resembles that remark! I haven't driven while putting on my makeup in over a year!!!

[GooooMarquette]

I'm not sure a robocar is in my future, but I sure want all you other schlubs to have one.  I want all the 16-22 year old idiots to have one.  And the 75 year olds.  Not to mention Nancy going to work putting makeup on while driving.  And Steven, on his way to work responding to his important emails.

I was once on an early morning bike ride when I saw a high school or college age girl driving (likely to school) looking down at an open textbook propped up on the steering wheel.

[MU82]

I was once on an early morning bike ride when I saw a high school or college age girl driving (likely to school) looking down at an open textbook propped up on the steering wheel.

Sounds like a fine, studious young lady. Probably class valedictorian.

It's called multi-tasking!

[g0lden3agle]

You know what makes the 472 figure really screwy?  I made it up.

I have zero doubt that robocars are safer per million miles driven.  That is likely true in 2018, and assuredly in 2028.

I'm not sure a robocar is in my future, but I sure want all you other schlubs to have one.  I want all the 16-22 year old idiots to have one.  And the 75 year olds.  Not to mention Nancy going to work putting makeup on while driving.  And Steven, on his way to work responding to his important emails.

Hey, all these test cars should have cameras rolling constantly, right?  Sure would love to see the footage of this accident.  (Minus the grisly bits.)

This is actually one of the problems with defending autonomous:  if you look at deaths per miles driven it will be hard for autonomous vehicles to meet or beat the current status quo of human drivers.  How much of an increase is something we're willing to accept and plan that autonomous can meet or beat that metric in the future?

[StillAWarrior]

This is actually one of the problems with defending autonomous:  if you look at deaths per miles driven it will be hard for autonomous vehicles to meet or beat the current status quo of human drivers. How much of an increase is something we're willing to accept and plan that autonomous can meet or beat that metric in the future?

I'm curious why you think that.  Actually, I think the opposite is true, and that's one of the main reasons I think autonomous will come faster than many are expecting.  Human beings are really inattentive and pretty horrible drivers.

[MU82]

I'm curious why you think that.  Actually, I think the opposite is true, and that's one of the main reasons I think autonomous will come faster than many are expecting.  Human beings are really inattentive and pretty horrible drivers.

Every human other than me, that is.

One of the fun things about us humans is that we all think the other person is the bad driver.

My 92-year-old father-in-law was still driving until just a couple of years ago, and he was convinced he was the only person who knew how to drive. Whenever he saw an accident, he'd say something like, "I guess dat lady screwed up." Didn't matter if it was a man or a woman, he always assumed it was "dat lady." (Dat's right: He's a sout-side Chicagoan.)

[StillAWarrior]

Every human other than me, that is.

One of the fun things about us humans is that we all think the other person is the bad driver.

My 92-year-old father-in-law was still driving until just a couple of years ago, and he was convinced he was the only person who knew how to drive. Whenever he saw an accident, he'd say something like, "I guess dat lady screwed up." Didn't matter if it was a man or a woman, he always assumed it was "dat lady." (Dat's right: He's a sout-side Chicagoan.)

I think I'm a pretty good driver.  I also think that I get distracted by my phone, passengers, my radio, billboards, nice cars, buildings, pedestrians, etc., etc., etc.  Honestly, I don't think the bar is particularly high for autonomous cars to surpass the safety of human drivers.

[mu_hilltopper]

I'm curious why you think that.  Actually, I think the opposite is true, and that's one of the main reasons I think autonomous will come faster than many are expecting.  Human beings are really inattentive and pretty horrible drivers.

+1.  Can you explain, g0ldeneagle?

[g0lden3agle]

I'm curious why you think that.  Actually, I think the opposite is true, and that's one of the main reasons I think autonomous will come faster than many are expecting.  Human beings are really inattentive and pretty horrible drivers.

My thoughts stem from a keynote I sat in late last year at Ford where they brought up a graph very similar to the one wiki has here: https://en.wikipedia.org/wiki/Transportation_safety_in_the_United_States#/media/File:US_traffic_deaths_per_VMT,_VMT,_per_capita,_and_total_annual_deaths.png

And the speakers, two professors at UofM, were very frank in the fact their opinion that we cannot expect autonomous to be rolled out to the masses aggressively and meet or beat the deaths per millions/billions miles driven right off the bat.

Maybe strictly looking at deaths is not fair, and there is some other metric capturing all non-fatal crashes as well that would have a larger opportunity for improvement that autonomous vehicles can go after.

[Pakuni]

Are autonomous vehicles safer?
The correct answer is, we have no idea.

The use of driverless cars has been tiny compared to those with human drivers, and almost exclusively in urban areas of western states under ideal driving conditions (i.e. no blizzards, downpours, narrow country roads shared by deer and farm equipment, etc.).
Even with those conditions, the rate of fatalities is, in fact, higher with driverless cars vs those with drivers.
But, in reality, the circumstances at this point don't allow for anything close to an accurate comparison.

[StillAWarrior]

My thoughts stem from a keynote I sat in late last year at Ford where they brought up a graph very similar to the one wiki has here:

And the speakers, two professors at UofM, were very frank in the fact their opinion that we cannot expect autonomous to be rolled out to the masses aggressively and meet or beat the deaths per millions/billions miles driven right off the bat.

Maybe strictly looking at deaths is not fair, and there is some other metric capturing all non-fatal crashes as well that would have a larger opportunity for improvement that autonomous vehicles can go after.

Thank you.  I'm curious why they think so.  I'm not arguing that they're wrong...what the hell do I know?!  I'm just curious what their opinion is based on.  Obviously, there must be an awful lot of deaths that are caused by reasons other than inattentive (e.g., distracted, sleeping, drunk, etc.) drivers because I'd think those would clean up quite a bit.  I haven't read much about how well the autonomous will handle challenging weather conditions.

[MU82]

Are autonomous vehicles safer?
The correct answer is, we have no idea.

The use of driverless cars has been tiny compared to those with human drivers, and almost exclusively in urban areas of western states under ideal driving conditions (i.e. no blizzards, downpours, narrow country roads shared by deer and farm equipment, etc.).
Even with those conditions, the rate of fatalities is, in fact, higher with driverless cars vs those with drivers.
But, in reality, the circumstances at this point don't allow for anything close to an accurate comparison.

Yep yep.

Facts are our friends.

[Chili]

I would look at flying as an example. If anyone knows any pilots out there please ask them this but in my hundreds of thousands miles flown almost all of that has been done by an autopilot. Even in bad weather now days pilots don't do very much but usually help with taxi, takeoff and landing. I am guessing that this revolution scared people back in the day but now days it's hardly given a thought to.

I am thinking now that the full adaptive cruise technology becomes more widely available more people will get exposed to computer driving. Right now it's primarily in luxury cars but it will be more available in about 2-3 years. Once people get a taste of it and see it's not the boogie man - it will accelerate at a multiplier pace.

[mu_hilltopper]

I have a feeling I'd find robocars to be very .. Sunday Driver. 

I mean .. I'm no lead foot, but I do keep to whichever lane is fastest, trying to go +10 of the speed limit.  (Got one ticket on my record in 1995.)

My last car had "adaptive cruise" which kept me XX feet behind the car in front of me, and I freaking hated it.    First .. the gap was likely very safe, but so big that people would constantly cut in front of you -- which means suddenly the cruise control would slow to get back to the proper gap.    Second, say you were going 80 and another car passed you and got in your lane going faster .. again, the cruise would slow you down momentarily, even though a human knows .. the guy who passed you is going to create a new gap.

I guess I figure that robocars will be very conservative drivers, always with the proper safety distances, always at the speed limit (yes?), never intuiting the 'faster' lane == Sunday Driver, just getting you safely from A to B.