Uber ~ Automated Driving Death

[oberkc]

And...sure a cool topic.  I suspect this conversation is going on all around the country and the world.  The question about car automation is much more complicated than the relative merits of the driver.  So many factors, laws, perceptions, confidence, mixed use, liability, preferences, enjoyment of driving (at least by some), etc....   I believe the easy part is the technology.

[Teken]

1 minute ago, apostolakisl said:

Oh dear.

Look, the thing is, the computer is going to whoop that pants off you and any other driver the world has ever known.  To deny the technology is to deny that a computer could ever win a trivia game like Jeapordy. . . as I'm sure a whole bunch of people were doing the day before Watson whooped those poor humans.  This is the Paul Bunyon story here and you're weilding the axe.  Paul Bunyon was a very proud woodsman.  The proud driver's of the world will being putting down their axes soon, my guess, 20 years from now all cars will be auto-pilot.  And I promise you, there will be a lot fewer auto related accidents/injuries in the world.

If you and I are still here conversing I'll take that bet.

I formally state that without dedicated infrastructure, rules, laws, and public education. This will not be safer than a human assuming they aren't focused on eating a salad while driving!  I'll bet you $100.00 USD that this technology will start a new era of problems never seen before. 

[oberkc]

54 minutes ago, Teken said:

If you and I are still here conversing I'll take that bet.

I formally state that without dedicated infrastructure, rules, laws, and public education. This will not be safer than a human assuming they aren't focused on eating a salad while driving!  I'll bet you $100.00 USD that this technology will start a new era of problems never seen before. 

I think you are absolutely right about the unintended consequences.  I can already see kids jumping in front of cars now, warching the reaction of the vehicles.  Other drivers will cut them off, counting on the overcautios programming to keep all safe.  Yes, we will find all sorts of fun things with these new vehicles.

[Teken]

3 minutes ago, oberkc said:

I think you are absolutely right about the unintended consequences.  I can already see kids jumping in front of cars now, warching the reaction of the vehicles.  Other drivers will cut them off, counting on the overcautios programming to keep all safe.  Yes, we will find all sorts of fun things with these new vehicles.

All joking aside I do hope this technology becomes common place and safe for everyone. The freedom a person who has mobility issues can't be over stated when speaking about ones independence. 

I'll leave this final thought for the day as *lets see what happens in 1, 2, years* in this field.

[asbril]

10 hours ago, larryllix said:

I have to agree with this. There is something inherently wrong with this idea, at this point in development.

The idea of the system is allow drivers to not be attentive, or be redundant, and yet as the reports linked show the system can disengage at any time when a discrepancy or error is found requiring an attentive human to take over. This requires a100% of the time attentive human. It also admits that the automatic driving system makes many mistakes. It doesn't matter if the system is correct 99.99999% of the time. This allegedly rare mistake, and the conceptual error,  may be the blame for taking this woman's life. It's not good enough. I haven't seen the video yet but from the video observers reports humans would nt have been able to avoid the accident.

It has to be remembered that most mechanical devices do NOT have the dynamic range of light levels the human eye has, so the video may not be an accurate representation that an attentive, in situ, driver would have seen. 

I have to side with Teken here at this point in the technology. Until the next levels are accomplished in the AI, the concept is flawed. Trouble is, how do we get past this level without high human and property risk?

Of course I agree that at this stage the concept is flawed. But to make progress we need to go through testing and this accident has not proved to me that we should not continue with the testing. However, they should come up with a method to keep back-up drivers focussed. 

[larryllix]

4 hours ago, lilyoyo1 said:

https://youtu.be/bp9KBrH8H04

Great! They can't always synchronise video and sound, and yet they can allow a 4,000 pound killing machine loose on it's own recognisance?

[larryllix]

2 hours ago, Teken said:

All joking aside I do hope this technology becomes common place and safe for everyone. The freedom a person who has mobility issues can't be over stated when speaking about ones independence. 

I'll leave this final thought for the day as *lets see what happens in 1, 2, years* in this field.

My conclusion about this is that the technology cannot be half-way. You can't rely on a machine that expects a human to rescue it's  bad decisions 0.000001% of time but  allows him to sleep the rest of the time. Humans do not react that fast, especially after a high voltage shock they should get.

 

The risk based concession analogy is absolute BS. Everything has risk but that doesn't make it OK to cross the street blindfolded, or work around dangerous machinery without human protection guards.   People getting hurt or killed "all the time", doesn't make it OK or forgivable.   This is a usual defence of the lazy and safety ignorant  thinkers, until it happens to them or their loved ones, then they scream "Regulations" and "why?". Allowing this is violating the prime directive of  most AI thinking groups.

 

As stated in previous posts, This is the future and resistance is futile It is going to happen.  Unfortunately with some possible attitudes the cost of people's lives will be dear.

[oberkc]

https://www.cnet.com/news/was-ubers-driverless-car-crash-avoidable-some-experts-say-the-self-driving-car-should-have-braked/

[mitchmitchell]

My conclusion about this is that the technology cannot be half-way. You can't rely on a machine that expects a human to rescue it's  bad decisions 0.000001% of time but  allows him to sleep the rest of the time. Humans do not react that fast, especially after a high voltage shock they should get.

 

The risk based concession analogy is absolute BS. Everything has risk but that doesn't make it OK to cross the street blindfolded, or work around dangerous machinery without human protection guards.   People getting hurt or killed "all the time", doesn't make it OK or forgivable.   This is a usual defence of the lazy and safety ignorant  thinkers, until it happens to them or their loved ones, then they scream "Regulations" and "why?". Allowing this is violating the prime directive of  most AI thinking groups.

 

As stated in previous posts, This is the future and resistance is futile It is going to happen.  Unfortunately with some possible attitudes the cost of people's lives will be dear.

Larry,

 

I think the trick will be to determine when the technology is reliable enough to save more lives than it costs. If you delay deploying it to get that last decimal point of reliability you might just wind up with more deaths due to the delay because of current low reliability of average human drivers.

 

Mitch

 

Sent from my SM-N950U using Tapatalk

 

 

 

[larryllix]

1 hour ago, oberkc said:

https://www.cnet.com/news/was-ubers-driverless-car-crash-avoidable-some-experts-say-the-self-driving-car-should-have-braked/

Thanks oberkc.

Since cameras do not have the dynamic range of a human eye and the street light was blinding the camera view with it's higher illumination level, I think an alert driver would have seen her reflective white shoes. This is not a realistic representation of what a human should have seen.

This appears to be at least three things at fault.

The woman may have been drunk as she closes in a collision path with a fully lit vehicele and appears not to even make any attempt to run or get out of the way. No basic life preservation attempt?

The driver, he likely would  have been able to detect her much earlier than we can in the video. (he may have had his eyes closed??)
......The false  trust of the current technology? I would be sure there would/should be many warnings and instructions or training to operate one of these. No? More diver fail?

The Huber, for failing to detect or react to likely many detection systems. One report said the vehicle slowed from 40 mph to 38 mph at impact. This appears to be a very simple case that was handled in the very basic beginnings of the software and sensor interpretation. Total fail.

[Teken]

Had a few moments to read the new link provided by oberkc which I find interesting. One minute the police are saying neither a human vs machine could have done anything different. Later due to public our cry and pressure these weak aszz cops walk back their comments and say its not determined.

Bottom line ~ stupid human . . .

Same stupid human with zero common sense probably rocking a head set looking down at her toes reading a text and got sch-mucked.

End of story . . .

This same moronic behavior happens in India, China, Russia each and every day of every second. I have zero sympathy for these people and the natural selection process of being eliminated is the end result.  

[asbril]

10 hours ago, Teken said:

Had a few moments to read the new link provided by oberkc which I find interesting. One minute the police are saying neither a human vs machine could have done anything different. Later due to public our cry and pressure these weak aszz cops walk back their comments and say its not determined.

Bottom line ~ stupid human . . .

Same stupid human with zero common sense probably rocking a head set looking down at her toes reading a text and got sch-mucked.

End of story . . .

This same moronic behavior happens in India, China, Russia each and every day of every second. I have zero sympathy for these people and the natural selection process of being eliminated is the end result.  

Teken, I take a less drastic and more practical approach, and I repeat what I have been saying all along this discussion....  We have to learn from experience. In this case I see (at least 2) lessons to be learned. Lesson 1 : Why did the radar/lidar not detect the victim ?  Lesson 2 : What can be done for the back-up driver to maintain focus ?

I bet you that lesson 2 is more complicated than lesson 1. Of course, first things is to move back-up driver's phone out of sight, but it may not be that easy for someone to maintain attention on outside without the physical (steering wheel, gas pedal, brake pedal) need to do so.

I also repeat my contention that so many things we use in life today (from airplanes to can openers...) are safer than before because of learning from accidents. The Uber accident in Arizona is crucial for getting to the next stage.

[dbuss]

A little levity and personal history....

I remember steering my Uncle's pickup truck while he was in the box throwing out hay for the cattle when I was six years old. With the manual throttle and what called Grandma gear there was no need to be able to reach the pedals. I do remember him jumping out of the box and coming up a grabbing the steering wheel as he didn't like the where I was headed. I guess you could say this was an autonomous driven vehicle with human intervention over 60 years ago.

I've been involved in three rearend collisions over the years. Two the fault of others and one my fault. One no injuries, one minor injuries and one life threatening injuries. I think an autonomous driven car would have prevented these crashes because in two cases the drivers were not paying attention and the other was in extremely dense fog.

I must say some of the questions being asked here remind me of the question.....

"Are you still beating your wife?"

No matter how you answer the question you're in trouble. It's best to just not answer the question.





Sent from my XT1650 using Tapatalk

[Mustang65]

Here are a couple articles that may be of interest from WIRED magazine.

 

The Unavoidable Folly of Making Humans Train Self-Driving Cars
https://www.wired.com/story/uber-crash-arizona-human-train-self-driving-cars/

 

https://www.wired.com/2016/06/self-driving-cars-will-power-kill-wont-conscience/

[asbril]

10 minutes ago, Mustang65 said:

https://www.wired.com/story/uber-crash-arizona-human-train-self-driving-cars/

Thanks Mustang56, this is exactly what I implied in my earlier posts.

[Mustang65]

1 hour ago, asbril said:

 Of course, first things is to move back-up driver's phone out of sight, but it may not be that easy for someone to maintain attention on outside without the physical (steering wheel, gas pedal, brake pedal) need to do so.

 

Personally, I believe that the only cell phone operation while the car is in motion should be:

911

Incoming calls (only if intercepted by the cars media center)

Outgoing calls (only if the called number is activated by voice commands) 

Display of cell phones is blanked out along with car display only shows incoming/outgoing voice activated number.

If you want to text, PULL OVER!

LAWS NEEDED

IF you are caught driving with a cell phone in hand.... $500 fine and points against the drivers license!

Just my thoughts

 

 

[asbril]

2 minutes ago, Mustang65 said:

Personally, I believe that the only cell phone operation while the car is in motion should be:

911

Incoming calls (only if intercepted by the cars media center)

Outgoing calls (only if the called number is activated by voice commands) 

Display of cell phones is blanked out along with car display only shows incoming/outgoing voice activated number.

If you want to text, PULL OVER!

LAWS NEEDED

IF you are caught driving with a cell phone in hand.... $500 fine and points against the drivers license!

Just my thoughts

 

 

Good suggestions but how easy to implement ?  How does phone know whether used by driver or passenger ? Though my car has integrated GPS I often use Waze on my Iphone for navigation. If there is a way to make your suggestions work then I am all for it.

[Mustang65]

11 minutes ago, asbril said:

Good suggestions but how easy to implement ?  How does phone know whether used by driver or passenger ? Though my car has integrated GPS I often use Waze on my Iphone for navigation. If there is a way to make your suggestions work then I am all for it.

EASY!!! The government says "No SELF DRIVING CARS" can be approved for sale in the US unless they meet the above conditions. Then the cell phone companies and the car manufactures will HAVE to work together to come up with a method for the above requirements. This of course would be a last ditch effort should the car and cell phone companies decide not to do it.

[paulbates]

Wayz stops providing info and puts up a prompt to make you 'self acknowledge' as being a passenger when it senses motion.

You can be driving when you press it, but then you've left an audit trail showing you were challenged,  physically ignored the prompt and lied. That typically comes down to those that are offended by any suggestion that they might need to rethink. (They're always right and can never be wrong).  In the end, that prompt probably makes Wayz legal team a little happier but changes nothing.

I agree with you that its a problem becoming worse. Not sure why after about 16 years of smart phones that could get to the interent its visibly worse now than ever. People swerving or out of they're lane like they're drunk.. except they're not, and their heads are pointed down operating their phone. Its more frequent now that I'm at the light behind someone, it turns green.. i've counted 10-1000 and their head is still down. I've stopped honking because they over-react to that. I simply pull around in another lane and drive ahead.

Paul

25 minutes ago, Mustang65 said:

Personally, I believe that the only cell phone operation while the car is in motion should be:

911

Incoming calls (only if intercepted by the cars media center)

Outgoing calls (only if the called number is activated by voice commands) 

Display of cell phones is blanked out along with car display only shows incoming/outgoing voice activated number.

If you want to text, PULL OVER!

LAWS NEEDED

IF you are caught driving with a cell phone in hand.... $500 fine and points against the drivers license!

Just my thoughts

 

 

 

[Mustang65]

6 minutes ago, paulbates said:

Wayz stops providing info and puts up a prompt to make you 'self acknowledge' as being a passenger when it senses motion.

 

Paul

 

All they have to do is add a voice response to their program. 

[paulbates]

Just now, Mustang65 said:

All they have to do is add a voice response to their program. 

To your earlier point, if I'm not sure where I am (I travel to different destinations in the US weekly so it happens), I pull over safely somewhere. set up gps to talk me through the rest of the way, then start driving again when I don't need to touch my phone any more.

[mwester]

As I consider this entire problem, I'm struck by the fact that one of the real challenges facing us all is the lack of a proper test environment.  We insist on zero-incidents as our baseline, with the result that we have no "control car" in our experiments to determine the effectiveness of any technology.  Consider a simple scenario -- a fictitious "Ultra-Auto-Drive model 3000" system, claiming to offer improved safety over the earlier "Ultra-Auto-Drive model 2000" it replaces.  How do you know?  The earlier model wasn't permitted to have any incidents either, so how can one claim that the model 3000 is better than the 2000?  Further, how do we know that all the technologies used in the very expensive model 3000 are actually useful -- perhaps one or another of the sensors is contributing nothing to safety, but we can't tell because we cannot run the vehicle *without* the entire battery of sensors and software for fear that we might just cause an incident...

The root cause is our unwillingness to accept the death or injury by an automated vehicle, despite the clear indications that in the long run, death and injury will be lowered.  The immediate threat takes priority over the long-term good.

So, a proposed solution:

I observe that we'll never change human perceptions, so that's a non-starter: we'll never be able to create a proper automated vehicle testing facility (AVTF) in any normal human environment.  Further, I observe that we'll never properly simulate human behavior in the real world -- humans do things that engineers can never fully predict.  So, we need real humans, and we need a dense urban environment, and we need to ensure that the general population doesn't feel threatened by the testing going on.  The popular media has provided us with a means to create this (rather often, actually - there are quite a lot of films based on this premise): we select a less-than-desirable city, surround it with a large chain-link fence and guard towers, and turn it into a high-security prison, which can double as the AVTF that we need.  I'd suggest Chicago -- representative weather, range of urban conditions, centrally-located, and if we keep the politicians on the inside when we fence it off, it'll be ideal.  (The downside is that unless we're doing testing to determine how automated vehicles behave under gunfire, we may have some adverse testing conditions to deal with.)

The clear advantage: when building the new *lower-cost* Ultra-Auto-Drive model 4000 unit, we'd have the ability to inject not just the model-4000-equipped test vehicle onto Lake Shore Drive, but we can also inject the "control vehicle", equipped with the older model -- after a number of test runs through real conditions, we'll clearly see if the elimination of specific sensors actually decreased the effectiveness of the unit, as well as if the newer sensors improved things.

And for those ridden with guilt over using the prison population as test subjects, well, all I can say is better that some drug dealer gets a few broken bones from an incident than someone's child coming home from school...

 

(The above is tongue-in-cheek, and I'm quite interested to see how many actually post enraged responses without reading far enough to read this line.  :-D)

[Teken]

2 hours ago, dbuss said:

A little levity and personal history....

I remember steering my Uncle's pickup truck while he was in the box throwing out hay for the cattle when I was six years old. With the manual throttle and what called Grandma gear there was no need to be able to reach the pedals. I do remember him jumping out of the box and coming up a grabbing the steering wheel as he didn't like the where I was headed. I guess you could say this was an autonomous driven vehicle with human intervention over 60 years ago.

I've been involved in three rearend collisions over the years. Two the fault of others and one my fault. One no injuries, one minor injuries and one life threatening injuries. I think an autonomous driven car would have prevented these crashes because in two cases the drivers were not paying attention and the other was in extremely dense fog.

I must say some of the questions being asked here remind me of the question.....

"Are you still beating your wife?"

No matter how you answer the question you're in trouble. It's best to just not answer the question.

Sent from my XT1650 using Tapatalk
 

When I read this reply I burst out laughing . . . Upon reflection, I was sadden because I know a few ladies that endured this sort of relationship. 

On Topic: This whole self driving technology does't bother me in the big picture. What bothers me is the fact the general populace have taken every known invention created by man and abused it. Some of you have already noted the vulnerable cell phone keeping in mind this technology has been present in some form for ages. Yet there was no such thing as *Distracted Driving*.

That's because there was no such thing as being able to send emoticons, pictures, reading books, listening to music, watching video's. Just this simple advancement has created a specific road hazard never seen before. This is exactly like those who insist upon drinking and driving and believe they are completely fine to do so. Now, people here and others around the world will take the exact same technology which allows them to be completely lazy, inept, and clueless.

To allow them to so called be more productive while on a long commute??   

Please . . .

If that isn't complete B.S. I don't know what is.

Nobody has even addressed who will be responsible for any road fatalities or property damage?!?! I find it hard to believe any sane company would be willing to place their companies lively hood on the chopping block. Upon reflection and being factual there isn't a law in the books that can say John Do is guilty for X.

Why???

Because the current laws state and define who has *Care and control of said vehicle* A person sitting in the back of the vehicle does not have immediate care and control of the vehicle. This is exactly the same laws stating if you're drunk and simply leave the keys in the ignition but sit in the back you're still legally liable. Now, the same can't be said when a robot is driving the vehicle because YOU ARE NOT DRIVING.

Thus, any imbecile can literally be drunk, stoned, romping in the back seat while Robby the robot is driving them from A to B. <-- In this scenario should the vehicle hit a person, vehicle, building, how could anyone or Government file suit or charge the guy in the back??

They couldn't based on current laws in the books. But one thing for sure all of the ambulance chasing lawyers would be lined up to sue GM, Tesla, Waymo, Uber, etc. I can tell you with 100% confidence the first time one of these companies get sued and lose millions of dollars they will either fold, or leave the market place. 

The idea is quite novel but the deployment of said concept is poorly executed. 

 

[dbuss]

When I read this reply I burst out laughing . . . Upon reflection, I was sadden because I know a few ladies that endured this sort of relationship. 
On Topic: This whole self driving technology does't bother me in the big picture. What bothers me is the fact the general populace have taken every known invention created by man and abused it. Some of you have already noted the vulnerable cell phone keeping in mind this technology has been present in some form for ages. Yet there was no such thing as *Distracted Driving*.
That's because there was no such thing as being able to send emoticons, pictures, reading books, listening to music, watching video's. Just this simple advancement has created a specific road hazard never seen before. This is exactly like those who insist upon drinking and driving and believe they are completely fine to do so. Now, people here and others around the world will take the exact same technology which allows them to be completely lazy, inept, and clueless.
To allow them to so called be more productive while on a long commute??   
Please . . .
If that isn't complete B.S. I don't know what is.
Nobody has even addressed who will be responsible for any road fatalities or property damage?!?! I find it hard to believe any sane company would be willing to place their companies lively hood on the chopping block. Upon reflection and being factual there isn't a law in the books that can say John Do is guilty for X.
Why???
Because the current laws state and define who has *Care and control of said vehicle* A person sitting in the back of the vehicle does not have immediate care and control of the vehicle. This is exactly the same laws stating if you're drunk and simply leave the keys in the ignition but sit in the back you're still legally liable. Now, the same can't be said when a robot is driving the vehicle because YOU ARE NOT DRIVING.
Thus, any imbecile can literally be drunk, stoned, romping in the back seat while Robby the robot is driving them from A to B. They couldn't based on current laws in the books. But one thing for sure all of the ambulance chasing lawyers would be lined up to sue GM, Tesla, Waymo, Uber, etc. I can tell you with 100% confidence the first time one of these companies get sued and lose millions of dollars they will either fold, or leave the market place. 
The idea is quite novel but the deployment of said concept is poorly executed. 
 

When I read this reply I burst out laughing . . . Upon reflection, I was sadden because I know a few ladies that endured this sort of relationship. 

My point exactly Teken.

That's what happens when a question is asked and a yes or no answer is required.

Sent from my XT1650 using Tapatalk

[Teken]

Just now, dbuss said:

When I read this reply I burst out laughing . . . Upon reflection, I was sadden because I know a few ladies that endured this sort of relationship. 

My point exactly Teken.

That's what happens when a question is asked and a yes or no answer is required.

Sent from my XT1650 using Tapatalk
 

Well, to be fair my question was to prove a point and illicit a honest reply from the forum members. If a person in good confidence and sound mind can't reason that its not a good idea to leave a child in a vehicle that has either been well proven vs not proven.

This immediately tells me there's something wrong.

Both Boeing and Airbus have proven over tens of thousands of miles and years of trials that their planes can fly unattended from point A to B. This is undisputed as it happens every second of every day as I write this. What took longer was to prove the same technology could reliably control a large vehicle from take off to landing in all weather conditions. Again, this has been proven by both companies because this is part of the larger fail over plan in case of a pilot failure.

Some of the conditions they wanted to address was as common as heart attack, stroke, sudden depressurization, and terrorist threats.

Yet not a soul has taken the plunge to get on board and fly A to B, why???

This same analogy directly relates to placing a child to animal in said automated vehicle. Nobody serious on this forum has spoken up to affirm the same besides offer counterpoints with caveats, if, but, etc.

That isn't a reply that instills confidence to me or those who have concerns the current deployment of this concept is totally flawed.