Author Topic: First pedestrian death by autonomous car...  (Read 27519 times)

0 Members and 1 Guest are viewing this topic.

Offline apelly

  • Supporter
  • ****
  • Posts: 1061
  • Country: nz
  • Probe
Re: First pedestrian death by autonomous car...
« Reply #25 on: March 20, 2018, 12:03:30 pm »
We know that it's impossible to pay attention 80% of the time (a number I made up based on my own driving) when your attention is required 0.1% of the time. People get bored. This is why the big boys are pushing for 100% autonomy.

But who's running the numbers here? How many deaths would there have been if shitty humans had driven all of America's autonomous hours?

People die. Sometimes in bad ways for no reason (insert mother pleading with congressman to ban swimming until you can swim) but anyway... The computer pays attention all the time, even if it isn't quite as good as an attentive human, it's already better than a drugged/tired/distracted one.
 

Online EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
Re: First pedestrian death by autonomous car...
« Reply #26 on: March 20, 2018, 12:07:58 pm »
Don't these cars have lidar, radar, etc that should have detected an object even if it was dark?  Supposed to be "safer" than a human driving right?  Seems pretty shitty if it didn't attempt to slow down or move out of the way slightly.  Should be interesting to see who comes up at fault.

I was driving home one night right across Sydney. It was torrential pouring rain, and as is normal in Sydney, roadwork galore, chopping and changing contraflow lanes, weaving in and out of stuff, stopping and going at lollipops, and dodging people in the city etc. And I was thinking, no 'effing way any autonomous car is going to do all that on a mass scale, not for a very long time. And here people are saying in only a few years time we'll all be driving fully autonomous cars. Rubbish.

 
The following users thanked this post: TheSteve, CNe7532294

Online EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
Re: First pedestrian death by autonomous car...
« Reply #27 on: March 20, 2018, 12:11:14 pm »
People die. Sometimes in bad ways for no reason (insert mother pleading with congressman to ban swimming until you can swim) but anyway... The computer pays attention all the time, even if it isn't quite as good as an attentive human, it's already better than a drugged/tired/distracted one.

And a drugged/tired/distracted human is still going to be way better at reacting to random unforseen anomalous events than some AI algorithm that thinks it knows and can handle any situation.
 

Offline apelly

  • Supporter
  • ****
  • Posts: 1061
  • Country: nz
  • Probe
Re: First pedestrian death by autonomous car...
« Reply #28 on: March 20, 2018, 12:28:35 pm »
People die. Sometimes in bad ways for no reason (insert mother pleading with congressman to ban swimming until you can swim) but anyway... The computer pays attention all the time, even if it isn't quite as good as an attentive human, it's already better than a drugged/tired/distracted one.

And a drugged/tired/distracted human is still going to be way better at reacting to random unforseen anomalous events than some AI algorithm that thinks it knows and can handle any situation.
No argument from me.

I'm guessing that despite that, inattention in "easy" situations, think freeway cruising, causes more injuries than those few complicated situations.

I want to see autonomous hours driven with no serious issues compared with human hours driven with no serious issues.
 

Offline orion242

  • Supporter
  • ****
  • Posts: 746
  • Country: us
Re: First pedestrian death by autonomous car...
« Reply #29 on: March 20, 2018, 12:55:05 pm »
I want to see autonomous hours driven with no serious issues compared with human hours driven with no serious issues.

Comparing all human hours, which would cover ALL road conditions to AV hours which are only in the best cases...bogus stats.  Your not going to find anything meaningful until AVs have alot more hours and how do you account for "disengagements"?  Is that a failure in the comparison, or its designed for that so that's perfectly acceptable no negative points?  Humans sure as hell can't tap out when the crap hits the fan.
« Last Edit: March 20, 2018, 12:56:54 pm by orion242 »
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 19520
  • Country: gb
  • 0999
Re: First pedestrian death by autonomous car...
« Reply #30 on: March 20, 2018, 01:38:03 pm »
Pedestrian crosses road in unexpected place in pitch black conditions, even human driver does not react. I have to say, some pedestrians have zero care for their safety, and seem to assume that because they can see the car, the car driver can see them. When the conditions often make the pedestrian virtually invisible. And then there are those people deliberately trying to get run over - commonplace in Russia... but then anything goes in Russia.
I agree. It's quite likely there was nothing the driver could have done.


Quote
The news reports eventually get around to some salient facts : Uber robot cars have been driving for nearly a year before this accident. One person gets killed on Arizona roads every 9 hours. Yup, humans killed 962 people (2016), robot car kills 1 person. So ban robot cars?
That doesn't mean anything. The question is: how many human drivers are there vs automated cars?

Quote
This is the question I raised before : will people accept robot cars, even if they kill a lot less people than human driven cars, but still kill a few?
I'd accept that. One big advantage is all the automated cars can learn from every single accident.

I'm still cynical that autonomous vehicles will ever take off. There are plenty of places where driver-less technology is already advanced enough, yet isn't widely implemented such as: guided bus ways and trains. Here in the UK there have been a few accidents due to excessive speed on guided bus ways, which probably wouldn't have happened, if they were automated.
 

Online EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
Re: First pedestrian death by autonomous car...
« Reply #31 on: March 20, 2018, 01:51:56 pm »
All I want to know is how the car decides where to park. If I order a Uber and stand outside, where will it stop to let me get in? Double parked, across a driveway, in a loading zone or 400m down the street. What if you drive to an event and they set up a carpark in a field, how will the car find a space? How will it get out if it rained and can't see other people churning up the mud and avoid getting bogged. How am I supposed to get out and push and tell the driver to go? Will it then stop and wait for me to get back in? Will it drive with no one inside anyway? Can I let it circle the block whilst I pick up my dry cleaning?

Yep, all these issues and more will hinder widespread adoption of fully automated humanless cars for general use for a long time.
All people need is one bad experience in a driverless Uber "Johnny Cab" and they'll likely never use one again, and they'll tell their friends about it.
Heck, I'd like to see an autonomous car figure out how to drive up to my office building and pick me up, it's a kinda complicated environment out the front. Simple stuff like watching people's faces and manerisisms to see if they are headed to their car to back out so you can get a spot is just one of countless examples of something that is trivial and natural for a human, but stupidly complicated for an AI car.
Just the two buildings I'm in alone have enough variability and complication to warrant their own rules. And visualising other buildings and carparks in my business park alone, I"m struggling to think of one that would be easy.
Maybe I'm underestimating the smarts of the latest AI cars and sensors, but I'd bet money that I'm not.
 

Offline mdijkens

  • Regular Contributor
  • *
  • Posts: 146
  • Country: nl
Re: First pedestrian death by autonomous car...
« Reply #32 on: March 20, 2018, 02:20:07 pm »
Fully agree.
Because cars seem to be able to drive autonomous under 'normal' 'expected' conditions, everyone talks like we're 90% there...
I think it's not even 10% yet.
How about al those situations that should not exist, but we encounter (unnoticed!) every day?
pedestrians and bikers in the city who behave 'at random' while we anticipate without even realizing because of facial expression or a small move of the head or arm?
In most big cities, if you wait with driving until the crossing is empty/safe, you'll still be waiting 3 days later
What if an accident blocked the road; we will without a doubt look for a way around it over the sidewalk or through the grass ...
what if road-signs are not placed correctly, we realize and act smart, creatively and responsible
etc.
 

Offline TK

  • Super Contributor
  • ***
  • Posts: 1722
  • Country: us
  • I am a Systems Analyst who plays with Electronics
Re: First pedestrian death by autonomous car...
« Reply #33 on: March 20, 2018, 02:44:26 pm »
self driving cars are not programmed with if-then algorithms.  They are based on ML (Machine Learning) software.  I live in one of the Uber cities and I see lots of Uber self driving cars around, day and night.  They are mapping the driving experience.  Recording every single sign, road condition, pedestrian movements.  They collect tons of data that is used to feedback the ML algorithms.  The backup drivers also provide important feedback at the end of the shift about any abnormal situation, any observation and it is also recorded and used as feedback.  Current AI is not what it used to be...
 

Offline MT

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: aq
Re: First pedestrian death by autonomous car...
« Reply #34 on: March 20, 2018, 04:53:58 pm »
The fact that Uber is a multi-billion dollar company, has nothing to do with it.

In AU and other more civilized countries perhaps not but in US hell yes.

Quote
Missy Cummings, a robotics expert(a.k.a a pointed professor by big time money and military industrial complex) at Duke University who has been critical of the swift rollout of driverless technology across the country, said the computer-vision systems for self-driving cars are “deeply flawed” and can be “incredibly brittle,” particularly in unfamiliar circumstances.

Companies have not been required by the federal government to prove that their robotic driving systems are safe. “We’re not holding them to any standards right now,” Cummings said, arguing that the National Highway Traffic Safety Administration should provide real supervision.

Federal transportation officials have relied on voluntary safety reporting to oversee the burgeoning industry, which has emphasized the life-saving potential of the technology in arguing against government mandates.

Arizona has aggressively courted driverless tech firms, based largely on its light regulatory touch. That approach has consequences, Cummings said. “If you’re going take that first step out, then you’re also going to be [the] first entity
to have to suffer these kinds of issues,” she said.

“The car cameras, the vision systems, they don’t perform inductively, meaning they can’t guess about the appearance of someone in a particular place and time,” Cummings said. “Pedestrians get hit by human drivers all the time for similar reasons,” though the exact cause of this crash remains unclear, she said.


« Last Edit: March 20, 2018, 05:02:50 pm by MT »
 

Offline orion242

  • Supporter
  • ****
  • Posts: 746
  • Country: us
Re: First pedestrian death by autonomous car...
« Reply #35 on: March 20, 2018, 05:39:17 pm »
Quote
In general, she said, autonomous vehicles have a variety of backup mechanisms to account for anomalies such as a pedestrian in the roadway.

Uber apparently needs a backup to the backup.

Quote
“Just because you map an area doesn’t mean your computer system is necessarily going to pick up a pedestrian, particularly one that wasn’t in a cross walk,” Cummings said

Peds using crosswalks lol, welcome to the typical city.  I see bikes blowing traffic lights and people walking out from between two parked cars daily as well.  Frankly maybe we need Uber AVs out here to cull the herd a bit.

Quote
“It’s going to be difficult to accept the deaths … but at some point you’ll start to see the curve bend,” he said. “The fact is these things will save lives and we need to get there.”

Why do we "NEED" to get there exactly?
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 6721
  • Country: nl
Re: First pedestrian death by autonomous car...
« Reply #36 on: March 20, 2018, 05:47:04 pm »
self driving cars are not programmed with if-then algorithms.  They are based on ML (Machine Learning) software.

The vision system isn't, but I am 100% certain that there is an expert system in there too. Trying to train some half assed neural net for it to obey traffic rules would be insanity, they are far too brittle and unpredictable.
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 6721
  • Country: nl
Re: First pedestrian death by autonomous car...
« Reply #37 on: March 20, 2018, 06:03:30 pm »
Pedestrian crosses road in unexpected place in pitch black conditions
The car has headlights and as far as we know she almost made it across before it hit her (ie. she crossed left to right and got hit on the right side of the car). Not exactly an olympic sprinter, so she must have been in the headlights for a bit.
Quote
human driver does not react.
A human driver doesn't see in IR though. As far as the car is concerned it's not much different from daylight.
 

Offline mdijkens

  • Regular Contributor
  • *
  • Posts: 146
  • Country: nl
Re: First pedestrian death by autonomous car...
« Reply #38 on: March 20, 2018, 06:20:19 pm »
self driving cars are not programmed with if-then algorithms.  They are based on ML (Machine Learning) software.

ML = if-then algorithms but then on premisses with probabilities. By feeding/learning it gets 'better' but not more 'creative' in unexpected situations.
ML is not HL and AI is not HI
The human brain rules big time!

ML is is very valuable and used in more and more domains, but it's not the holy grail; it has is place amongst everything else
Don't be fooled by the latest greatest hypes: We see them coming with big buzz and they go silently or find the humble place they deserve.
 

Offline Halcyon

  • Global Moderator
  • *****
  • Posts: 5679
  • Country: au
Re: First pedestrian death by autonomous car...
« Reply #39 on: March 20, 2018, 07:24:51 pm »
Don't these cars have lidar, radar, etc that should have detected an object even if it was dark?  Supposed to be "safer" than a human driving right?  Seems pretty shitty if it didn't attempt to slow down or move out of the way slightly.  Should be interesting to see who comes up at fault.

I was driving home one night right across Sydney. It was torrential pouring rain, and as is normal in Sydney, roadwork galore, chopping and changing contraflow lanes, weaving in and out of stuff, stopping and going at lollipops, and dodging people in the city etc. And I was thinking, no 'effing way any autonomous car is going to do all that on a mass scale, not for a very long time. And here people are saying in only a few years time we'll all be driving fully autonomous cars. Rubbish.

True, however one could argue that you're still safer with an Autonomous car as opposed to some (many?) drivers out there. There are certain populations of Sydney and without being racist, are just awful drivers, beyond awful, absolutely abominable. I've seen some shit you just wouldn't believe unless you witnessed it yourself. Then there are taxi drivers... for someone who spends their day driving, they ought to be some of the best, well practised drivers on our roads... they are not.

I'm finding as intelligence among the human race drops, so does their driving ability. In Australia, you don't even need to be able to read English to hold an unrestricted driving licence.  :o
« Last Edit: March 20, 2018, 07:26:38 pm by Halcyon »
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26906
  • Country: nl
    • NCT Developments
Re: First pedestrian death by autonomous car...
« Reply #40 on: March 20, 2018, 07:42:14 pm »
Don't these cars have lidar, radar, etc that should have detected an object even if it was dark?  Supposed to be "safer" than a human driving right?  Seems pretty shitty if it didn't attempt to slow down or move out of the way slightly.  Should be interesting to see who comes up at fault.

I was driving home one night right across Sydney. It was torrential pouring rain, and as is normal in Sydney, roadwork galore, chopping and changing contraflow lanes, weaving in and out of stuff, stopping and going at lollipops, and dodging people in the city etc. And I was thinking, no 'effing way any autonomous car is going to do all that on a mass scale, not for a very long time. And here people are saying in only a few years time we'll all be driving fully autonomous cars. Rubbish.
True, however one could argue that you're still safer with an Autonomous car as opposed to some (many?) drivers out there. There are certain populations of Sydney and without being racist, are just awful drivers, beyond awful, absolutely abominable. I've seen some shit you just wouldn't believe unless you witnessed it yourself. Then there are taxi drivers... for someone who spends their day driving, they ought to be some of the best, well practised drivers on our roads... they are not.

I'm finding as intelligence among the human race drops, so does their driving ability. In Australia, you don't even need to be able to read English to hold an unrestricted driving licence.  :o
I agree with that. Some people suck bad at driving through less than optimal circumstances. A fleet of self driving cars at least has the ability to share information about the road ahead so the cars behind don't have to re-analyse the entire situation. Having written that... in one occassion a Telsa in front of me nearly went full stop when coming up to road works. I'm pretty sure the driver had the auto pilot on because a human driver would never be that stupid.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline raptor1956

  • Frequent Contributor
  • **
  • Posts: 869
  • Country: us
Re: First pedestrian death by autonomous car...
« Reply #41 on: March 20, 2018, 08:51:04 pm »
People die. Sometimes in bad ways for no reason (insert mother pleading with congressman to ban swimming until you can swim) but anyway... The computer pays attention all the time, even if it isn't quite as good as an attentive human, it's already better than a drugged/tired/distracted one.

And a drugged/tired/distracted human is still going to be way better at reacting to random unforseen anomalous events than some AI algorithm that thinks it knows and can handle any situation.


I don't know about that ... since the autopilot is looking at several cameras, radar, lidar, and also ultrasonic sensors it tends to have a vastly great situational awareness than a person.  That and the fact that the processing power available greatly exceeds the typical distracted human driver.  To be sure, when encountering something unexpected humans are infinitely superior to AI, but in order to properly respond you need to see it first.  Given the info we have it seems that had the car been driven by a human the outcome would have been the same.

I would like to know if the Uber car have infrared cameras as that would be particularly useful in seeing a human or animal in darkness like that.  The advantage AI has is being able to process input from many more sensors and 100% concentration.  Again, I think IR is key and without it AI is not going to fair much better at night.


Brian
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 6721
  • Country: nl
Re: First pedestrian death by autonomous car...
« Reply #42 on: March 20, 2018, 08:52:45 pm »
It has a greater array of inputs, it has zero awareness.
 

Offline raptor1956

  • Frequent Contributor
  • **
  • Posts: 869
  • Country: us
Re: First pedestrian death by autonomous car...
« Reply #43 on: March 21, 2018, 01:22:04 am »
It has a greater array of inputs, it has zero awareness.

OK, perhaps you can explain what you mean by "zero awareness".  Seems to me that if it had zero awareness it would not work at all, but again, perhaps you can clarify you comment.


Brian
 

Offline fubar.gr

  • Supporter
  • ****
  • Posts: 366
  • Country: gr
    • Fubar.gr
Re: First pedestrian death by autonomous car...
« Reply #44 on: March 21, 2018, 07:23:01 am »
Fully agree.
Because cars seem to be able to drive autonomous under 'normal' 'expected' conditions, everyone talks like we're 90% there...
I think it's not even 10% yet.
How about al those situations that should not exist, but we encounter (unnoticed!) every day?
pedestrians and bikers in the city who behave 'at random' while we anticipate without even realizing because of facial expression or a small move of the head or arm?
In most big cities, if you wait with driving until the crossing is empty/safe, you'll still be waiting 3 days later
What if an accident blocked the road; we will without a doubt look for a way around it over the sidewalk or through the grass ...
what if road-signs are not placed correctly, we realize and act smart, creatively and responsible
etc.

According to this article, self driving cars need special parking spaces at the time being:


Quote
Nissan notes autonomous vehicles can’t yet navigate public pick-up and drop-off zones as a human driver would be able to do, given many people stop illegally to drop off or pick up passengers _ and that’s something an autonomous vehicle isn’t programmed to do. Instead dedicated parking bays will be needed.

Online BrianHG

  • Super Contributor
  • ***
  • Posts: 7733
  • Country: ca
Re: First pedestrian death by autonomous car...
« Reply #45 on: March 21, 2018, 08:09:44 am »
The news reports eventually get around to some salient facts : Uber robot cars have been driving for nearly a year before this accident. One person gets killed on Arizona roads every 9 hours. Yup, humans killed 962 people (2016), robot car kills 1 person. So ban robot cars?

I got 1 minor issue with that comment, how many cars are on Arizona's roads VS how many autonomous cars.
Now, of course, there is not yet a large enough sample set to do a proper evaluation, but, things are looking horrible for the autonomous cars.

Accidents/deaths will fall into my little 3 categories:
1. Death in a situation where neither a human driver or a computer driver could have prevented a death.
2. Death in a situation where a human driver would have not caused a death.
3. Death in a situation where a computer driver would have not caused a death.

     Unfortunately, leaving life in the hand of a computer technology, especially when new, will always get the thumbs down compared to when a human caused such an accident.  To be human means to make errors.  Having a machine do the work, it must be foul-proof.
     Worse, the programming of these cars is basically trying to drive by rules.  They are not designed specifically to watch out for human lives, and if a pedestrian's life is in danger where braking isn't enough, are these computer drivers programmed to turn off the road into the sidewalk if it is clear of pedestrians and if it is low enough to drive up on, or turn to the on-coming traffic lane if no cars are coming from that direction to avoid the pedestrians?  Is it programmed to recognize a cliff or ditch on one side to turn the other way to avoid the pedestrian & not kill the driver by driving off a cliff?  What about slamming on the park brake or run the transmition in reverse on dry pavement where there is enough traction to stop even faster?  What about honk the horn and flash the high beams if the computer sees a beginning to cross the street to stop them and save the day?


What about me as the passenger of the self driving car?  How would I feel if I chose to buy one of these full self driving cars, and I'm sitting there & my computerized car just killed someone in front me where if I owned a normal car, I could have prevented this loss of life?  How do you think I would feel?  I bought a machine which just mowed over a mother and/or child...  For now, self driving technology is to be avoided like a plague until AI reaches the kind of nuance decision human pedestrian avoidance capabilities of my above paragraph.  Sadly, I know that the amount money involved means there will be a lot of crap exponentially growing for around 2 decades before the cars get that smart.
« Last Edit: March 21, 2018, 09:01:21 am by BrianHG »
 

Offline edy

  • Super Contributor
  • ***
  • Posts: 2385
  • Country: ca
    • DevHackMod Channel
Re: First pedestrian death by autonomous car...
« Reply #46 on: March 21, 2018, 02:01:59 pm »
There is the issue of sharing of the road between autonomous vehicles (AV's) and regular human-driven vehicles, and the mixed type of traffic (pedestrian, bike, light cars, light trucks and heavy trucks) all using the roads which is bound to create numerous problems as well. While I think AV technology will eventually find a place in the future, I don't think it will take off until there are dedicated lanes or roads for it (with higher than normal "protections" in place to avoid non-AV's from using the same areas) or until almost all vehicles on the road are AV's and have the ability to talk to each other and regular speeds and follow strict rules, spacing, distancing and so on.

Once you throw in unpredictable humans who would like to "cut off" an AV, or signal last minute (or not at all) or introduce all sorts of surprises into the equation, we will end up with much confusion on the part of AV's and bloated algorithms that probably will have flaws and miss situations.

I can see AV technology working well when you want it to take over driving on long boring routes on highways, where you may need to drive on a designated "AV lane" and you put it on auto-pilot until you are ready to take control and drive through less predicatable situations. I can see urban driving with AV where speeds are relatively low and using small light compact cars that will buzz around, again along certain lanes and routes, to get predictably and safely around a city. However, the number of potential objects that such a car would encounter in an urban situation is also going to create many situations where accidents can occur especially when mixed in with unpredictable aggressive human drives, pedestrians and bicyclists in urban environment.

I'm thinking for now at least, AV will find it's place in certain niches but I cannot see it taking over the masses. Even if the technology is perfected, there are legal issues and who to blame with accidents and the human trust factor that needs to be overcome.

I'm sure we can design airplanes that can take off, fly and land completely autonomously today. There is no need for a human pilot right? Will flying be safer having a completely computer-run airline handling everything from start to finish? Don't many of the accidents involving planes boil down to human error? How many are mechanical and how many are pilot related? Yet we would still be happier having a flawed human pilot flying us than a "perfect" computer algorithm.



YouTube: www.devhackmod.com LBRY: https://lbry.tv/@winegaming:b Bandcamp Music Link
"Ye cannae change the laws of physics, captain" - Scotty
 

Offline GlennSprigg

  • Super Contributor
  • ***
  • Posts: 1259
  • Country: au
  • Medically retired Tech. Old School / re-learning !
Re: First pedestrian death by autonomous car...
« Reply #47 on: March 21, 2018, 02:16:55 pm »
It's terrible of course!!!!... but what was not delved into too much... yet......
is that this unfortunate victim was walking with, (wheeling) this bicycle along side themselves,
on the sidewalk at walking pace, and suddenly walked out in front of the traffic as if to cross
the road...   The "Vehicle" did not spot such danger ahead in the 1 sec it had to react.....
The brakes were immediately applied, and although a collision occurred, the bike ended up
being located/knocked about 10 feet in front of the now stopped vehicle......

Yes... a 'real' person was also present "behind the wheel", but even HE could not react faster....
Is there room for improvement?....  YES....  Different result with purely human driver...  no.....
Diagonal of 1x1 square = Root-2. Ok.
Diagonal of 1x1x1 cube = Root-3 !!!  Beautiful !!
 

Offline orion242

  • Supporter
  • ****
  • Posts: 746
  • Country: us
Re: First pedestrian death by autonomous car...
« Reply #48 on: March 21, 2018, 02:24:19 pm »
Was there a report that the brakes where applied?  I haven't heard anything other than it plowed right in.  And with all the sensors strapped on top, shouldn't it have detected the object even in complete darkness from some distance??

Hard to say without seeing the video.  Myself anytime I see someone walking on the side of the road I'm moving over to on top of the center line or into the on coming lane if possible.  We have ass clowns down here that will ride bikes on snow/icy roads.  They could slip and fall over at any minute.  Is the AI smart enough to do the same?  Can it sort out a mailbox from a child standing on the side of the road and do the same?
« Last Edit: March 21, 2018, 02:25:50 pm by orion242 »
 

Offline orion242

  • Supporter
  • ****
  • Posts: 746
  • Country: us
Re: First pedestrian death by autonomous car...
« Reply #49 on: March 21, 2018, 02:29:39 pm »
According to this she walked right in front of the AVs path of travel and it never slowed down.

https://www.nytimes.com/interactive/2018/03/20/us/self-driving-uber-pedestrian-killed.html

Being equipped with both lidar and radar, it shouldn't have had much trouble spotting her.  Especially since she cross right in front!  Looking at it more, there are street lights on that portion of the road.  You can see their shadows on the road in their picture.  Given that, the headlights...even the damn person behind the wheel should have noticed her.  Again, if your used to the AV driving 99% of the time...how vigilant are you really for the 1% when the unexpected comes into the equation?
« Last Edit: March 21, 2018, 02:35:23 pm by orion242 »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf