Self Driving Uber Kills Pedestrian

Here are some statistics about deaths in traffic in the US.

https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in_U.S._by_year

It is surprising that numbers are not coming down but look fairly constant despite joint efforts to make traffic safer. A total of 37,461 people died in 2016. That is 1.18 persons per one million miles traveled and 11.59 people per 100,000 inhabitants. Is it not remarkable how socially acceptable these numbers seem? Or are you aware of any outcry? A war on traffic maybe? Is this 'just the price we have to pay for our mobility'? Do you think that this will change in future? And even more interesting: what will happen when autonomous vehicles come into the equation?

I guess we need to answer these questions before we let autonomous vehicles loose onto the roads. I am an automation engineer by trade and I believe that there will be a technical solution in future to bring fatality numbers down by 99% when traffic is eventually fully automated. It is out of question if this will happen. The question is more when will it happen and how will the transition period look like. I am just guessing here but I think that once we reached a threshold of around 25% of fully automated transport the change will come super quick. It will become apparent to insurers, the public and even the greatest sceptic that humans are really bad at driving cars and that killing people is not normal or just another tragedy. The airline industry has shown us how its done and land traffic will follow, and that is something I have no doubts about it. I am sure that in 10, 20, 30, 40 or maybe 50 years time we will look back at the quoted fatality numbers and call it the dark ages of human transport.
 
Punx0r said:
kerim said:
A total of 37,461 people died in 2016. That is 1.18 persons per one million miles traveled

It's 1.18 fatalities per hundred-million miles :)


... or 102+ persons killed each day... in the USA alone... (Never mind other countries...) And injuries and property damages? (Like, how many suffer from being "overweight"...) :wink:
 
RemoteRoad said:
Strawman, Bill. You listed three stories of tens of thousands on the incident.
I listed the first three that popped up as a result of a Google search. Since it's a very common search engine, those are the stories that will often be seen first by people searching for it.
Furthermore, business papers are exceptional in the media biz. They tend to be far more objective.
Right. So does AP, AFP, BBC, Reuters and NPR. NBC, ABC and CBS provide more popular content. CNN tends more towards infotainment/spectacle and FOX tends towards political propaganda. Every source has their own audience and own agenda.
Further, headlines and a few quotes don't reveal the sum effect. I've seen unnecessary and less than fully relevant character attacks on the pedestrian and now the "monitor-driver" in nearly every article I've read.
Sounds like you might be imposing your own biases on the news you are reading; seeing what you want to see.
 
LockH said:
A total of 37,461 people died in 2016. That is 1.18 persons per one million miles traveled

It's 1.18 fatalities per hundred-million miles :)

... or 102+ persons killed each day... in the USA alone... (Never mind other countries...) And injuries and property damages? (Like, how many suffer from being "overweight"...) :wink:

Are you kidding me? Obesity contributes to heart disease, stroke, and diabetes. The problem might be as much as ten times as large as car fatalities depending oh how you want to factor the degree to which obesity and related behaviors contribute to the problem. And that doesn't factor the number of car related deaths from the additional vehicle miles for food delivery and grocery shopping. ;^)

I'm tellin' ya. My grand plan is the way to go.

https://www.nhlbi.nih.gov/news/2017/high-number-deaths-heart-disease-stroke-and-diabetes-linked-diet

"Of the 702,308 adult deaths due to cardiometabolic diseases, 318,656, or about 45 percent, were associated with inadequate consumption of certain foods and nutrients widely considered vital for healthy living, and overconsumption of other foods that are not."
 
kerim said:
Here are some statistics about deaths in traffic in the US.

https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in_U.S._by_year

It is surprising that numbers are not coming down but look fairly constant despite joint efforts to make traffic safer. A total of 37,461 people died in 2016. That is 1.18 persons per one million miles traveled and 11.59 people per 100,000 inhabitants. Is it not remarkable how socially acceptable these numbers seem? Or are you aware of any outcry?<snip>

The numbers are historically quite low per the wiki link you've provided. So comparatively, I can see why there is no general outcry.

The trend downward was fairly steady for many decades up until about 2010. Fatality rates have halved twice in my lifetime and more than three times in my parent's. Then the reduction plateaued. Hmmmm... What does that coincide with? Smart phone usage? It may not be, but it certainly might be. And a lot of people are crying out about the distracted driving problem.

http://www.pewinternet.org/fact-sheet/mobile/
 
Hehe... "My grand plan is the way to go."... Still hoping for any link to the "grand plan"? During my recent extended vacation through the local "health care system"... various places with signs watt still say "hospital"... I commented to one doctor on the number of... "large" folks... (used to be termed "obese"... now its just "pleasantly plump"...) all staggering around in one waiting room. ... and the Doc just rolled his eyes... :roll:
 
So far, the cumulative fatality rate is 100%.

News media is the primary driver for people's responses. If your news carried 100 grisly photos of car deaths every day there might be a concern about it. Since that's not on their agenda they only barely report local deaths, so the frequency is low. But when there's news that fits their agenda anywhere in the country, or even the world, they milk it hard.
 
spinningmagnets said:
The cars that seems to have to most survivable strikes against pedestrians had a low bumper with a sloping hood that would "sweep you off your feet"
Lotus
ftw
AGAIN!
Punx0r said:
She certainly did herself no favours wearing dark clothing and if she failed to properly judge the speed of traffic, but the car driver carries the greater responsibility to avoid such a collision. If you follow the driving maxim that you should drive at a speed where you can safely stop in the distance you know to be clear, then the car is 100% at fault. The woman didn't unexpectedly step off a kerb as the car was passing

in that clip
right at the end
just b4 impact.

you can clearly see that she's standing in the way of progress.
 
When I look at this video, the first thing I think about is where is the raw video? This seems to be just someone with an iPhone recording it from a screen. It even has all the sides with the numbers blurred out.
The raw video should also not include the blurred out numbers and white line border. White line borders make it harder for the dark parts of the image to show up.
I assume the raw video would be at least 1080p quality that is better than the $30 1080p dashcams you can buy off ebay these days which do an amazing job for the money. Like this one https://youtu.be/DGXiXz6QXQ4?t=14m59s

Also if this was even a remotely decent investigation they would be releasing a video of what the LIDAR feed saw, like this below shown in google cars video.
I hope the authorities still have the Uber car so that more capable investigators have a look at all this. LIDAR video is what the car REALLY SAW..

All up it seems clear to me that the Google self-driving cars are 10 times smarter and safer than these Uber cars. These Uber cars seem to just drive around based on luck on not hitting anyone than anything else.

f1TnF3w.gif


The Google cars are much much smarter, I hope they don't allow Uber cars in Australia for a long time until they catch up to what the Google cars can do.
https://youtu.be/tiwVMrTLUWg?t=11m5s

What I like about this Google car is the LIDAR could see a cyclist coming from way over from the other side of the road and see it was an issue, this is completely different from the UBER car.
https://youtu.be/tiwVMrTLUWg?t=12m33s
 
Any FLIR device would have caught the woman walking across the street. I wonder how fast the car was driving, plus I do not see any reflectives on the bike.

Someone stated 40mph, or 38mph in a 35mph zone.
 
IMO, this particular incident had little to do with self driving car. I don't agree that most people would have stopped or swerved in time. Some would, but many would have just braked hard, still hitting the ped. Quite possibly not fatal then, but maybe so. But the bottom line is the cars lane was clear till the last second, giving the car no reason to think it must stop. The hard thing for self driving cars, or human operators too, is anticipating that some person will really be that stupid. Had she been seen by the human driver, hard to say if slam on brakes would have even been the decision. Easy to think the person would just pause in the other lane. She did not.

I nearly hit two women pushing babies in strollers once, because it never occurred to me they would step into the street while their heads were turned away from me. Had I seen faces, I would have thought they were about to do it. But no look at all fooled me. They just stepped out without checking at all, with their infants. UNBELIVEABLE! Had it been night and no streetlight, I'd have killed 4 people.

This person made a big mistake crossing as she did. She had the least situational awareness compared to the car. She only had to see the car, well lit, coming, and pause in the other lane. She just walked right into that car, that was very close as she entered that lane. The black coat did not help things much, but I do understand that for the homeless, black can be a survival outfit.

Suicide is possible here, likely not, but possible. Locally we have had old sick people just walk right into the train.
 
dogman dan said:
IMO, this particular incident had little to do with self driving car. I don't agree that most people would have stopped or swerved in time. Some would, but many would have just braked hard, still hitting the ped. Quite possibly not fatal then, but maybe so. But the bottom line is the cars lane was clear till the last second, giving the car no reason to think it must stop. The hard thing for self driving cars, or human operators too, is anticipating that some person will really be that stupid.

That's a moving violation called "driving too fast for conditions", or IMO reckless driving. These sorts of tragedies are the side effect of allowing traffic to move at deadly speeds where people are present. We don't have to allow it, the first cars were prohibited from doing it, and I will move to the first halfway decent American city that develops the moral fiber to prohibit it again.

It's not ethical to choose a deadly form of transport and then obligate non-deadly road users to make way for you, and it's not ethical to blame the victim when the inevitable happens. We're just accustomed to this brand of unethical behavior.
 
I had a somewhat similar thing happen a couple years ago while I was riding my ebike. The pedestrian was in the shade, halfway around a corner as I came in the sun, Due to the shade and the curve I could not see them until I was nearly upon them. This in an area where no pedestrians normally exist. The pedestrian seemed to pay no attention to an oncoming electric bike travelling at 28 mph. A rapid course adjustment allowed me to swing behind them. There was insufficient time to brake. In situations like this braking is rarely adequate. I got a huge adrenaline surge afterward. It was a close call.

It is not so unusual for the human eye and brain to figure out a tricky situation. But it is not 100% successful.

I keep wondering if the pedestrian had earphones in. Another Iphone statistic. But we don't know. I see plenty of them riding bicycles. A lot of motorcyclists and car drivers as well.

How do the self driving cars do at recognizing less common road hazards like a 4x4 laying in the road? A new pothole? A fallen cyclist or pedestrian? Emergency flares and cones? Speed bumps? modified lanes in construction areas?

Handing the special cases is hard. That's always the difficult part of software.
 
markz said:
Any FLIR device would have caught the woman walking across the street.
FLIR works only when the temperature of the object is different from the background temperature. On a road shortly after sunset that would likely not be the case. Later at night it would be.
 
Alan B said:
I keep wondering if the pedestrian had earphones in. Another Iphone statistic.

Why is anything the ped does even at issue? Remove the hazard and the peds can walk and listen however they want-- they're not threatening anybody. On the other hand, the car driver going 40mph next to a sidewalk (or e-biker going 30mph next to a sidewalk) is sooner or later going to hurt somebody no matter what he or she does to avoid it.

Keep in mind that anything an adult with ear buds can do, a child/animal/disabled person/runaway baby stroller can do even better. Is it their fault too, when it was only the car driver who ever brought in the risk of grievous harm?
 
Chalo said:
Keep in mind that anything an adult with ear buds can do, a child/animal/disabled person/runaway baby stroller can do even better. Is it their fault too, when it was only the car driver who ever brought in the risk of grievous harm?
In the case of the child, no, the child's parents would be at fault for letting their child play on a road/freeway (or a railway, or a bike path, or a runway.) Same goes for an animal; it is the animal's owner that would be at fault.
 
Keep blaming the victims, and the real culprits will keep providing you more victims to blame.
 
Chalo said:
Keep blaming the victims, and the real culprits will keep providing you more victims to blame.
How unfortunate that you didn't read my post before replying to it. It must be difficult for you to communicate with people on line using that approach.
 
Clearly the driver monitoring the self driving car was not monitoring very closely. If their purpose is to take over in an emergency their alertness was inadequate.

Pedestrians do have some responsibility in safety. That's the way things work. Even pedal bicycles cannot really operate if pedestrians walk in front of them randomly without responsibility.
 
Alan B said:
Pedestrians do have some responsibility in safety. That's the way things work. Even pedal bicycles cannot really operate if pedestrians walk in front of them randomly without responsibility.

That balance of responsibility only works if the non-lethal road users have some say over whether others get to be lethal road users. They don't have a say in it, so the burden must lie with the self-centered dickholes who think risking others' lives for their momentary convenience is acceptable.
 
Chalo said:
That balance of responsibility only works if the non-lethal road users have some say over whether others get to be lethal road users.
Of course they do. They can vote directly for ballot measures that shut down roads to vehicles, and can vote in legislators who will ban cars, or who will make the selection process for drivers much tougher.
 
USA News report quotes a autonomous vehicle expert ; " most LiDAR systems....have a blind spot of around 30 feet immediately surrounding the vehicle." Heck, that is where I want the operator to be focused while passing me ! https://www.usatoday.com/story/tech/2018/03/23/top-robotics-expert-uber-crash-questions-whether-sensors-worked/451420002/
 
Chalo said:
Keep blaming the victims, and the real culprits will keep providing you more victims to blame.

poor little robocar iz a victim too only doing as its told.
running in eco-mode it's programmed to maintain a steady even pace with no excessive braking to maximize gaz mileage.

Alan B said:
Clearly the driver monitoring the self driving car was not monitoring very closely. If their purpose is to take over in an emergency their alertness was inadequate.
yeah well he's kickin it eco mode too.
if this is an uber-eats run & the delivery didn't arrive there may very well be another victim.

c'mon we all know who's to blame here.
fire up the torches, march on up to karl uberfeldt's house.

kobayashi maru happens.
 
billvon said:
Chalo said:
That balance of responsibility only works if the non-lethal road users have some say over whether others get to be lethal road users.
Of course they do. They can vote directly for ballot measures that shut down roads to vehicles, and can vote in legislators who will ban cars, or who will make the selection process for drivers much tougher.

That kind of democracy only works when you don't have 95% of the people born into and indoctrinated as road pirates.

One day perhaps....
 
Back
Top