Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Self-driving Uber car kills Arizona woman crossing street

reuters.com

Post reply on HN
Print view View original

911–920 of 1001 posts

Re: Self-driving Uber car kills Arizona woman crossing street

#911

Post by smilekzs »

rboyd wrote: post #847

FTA "Elaine Herzberg, 49, was walking her bicycle outside the crosswalk". The article says "as soon as she walked into the lane of traffic she was struck". The police sergeant also "said he believed Herzberg may have been homeless". These three things may together indicate that even a human driver might have struck and killed the lady if she entered into the roadway unsafely. I'm not sure how everyone (especially on…

Anecdote: About a year ago when I was driving at night I almost hit someone, apparently homeless with dark-colored clothes, pushing a covered shopping cart across a 5-lane street with no crosswalk. Fortunately I was in one of the middle lanes so I noticed them and slammed the brakes in time (while risking getting myself rear-ended).

Re: Self-driving Uber car kills Arizona woman crossing street

#912

Post by BucketSort »

icc97 wrote: post #713

One aspect that comes from this is that now car crashes can be treated more like aircraft crashes. Each self-driving car now has a black box in it with a ton of telemetary. So it's not just "don't drink and drive", knowing that they'll probably reoffend soon anyway. Every crash and especially fatality can be thoroughly investigated and should be prevented from ever happening again. Hopefully there's enough data in th…

Black box with a ton of telemetry being piped into black box models though.

Re: Self-driving Uber car kills Arizona woman crossing street

#913

Post by JimboOmega »

femto wrote: post #906

> “They are going to attempt to try to find who was possibly at fault and how we can better be safe, whether it’s pedestrians or whether it’s the vehicle itself.” Autonomous vehicles should automatically be 100% liable in every accident in which a person is injured, and it should be paid for by a levy on every vehicle. A levy is fair because the per vehicle risk is much more uniform than with a driver behind the whee…

Given current road and vehicle design, that's an incredibly impractical goal. There simply isn't enough separation between pedestrians and roadways; to keep injury risk negligible you'd also have to keep speed incredibly low (way below speed limits) on all but the best limited access highways.

If the vehicle can only crawl at 5MPH because you never know when a pedestrian might just dive in front of you... that's just impractical for real vehicle operation.

It's one thing to hit the brakes because somebody decided to jaywalk without looking; another to expect there is always enough space to stop in those situations.

Re: Self-driving Uber car kills Arizona woman crossing street

#914

Post by lev99 »

ejstronge wrote: post #895
raphaelj wrote: post #832

To put this accident in perspective, Uber self-driving cars totaled about 2 to 3 millions miles, while the fatality rate on US roads is approximately 1.18 deaths per 100 millions miles [1]. [1] https://www.nhtsa.gov/press-releases/usdot-releases-2016-fat...

The appropriate comparison would be to ask how many pedestrians were struck by a car and killed. Considering that humans would likely slow down if they see a pedestrian - even if one appeared suddenly - this is even more disconcerting.

It's more specific, but I care about all fatalities, not just car-pedestrian fatalities.

Re: Self-driving Uber car kills Arizona woman crossing street

#915

Post by web007 »

yelloweyes wrote: post #851
IgorPartola wrote: post #390

This is what's going to happen. If you've ever seen a machine learning algorithm in action, this isn't surprising at all. Basically, they'll behave as expected some well known percentage of the time. But when they don't, the result will not be just a slight deviation from the normal algorithm, but a very unexpected one. So we will have overall a much smaller number of deaths caused by self driving cars, but ones that…

Very well put. It boggles my mind that a forum full of computer programmers can look at autonomous cars and think "this is a good idea". They are either delusional and think their code is a gift to humanity or they haven't put much thought into it.

I don't believe I'm delusional, my code is certainly medium-to-okay. I've put a lot of thought into this. I think autonomous cars are a very good idea, I want to work on building them and I want to own one as soon as safely possible.

Autonomous cars, as they exist right now, are not up to the task at hand.

That's why they should still have safety drivers and other safeguards in place. I don't know enough to understand their reasoning, but I was very surprised when Waymo removed safety drivers in some cases. This accident is doubly surprising, since there WAS a safety driver in the car in this case. I'll be interested to see the analysis of what happened and what failures occurred to let this happen.

Saying that future accidents will be "unexpected" and therefore scary is FUD in its purest form, fear based on uncertainty and doubt. It will be very clear exactly what happened and what the failure case was. Even as the parent stated, "it saw a person with stripes and thought they were road" - that's incredibly stupid, but very simple and explainable. It will also be explainable (and expect-able) the other failures that had to occur for that failure to cause a death.

What set of systems (multiple cameras, LIDAR, RADAR, accelerometers, maps, GPS, etc.) had to fail in what combined way for such a failure? Which one of N different individual failures could have prevented the entire failure cascade? What change needs to take place to prevent future failures of this sort - even down to equally stupid reactions to failure as "ban striped clothing"? Obviously any changes should take place in the car itself, either via software or hardware modifications, or operational changes i.e. maximum speed, minimum tolerances / safe zones, even physical modifications to configuration of redundant systems. After that should any laws or norms be changed, should roads be designed with better marking or wider lanes? Should humans have to press a button to continue driving when stopped at a crosswalk, even if they don't have to otherwise operate the car?

Lots of people have put a lot of thought into these scenarios. There is even an entire discipline around these questions and answers, functional safety. There's no one answer, but autonomy engineers are not unthinking and delusional.

Re: Self-driving Uber car kills Arizona woman crossing street

#916

Post by dilanj »

According to the latest info, it doesn't look like the system was at fault at all. Seem like the likely homeless woman abruptly stepped into traffic. Every accident still of-course is terrible, but some maybe are unavoidable. https://www.sfchronicle.com/business/article/Exclusive-Tempe...

Re: Self-driving Uber car kills Arizona woman crossing street

#917

Post by LyndsySimon »

lev99 wrote: post #914
ejstronge wrote: post #895

Earlier quoted context omitted.

The appropriate comparison would be to ask how many pedestrians were struck by a car and killed. Considering that humans would likely slow down if they see a pedestrian - even if one appeared suddenly - this is even more disconcerting.

It's more specific, but I care about all fatalities, not just car-pedestrian fatalities.

At some point, autonomous vehicles will be carrying cargo instead of passengers. When that happens, fatalities per mile driven will no longer be a valid comparison between manned and unmanned vehicles, as catastrophic accidents will have fewer people involved.

Re: Self-driving Uber car kills Arizona woman crossing street

#918

Post by atopuzov »

mindslight wrote: post #839
fudged71 wrote: post #766

Earlier quoted context omitted.

Although it's comforting that this exact situation shouldn't happen again in an Uber autonomous car... there is no mechanism to share that learning with the other car companies. There seriously fucking needs to be a consortium for exactly this purpose: sharing system failures. Also my problem with this is that a human death is functionally treated as finding edge cases that are missing a unit test, and progressing th…

> We need to avoid treating deaths as progress in the pursuit of better things Au contraire. Go read building codes some time. There's a saying that they're "written in blood" - every bit, no matter how obvious or arbitrary seeming, was earned through some real-world failure. The death itself isn't progress, of course. But we owe it to the person to who died to learn from what happened.

This reminds me of the story behind the iron ring and the collapse of the second narrows bridge in Vancouver, BC.

http://www2.ensc.sfu.ca/undergrad/euss/enscquire/Vol10No2/pa...

Re: Self-driving Uber car kills Arizona woman crossing street

#919

Post by DoreenMichele »

rboyd wrote: post #847

FTA "Elaine Herzberg, 49, was walking her bicycle outside the crosswalk". The article says "as soon as she walked into the lane of traffic she was struck". The police sergeant also "said he believed Herzberg may have been homeless". These three things may together indicate that even a human driver might have struck and killed the lady if she entered into the roadway unsafely. I'm not sure how everyone (especially on…

The article says she died from her injuries some time later. So, she didn't die instantly on impact.

I know more than average about homelessness. A high percentage of homeless people have serious health problems. If she was homeless, she may have already been in very poor health. That may have contributed to her death.

I can think of potentially other contributing factors, but I am leery of coming across like I am badmouthing homeless people. They are subjected to enough hostility and prejudice and I am usually trying to advocate on their behalf.

Re: Self-driving Uber car kills Arizona woman crossing street

#920

Post by femto »

JimboOmega wrote: post #913
femto wrote: post #906

> “They are going to attempt to try to find who was possibly at fault and how we can better be safe, whether it’s pedestrians or whether it’s the vehicle itself.” Autonomous vehicles should automatically be 100% liable in every accident in which a person is injured, and it should be paid for by a levy on every vehicle. A levy is fair because the per vehicle risk is much more uniform than with a driver behind the whee…

Given current road and vehicle design, that's an incredibly impractical goal. There simply isn't enough separation between pedestrians and roadways; to keep injury risk negligible you'd also have to keep speed incredibly low (way below speed limits) on all but the best limited access highways. If the vehicle can only crawl at 5MPH because you never know when a pedestrian might just dive in front of you... that's just…

I disagree. Pedestrian density is generally low enough that the car could generally be doing 60km/h and not hit a "worst case" pedestrian.

For example, how many people do you see walking along freeways? If a person is on a freeway then cars should be slowing down for them. It's a rare enough event that cars could be programmed to avoid all collisions with minimal impact on average speeds.

If a car is around people then it should be going at a low speed. If there are enough people around that it can't get speed up, then that car is effectively in a "shared" or "local traffic" area and should be going slowly.

I'd argue that the effect on average speeds will be similar to having low speed limits on residential streets:a slower period at journey beginning/end but the majority of the journey would be at speed on sparsely populated arterial roads.

Here's another argument: intrinsically safe programming might even increase average speeds? Why do we even need traffic lights and pedestrian crossings with autonomous vehicles? If vehicles were 100% safe, we could do away with dedicated crossings, and there would be no need for vehicles to waste time at red lights, merely slowing down on the relatively rare times when it is necessary.

Post reply on HN