Human Driver Could Have Avoided Fatal Uber Crash, Experts Say (bloomberg.com)

An anonymous reader shares a report: The pedestrian killed Sunday by a self-driving Uber SUV had crossed at least one open lane of road before being hit, according to a video of the crash that raises new questions about autonomous-vehicle technology. Forensic crash analysts who reviewed the video said a human driver could have responded more quickly to the situation, potentially saving the life of the victim, 49-year-old Elaine Herzberg. Other experts said Uber's self-driving sensors should have detected the pedestrian as she walked a bicycle across the open road at 10 p.m., despite the dark conditions. Herzberg's death is the first major test of a nascent autonomous vehicle industry that has presented the technology as safer than humans who often get distracted while driving. For human driving in the U.S., there's roughly one death every 86 million miles, while autonomous vehicles have driven no more than 15 to 20 million miles in the country so far, according to Morgan Stanley analysts. "As an ever greater number of autonomous vehicles drive ever an ever greater number of miles, investors must contemplate a legal and ethical landscape that may be difficult to predict," the analysts wrote in a research note following the Sunday collision. "The stock market is likely too aggressive on the pace of adoption."

Comments

I probably would have hit her

[ZorinLynx]

Based on the video I saw, she was practically invisible until she entered the car's headlight beams. The road was poorly lit, and she had dark clothing, no reflectors on the bike and no lights.

I don't see how I could have stopped or swerved in time to avoid her in that brief window.

Believe me, I don't care for self-driving cars at all, but I have to remain unbiased here because I know I would have hit her in the same situation.

Be safe out there, people. Put lights on your bike or yourself when you're out there on the road at night.

Re:I probably would have hit her

[quantaman]

Based on the video I saw, she was practically invisible until she entered the car's headlight beams. The road was poorly lit, and she had dark clothing, no reflectors on the bike and no lights.

I don't see how I could have stopped or swerved in time to avoid her in that brief window.

Believe me, I don't care for self-driving cars at all, but I have to remain unbiased here because I know I would have hit her in the same situation.

Be safe out there, people. Put lights on your bike or yourself when you're out there on the road at night.

Then you should drive slower because I would have avoided her.

When I drive at night I drive at an appropriate speed so I can stop in time if my headlights detect something on the road ahead of me. And if my headlights and eyes were as terrible as the crappy video we've been shown I would have been driving very slowly indeed.

Re:I probably would have hit her

[Carewolf]

I would not have. I have been in situations like that many times. The low light is far enough that you have time to avoid stationary objects in the road, or things moving slowly onto the road.

Re:I probably would have hit her

[Luthair]

Its hard to be certain, the video is rather low quality and typically cameras struggle to capture image at night. Even in the low quality video I saw on BBC's site you can see white shoes moving in the shadow which makes me suspect the person was more visible than the video would have you believe.

Perhaps more concerning - the video released of the person supposedly monitoring the car spent an awful lot of time looking down not ahead and out the window.

Re:I probably would have hit her

[Sperbels]

You don't just cross roads like that without looking while wearing black in the middle of the night.

I'm not even sure how the pedestrian missed the car. It's not like there was a lot of other things to look at in the darkness besides the very obvious headlights coming quickly at you.

Re:I probably would have hit her

[AlanBDee]

But consider this. Next time you're a passenger at night on a poorly lit road take out you phone and record the road. I'll wager that your real eyes can see better in the low light then your phone or the camera attached to the Uber car.

Morally, I think the woman is at fault for crossing the road in the dark without looking for oncoming cars and not having any kind of light. But the autonomous car's other sensors should have picked her up anyway. To be successful autonomous cars need to be significantly better then the average driver, they need to be better then the best drivers out there.

I am bias, I can't wait for autonomous cars to come to market. But even I have to admit that it should have seen her coming with plenty of time to spare.

Re:I probably would have hit her

[TexasDiaz]

When you live in rural North Georgia or rural New Hampshire for as many years as I have, there's one rule you learn - don't swerve to avoid obstacles in the road (namely deer). You often times will kill yourself trying to avoid the deer much more than you would if you just hit it. So I'm trained, if something jumps out at me, I'm gonna get ready to hit it. A human driver may have been able to swerve out of the way if they were alert and ready to perform the necessary maneuver, but I'll be goddamned if I'm gonna do that. I'll do everything in my power to not kill something, but I won't kill myself trying.

Re:I probably would have hit her

[ChrisMaple]

Speed limits are set according to fixed rules that have been set by carefully examining statistics and the theoretical capabilities of cars and drivers. A self-driving car would be obeying the speed limit (well, this is Uber, maybe not). A human driver would assume that driving at the posted limit was safe for all but the most severe conditions (dense fog, or heavy snow, icy road, and night).

A cyclist crossing the road on foot, wearing dark clothing, should be able to see approaching headlights from hundreds of yards away. This seems like a case of extreme bad judgement on her part.

Re:I probably would have hit her

[xevioso]

And I think this is actually kindof the crux of the problem here...we are now at the spot where we are going to expect vehicles to decide, either hit that person, or swerve or brake and injure the driver or someone else. Those sorts of split second decisions people rarely have to make, but they *do* occasionally have to make them.

I'm of the opinion that we use applied philosophy...that is, given a person the option of two different types of software in their cars, and they have to decide which one they want.

1) Given the choice between the car hitting a pedestrian, or taking an action that will injure the driver or cause other, unintended consequences, you choose the vehicle that would hit the pedestrian, or,

2) Given the choice between the car hitting a pedestrian, or taking an action that will injure the driver or cause other, unintended consequences, you choose the vehicle that would make the chouce that would injure the driver or cause other unintended consequences.

That way, liability is limited as you made the choice at the beginning.

Re:I probably would have hit her

[AmiMoJo]

It's the sensor failures that really worry me. Radar should have seen her, the lidar should have seen her. The cameras should have seen her - most autonomous cars use cameras with some IR vision capability so they can see at night.

The cameras on my Nissan Leaf have better night vision than the one in the video, which makes me think it's not the one the system uses.

Re:I probably would have hit her

[reboot246]

Finally, we hear from an experienced human driver. I also drive quite a bit at night, and the one thing you have to do is drive within your headlights. If you're going too fast and out-driving your headlights, you're asking for disaster. One trick is turning down your dash lights. It's hard enough to see at night without an extremely bright light in your face.

There are more things out there to hit at night besides humans. I've come close to hitting dozens of deer, bunches of dogs (large and small), a cow, a couple of trees that had fallen across the road, and all kinds of garbage (ladders, chunks of metal scrap, etc.).

Hell, even in the daytime you can be going the speed limit and STILL be going too fast for the road conditions. Ever top a hill at 55 on a 2-lane road and all of a sudden there's a garbage truck stopped in front of you, or a very slow-moving farm tractor, or a road crew? Ever go around a sharp curve at the speed limit and see the road washed out in front of you or a huge log across the road? How about a bridge you've crossed hundreds of times, but never in a heavy rain, so you didn't know that it became as slippery as ice during a heavy rain? Oh, and try all of those daytime hazards at night and see how you fare.

I've driven about 2 million miles. I've probably seen it all and it can be very scary out there. Would I trust my life to a self-driving car? No way in hell.

Re:I probably would have hit her

[LordKronos]

Based on the video I saw, she was practically invisible until she entered the car's headlight beams. The road was poorly lit, and she had dark clothing, no reflectors on the bike and no lights.

I don't see how I could have stopped or swerved in time to avoid her in that brief window.

Then I suggest you try driving by looking out the windshield, and not at a crappy video of whats in front of you. I say that not as a joke. People seem to keep judging this situation by the video we see, but the video quality is pretty much crap. I guarantee the human eye would capture much better detail (both in terms of resolution as well as shadow detail) then what we see in that video. The video is absolute crap, so please don't say what you couldn't have done based on it.

Re: I probably would have hit her

[sexconker]

Thereâ(TM)s a difference between the speed limit and the speed appropriate for the conditions. The posted speed is a limit, not a target.

Not really. Many states have a basic speed law that dictates the speed limit is the safe and appropriate speed. The speed appropriate for the conditions IS the speed limit.

Re:I probably would have hit her

[thegarbz]

Speed limits are set according to fixed rules that have been set by carefully examining statistics and the theoretical capabilities of cars and drivers.

In all countries other than the USA, where they are often set by policy and manipulated politically.

A human driver would assume that driving at the posted limit was safe for all but the most severe conditions (dense fog, or heavy snow, icy road, and night).

Maybe in the USA. In many other countries we are told to "drive to conditions". The speed limit has never been a defense against an accident. Breaching it however has always been a contributory factor. Take for instance the european right of way rules. Literally give way to anything coming from your right. In an average built up area with a speed limit of 50km/h you never are able to get that fast despite that being the limit set. Hell there's a road I go down on the way to work I don't even comfortably take at the posted crawling pace of 30km/h due to the number of blind corners, density of cars and people.

This seems like a case of extreme bad judgement on her part.

That however I agree on. Road rules don't bring you back to life.

Re:I probably would have hit her

[thegarbz]

the video released of the person supposedly monitoring the car spent an awful lot of time looking down not ahead and out the window.

It's truly amazing watching other people drive. You'll be amazed if you take the time to focus on them how little time they actually spend looking forward.

Re:I probably would have hit her

[Cederic]

Maybe you should buy a car that can do more than go in a straight fucking line.

Re:I probably would have hit her

[jeremyp]

If you had hit the mother deer, the baby deer would likely also have died later. Also, an adult deer coming through your windscreen is probably not survivable.

Re:I probably would have hit her

[jeremyp]

Night time isn't ideal conditions, especially not if you really can only see as far as the video makes it appear.

Re:I probably would have hit her

[Green Mountain Bot]

And if you hit something due to overdriving your headlights, then you are at fault for not driving in an appropriate manner for the conditions.

Re:I probably would have hit her

[RespekMyAthorati]

we are now at the spot where we are going to expect vehicles to decide,
either hit that person, or swerve or brake and injure the driver or someone else

No need to make a person-oriented decision.
If the car detects an obstacle, whether a child, an adult , a dog or a traffic cone,
the car just has to solve one problem: is there a free lane that I can swerve into safely?
If so, do it.
If not, hit the brakes as fast and as hard as it can.

Uber's implementation sucks

[reve_etrange]

raises new questions about autonomous-vehicle technology.

No, it raises further questions about Uber's poor, perhaps criminally negligent, implementation. In the last year Uber's had more, and more serious, accidents than I think every other driverless program combined. Google/Waymo has been testing in San Francisco - not Tempe - for years with nothing comparable.

Shouldn't have happened:

[foxalopex]

I actually watched the set of videos and there's two major things to note:

First of all the safety or backup driver appeared to be distracted. Although in all fairness if you're suppose to sit there hours on end without taking an active roll at driving this is probably going to happen. This is why google believes in all or nothing approach, half-baked systems are going to get people killed. While this wouldn't save the cyclist from being hurt, quick reflexes may have saved it from being fatal.

Second, LIDAR works by projecting a super high speed panning laser that maps out the 3D spacial environment. It causes the computer to produce a 3D model of the surroundings. This should NOT be affected by the dark! Unless Uber decided not to use LiDAR which would be a dangerous move. If they're using LiDAR the only explanation is the AI image recognition system failed to recognize the cyclist which is weird considering an object that BIG moving should register as a collision threat. Google has noted that in their own self-driving program the computer can sometimes panic over a flying piece of newspaper while a normal driver wouldn't because it looks like an object heavy enough to threaten the car.

Re:Shouldn't have happened:

[Anonymous Coward]

Uber do use LIDAR. Looks like there was a hardware or software issue. The car was also breaking the speed limit at the time.

Re: Shouldn't have happened:

[cyber-vandal]

The driver looks like they have consumed many rolls but are not at all active

Re:Shouldn't have happened:

[Solandri]

LIDAR (and radar, and sonar) is one of those things which sounds great when you consider the car in isolation by itself. But isn't so great once you have multiple cars on the road all using it. I'm already noticing the problem with the sonar-based parking sensors. Despite using CHIRP (sonar frequency which varies over time), there's still enough random overlap from nearby cars that the parking sensor will occasionally trigger due to other cars which also have parking sensors. Usually it's at a red light when a lot of cars are bunched up. But I've had it trigger during normal driving and even a couple times on the freeway.

A passive system, like generating a 3D model of the surroundings based on the slight parallax between two eyeballs, does not interfere with other similar or identical systems sharing the road. So unless someone can come up with a bulletproof way of preventing two laser / radar / sonar systems from interfering with each other, the end goal should be an entirely passive environmental mapping system.

Re:Shouldn't have happened:

[kiviQr]

I wonder if they removed LIDAR after Waymo lawsuit?

Re:Shouldn't have happened:

[mjwx]

I wonder if they removed LIDAR after Waymo lawsuit?

I doubt it, LIDAR is likely to be someone elses tech bought off the shelf. Not sure about Uber but Alphabet (Google) uses Helodyne units which I've used for aerial terrain survey. Phenomenally accurate units except if its raining, snowing or there's cloud in the way.

That's video

[Anonymous Coward]

....she was practically invisible until she entered the car's headlight beams

Human vision is MUCH more sensitive than cameras. What looks dark in the video wouldn't be so bad to a human. That's why they use all those lights when shooting video.

So, it wasn't as dark as it appears.

Re:That's video

[K. S. Kyosuke]

Yes, that's why night vision devices use extracted human eyeballs instead of optoelectronic components.

Re:That's video

[Zaelath]

Night vision has amplification, this camera did not.

What the car should have had was infra-red and if it doesn't then I can't see how you can suggest it's fit to use at night in any conditions.

If it did have infra-red she would have been a massive bright spot on a black background moving across the cars path, and it reacted by ... doing nothing.

Why didn't SHE see the car?

[petes_PoV]

I watched the video and it looked to me like (in the visible spectrum at least) she literally appeared out of the shadows less than a second or two before she was hit.

But the other question would be why didn't she see the oncoming vehicle. It had its lights on and even coming around a bend, the light it threw onto the roadway would be visible long before the car itself appeared.

Even if one party in a collision is not at fault, that doesn't mean they couldn't have avoided it.

Re:She fucking Darwined herself

[Carewolf]

Note, the place is apparently a pedestrian crossing zone just without zebra stripes. You see the marking if you check Google Streetview.

This is why I ready /.

[Prien715]

Kudos to everyone in the last story who commented LiDAR being able to see the pedestrian and the crash being totally avoidable. Comments have also been more accurate than the news in the recent Intel and AMD (non-story) about security.

The fact that the highly moderated comments is more accurate almost any news outlet is why I keep coming back. That and I'm *still* looking for Natalie Portman's brand of hot grits.

Re:False Equivalency

[viperidaenz]

Clearly we need more people to die before the statistics can be compared.

People need to die

[Firethorn]

That's actually true, it is the way statistics work.

Let's say that AI cars have a 'true rate' of 100M miles between fatal accidents, somewhat better than human. Or they could be 50M miles, somewhat worse. The technology is still in development, so who can say.

The fatal accident can happen anywhere in that 100/50 million miles. It could happen mile 1. Mile 20M. Mile 99,999,999.

Spreading that even more, you could have 200M miles with 2 deaths, and have both happen in the first 100M, second 100M, etc... It's really 1/100,000,000th chance of death per mile. You could get multiple deaths, or no deaths.

The only way to nail the real rate down with any degree off accuracy is to have multiple occurrences. Which in this case means more deaths.

Extremely boring

[140Mandak262Jamuna]

I mentioned it earlier. The idea of a human being in the driving seat would be alert enough to override the autonomous mode is really really stupid.

If there is no need to steer, the attention wanders and it is impossible to stay alert. This was discovered almost 100 years ago in the railroads. The engineer had the exacting task of watching for grades and monitoring speed, especially those days with weak steam locomotives that responded very slowly. Still they would get bored and fall asleep. They invented the dead man's treadle. The engineer must keep it pressed, or the locomotive will stop. Even now there are various techniques to check and keep the engineers alert on railroads.

With that much of history, it is stupid for autonomous cars to just leave the driver there. They should have active devices that do challenge and response to make sure the human operator stays alert. Else it is a waste to put a human being there.

Re:Human pedestrian could have avoided fatal crash

[Kjella]

Seriously. Pay attention when crossing the road, especially at night.

Seriously. Pay attention when driving a car, especially at night.

Unfortunately paying attention as a driver is no immunity from getting hit as a pedestrian. It doesn't matter if I have the right of way, I'll be the one injured, crippled or dead. And when you know that by far most of the adult population have a driver's license so when you're scraping the bottom of that barrel there's some pretty terrible drivers out there. Looking out for yourself is simple self-preservation, not matter how much the rules say you shouldn't have to.

Re:She fucking Darwined herself

[dabadab]

It looks like a crossing (and it certainly should be) but it's actually not:
streetview

Re:Better

[chubs]

So long as it hits fewer ladies with bikes than humans do, yes. It doesn't need to be perfect to be ready for adoption. It just needs to be better on average than humans. If 1/10 fatal accidents involving an autonomous vehicle could have been prevented by a human driver, while 2/10 fatal accidents involving human drivers could have been prevented by autonomous vehicles, then the obvious choice is that we should obviously switch to autonomous vehicles for a total reduction in fatal accidents. I absolutely call that progress. Yes, it's going to suck for those 1/10 people who die as a result, but it'll be huge for those 2/10 people who had their lives saved.

Speaking from experience

[DrYak]

Speaking from experience :

A.
I drive quite a lot when on vacations/week-end, including often on nights, including sometime in fog.
- The human supervisor *should* have turned on the high beams. It seems to me that only the low beams were active, reducing the visibility range. (This might have affected the camera part of the sensors). The super visor is supposed to supervise the self-driving car and thus should be able to see in order to anticipate and compensate bugs, instead of relying the whole thing to work.
- The human supervisor *should* have instructed the car to drive at a speed within the supervisor's visible range (with the low beam only, the visibility is extremely short, the speed should have been kept low).
- The human supervisor *should* have kept eyes on the road (Uber is testing a new technology, bugs are bound to happen.

Definitely the supervisor was doing a couple of things wrong. But even if all the above were followed, that probably wouldn't have saved the bike rider.

B.
I bike to work almost every single working day (welcome to europe), I'm used to bike at night, etc.
- She was wearing dark clothes. It's not a problem per se but you have to keep in mind you're a bit less visible. (Some people here around even where reflective jackets when biking).
- She didn't have any reflector on the bike. That makes the bikes drastically less visible. Usually most bike riders have a good quantity of retroflective reflectors on the bike (plastic on the wheel spokes, sticky bands on the bike body), etc.
- She had absolutely no light. That's a very high danger of collision. Enough for cops here around to pull you if they catch you with an unlit bike. Nearly every one will use a good battery-powered headlight/taillight, often a blinking tail light (legally dubious but every one use them for visibility and police tolerate them). Some bike rider look for "always on" solution to be better visible (magnetic induction tiny lights that are at least visible, even if not very good at lighting. Or on-axis dynamo that powers good lights). Some almost turn their bikes into christmas-tree like light shows just for lulz.
- No helmet (could help diminish the results of an impact, also most of them have reflectors, and some even feature built-in lights).
- She should have seen the headlights from far away. If she counted on the car slowing down, standard practice (massively advertised at the beginning of each school year here around) require to establish eye contact (to make sure that the driver has seen you and will slown down) which is impossible to do by night.
- In the absence of eye contact, she should have waited for obvious signs that the driver will slow down (e.g.: the driver already starting to slow down and blinking the high beams to ack).
- Other wise she should have assumed to not having been seen (specially given the clothes she's wearing).

With my experience with bike I would never attempt to cross the way she did given all the above. That looks absolutely suicidal to me, there's now sane way to expect even a well behaving driver to have avoided the collision.

There must be some reason for her absolutely not paying attention : - being way too much absorbed into some phone conversation over earphones ? - emotionally distressed and not able to focus and pay attention ? - drunk ? (which here around could be a reason for the police to take away your *other* licenses : car driver, boat, etc.)

Re:There was no dilema

[Cederic]

I will never swerve on a two lane road. That's a conscious decision I've already made.

You're a fucking shit driver then, get off the road.

People swerve safely a lot. A hell of a lot. Often at high speed. They know how their car handles and they maintain situational awareness that lets them know whether the other lane is clear.

As an example,

It can easily get you killed, through a roll, head on, or into a tree, building, fire hydrant

If you can't swerve without rolling your car, learn to drive.
If you can't tell whether there's an oncoming vehicle that makes swerving dangerous then you don't know what's happening on the road. Learn to drive.
If you can hit a tree, building or fire hydrant by swerving into the empty lane next to you, you've definitely done it wrong.

In short, you need to learn to drive.