Police Release First Video From Inside the Uber Self-Driving Car That Killed a Pedestrian (recode.net) 698
An anonymous reader quotes a report from Recode: Three days after an Uber self-driving vehicle fatally crashed into a pedestrian in Tempe, Ariz., police have released video footage of what the vehicle saw with its cameras moments before running the woman over, and what happened inside the vehicle, where an operator was at the wheel. The video footage does not conclusively show who is at fault. However, it seems to confirm initial reports from the Tempe police that Herzberg appeared suddenly. It also showed the vehicle operator behind the wheel intermittently looking down while the car was driving itself.
Yeah, it was her fault (Score:5, Insightful)
That said, I'm glad I was correct in my knowledge of what those "safety drivers" actually do all day.
Re:Yeah, it was her fault (Score:5, Interesting)
The scariest part is that the pedestrian does not react to the car at all before being struck. As if she either gave no effort to check for oncoming traffic, or if she just had the mindset of "I have the right of way, the vehicle will stop for me".
We'll never know what her motive was for crossing at such a poor time, and it's a tragedy that this happened, but her choice to cross there was baffling.
Also the driver was "intermittently" looking down? No, the driver was looked up twice for a brief moment twice in the video with very long periods of staring down. This may have been unavoidable regardless, but until self driving cars are more reliable, taking your eyes off the road like this is not a good idea.
Re: (Score:2)
The scariest part is that the pedestrian does not react to the car at all before being struck.
I think the woman must have been very confident that the car would have seen her and stopped.
Re: (Score:3)
The scariest part is that the pedestrian does not react to the car at all before being struck.
I think the woman must have been very confident that the car would have seen her and stopped.
That's because a human driver would have seen her and swerved. She was more than halfway across the lane, after all. I've missed many pedestrians in similar conditions and similar positions in the lane. Lots of people have.
Re:Yeah, it was her fault (Score:5, Insightful)
Re:Yeah, it was her fault (Score:4, Insightful)
It is actually confusing that the car did not do an emergency break.
Would still have hit the woman, but she perhaps had survived.
Hard to judge, but I think she showed up in the light at about 15 yards distance. The car was driving about 35mph, over 15 yards it should have braked below 30mph ... not sure if that had helped much as that is roughly equivalent to a drop from 25 yards height (I'm to lazy to calculate it exactly).
Comment removed (Score:5, Informative)
Re:Yeah, it was her fault (Score:5, Informative)
The typical response/reaction time is 0.3s
That's only true when you're anticipating an event, with max attention, and have already mentally prepared your response. Response in normal driving circumstances is much slower:
http://www.croberts.com/respon... [croberts.com]
McGee et. al. (1) reported that perception time is the sum of eye movement time, fixation on the hazard time delay, recognition time delay and muscle response delay time. They found that for the 85th percentile of drivers, eye movement delay was 0.09 seconds, fixation delay time was 0.20 seconds, recognition delay time was 0.50 seconds, decision time 0.85 seconds, muscle response delay was 0.31 seconds and brake reaction time was 1.24 seconds. The sum total of these times, the response time, was 3.19 seconds. The 85th percentile is often chosen as the upper bound for design analyses
Re: (Score:3)
Human response time, yes... but that will never address the challenge here. If Uber’s neural network needs more than 0.1s for decision making then it never should have been allowed on the road. Add to that the fact that their initial detection and action window should have been significantly longer with LIDAR ranges, and you appear to have a system that is not road-worthy.
Re: (Score:2)
Stole that part from Google.
Had to start over, and that parts hard.
Re: (Score:2)
A human simply cannot transition from "not driving for hours on end" to "driving" in a split second. It's been studied quite a bit. Having a human at the drivers' seat in a Level 3+ autonomous system is just safety theatre. Even at Level 2 where you have steering wheel holding requirements and sometimes driver attention montors, and the system doesn't try to handle many common traffic features (so the driver doesn't have the option of inattention), distraction is a real concern. With Level 3+, it's unavo
Comment removed (Score:5, Informative)
Convinces me Uber is at fault because of 1/R^4 (Score:4, Interesting)
I had heard reports that the video showed her popping out of no where. Absolutely that is not what it shows. It shows here suddenly coming into the headlights lit region, not appearing from behind a bush.
What's the cardinal rule of driving at night or snow storms? NEVER outdrive the range of your headlights. That is, your stopping distance abolutely positively has to be within your range of sight. Anything else is completely irresponsible.
So that's clearly what happened here. The woman in the video just appears like magic in a couple of frames from dark to in the head light. That means the leading edge of the headlight zone was something less than 1/2 of a second from impact. No way can you stop in that time.
This is defacto outdriving your headlights. Uber is guilty. Case closed.
Now moving on to technical details this also shows. I think part of this is that the dymanic range of the camera sucks. I am fairly sure my own eyes would have been able to see further into the dark. Those black pixels ar not just dark they are completely saturated on the dark end. Nothing is resolvable in them which is why the appearance time is so short. This is a serious problem for all systems as the dynamic range of most cameras is very limited, especially when were dealing with 1/R^4 light fall off ( 1/^R^2 light outbound and then 1/R^2 reflected. Thus a 256 bit sensor is effectively a 16 bit dynamic range sensor. And if you were to account for glints and such then it's even less. No wonder she pops out.
Secondly, where the hell was the lidar here? Shouldn't that have spotter her?
Uber is flagrantly at fault.
Re:Convinces me Uber is at fault because of 1/R^4 (Score:4, Insightful)
The video appears to be deceiving. It is almost like it was purposely dimmed before being released. A human behind the wheel would be able to see much more then what is shown on the video. Look at the buildings in the background, the ditch further down the road... it is all black. No more then 50' from a street light and everything is black. The human eye is so much better then that. If the driver was watching, he would have seen her. Any video system should have also been able to see her. Uber has no excuse - the cyclist was technically at fault but the Uber car should never have hit her. The car never even slowed down.
Deer are harder to see then a cyclist with reflective shoes - most drivers would have avoided a deer in this situation.
Re: (Score:2)
Re: (Score:2)
yeah, the video only shows two white stripes ahead of the driver on the left but much more on the right.
You also have to notice that Uber's plan for driver attentiveness appears to have allowed quite a bit of inattention.
Re: (Score:3)
What's the cardinal rule of driving at night or snow storms? NEVER outdrive the range of your headlights. That is, your stopping distance abolutely[sic] positively has to be within your range of sight. Anything else is completely irresponsible.
Objects approaching a traveling vehicle from the side can intersect the path of that vehicle nearer than the furtherest projection of the headlight beam. This is a case where your rule fails to protect against night-time collisions. In the linked video the homeless woman with the grocery cart approaches the path of the oncoming Uber car from the side, not from the front.
I once I collided with a raccoon in my rx-7 because, though I was traveling at a safe speed, the raccoon ran in front of my car from t
Re:Convinces me Uber is at fault because of 1/R^4 (Score:5, Insightful)
Or not. http://www.ncsl.org/research/t... [ncsl.org] in Arizona "Pedestrians must yield the right-of-way to vehicles when crossing outside of a marked crosswalk or an unmarked crosswalk at an intersection."
Re: (Score:2)
Right-of-way doesn't make it legal, it's still jaywalking.
Re:Convinces me Uber is at fault because of 1/R^4 (Score:5, Funny)
"Pedestrians have the always have the right of way."
We'll engrave that on your tombstone.
Re: (Score:3)
What is relevant is that the pedestrian was not paying any attention whatsoever to traffic.
Re:Convinces me Uber is at fault because of 1/R^4 (Score:5, Insightful)
Re:Yeah, it was her fault (Score:4, Insightful)
The cyclist should have had wheel reflectors and front and back lights at a minimum, as well as reflective clothing.
This does not, however, mean the driver -- or should I say human attendant-- is not at fault as well for (apparently) texting. This kind of road is where you need to be especially alert because of the combination of poor lighting and high speed.
Re: (Score:3)
High speed is speed in which things happen faster than you can react effectively to. So it depends on context.
This is something I tried to drill into my kids before they learned to drive: absolute speed is not a reasonable measure of safe speed. On a dry interstate around noon 90 mph would not be unsafe if other cars are traveling at that speed. If it is icy, 40 mph maybe too fast.
Anytime you cannot see the road surface ahead you need to slow down so you can react to something outside your vision. Thi
Re:Yeah, it was her fault (Score:5, Interesting)
You're kidding right? It is impossible to tell from that video.
WTF happened to LIDAR and sub millisecond braking reactions? The woman stepped out of a shadow at a point where a human would've struggled to brake hard enough to stop, but a machine should've been able to sense via lidar an object moving ACROSS ANOTHER LANE in a trajectory that would end in front of it fast enough to at least brake enough to turn a death into an injury.
I don't think this is a problem with autonomous cars in general, but a problem with Uber's 'I got mine, fuck everyone else' mentality towards everything. I doubt they're prioritizing pedestrian safety whatsoever.
Re:Yeah, it was her fault (Score:4, Insightful)
The car obviously had no lidar.
The lady was about 15 yards away ... total distance to brake from 35mph is about 20 yards in perfect conditions (not counting reaction time, which would eat already 10 yards), and 40 yards in general.
I don't think this is a problem with autonomous cars in general, but a problem with Uber's 'I got mine, fuck everyone else' mentality towards everything.
True. A car without LIDAR and various RADARs and ultrasonic sensors for road texture is not really self driving ready. This car basically only had an auto pilot, lane detection and sign detection. Pedestrian detection failed due to bad light/camera conditions.
Re: (Score:2)
Re:Yeah, it was her fault (Score:5, Informative)
The pedestrian didn't seem to notice the car. The car appears to be a Ford Fusion (probably hybrid). If it was in charge-sustaining mode, the car might have been very difficult to hear.
The pedestrian was wearing a yellow hat and pushing a pink bicycle. Her shirt was dark though. She was visible to the camera for only about .77 seconds prior to impact. A human being would take 0.5-2 seconds to react to the object in the road once it became visible. Depending on the human, the pedestrian might have been visible for a couple of seconds longer than we see her in the footage, but the safety driver appeared to be distracted.
The reaction time of the autonomous car should be milliseconds. Assuming that the dashed lane markers are fairly evenly spaced, the car doesn't appear to have decelerated at all from my perspective. According to the police, the car was traveling 38 MPH, or roughly 61 km/h. On dry pavement with decent tires, the stopping distance in meters without accounting for any reaction time should be about (s^2)/(250*.8) with s = speed in km/hr... so, about 18 meters, or to be generous, 60 feet.
See https://korkortonline.se/en/th... [korkortonline.se] .
Judging from the aerial layer on Google maps, the distance between the beginning of a lane marker and the beginning of a subsequent lane marker is 30 feet or so. From this, I think the first time you see the victim in the video she's about 43 feet away (.77 seconds at 38 MPH).
Here's the thing though... the LIDAR should have seen this in time to at least swerve to avoid. The LIDAR should also have seen the victim before the victim was visible in the headlights. In my state, the driver has the responsibility to swerve to avoid even if there isn't enough time to stop. It's obvious that there was nobody in the left lane (even in the blind spot, which isn't blind with LIDAR).
This really seems like an example of where an autonomous car could have saved a life that would have been lost due to a human driver's natural limitations, but it failed to do so. The car should have been able to see hundreds of feet, and the car should have had practically zero reaction time. Just as you would be lenient in judging and older driver for longer reaction times, I think we should hold the autonomous car to a higher standard.
This thing was a test vehicle. The debug-level logging of the incident should be made public so that if there was a bug that killed this woman, the truth will be known.
Re: (Score:2)
Re: (Score:3)
All may be true, but you simply do NOT step into a dangerous roadway where cars are whizzing by at 40MPH without looking and listening, unless you have some serious mental issues of the self destructive kind. The car would have been visible to her with it's blaring headlights and likely would have been pretty noisy traveling at nearly 40MPH, even if it was totally electric powered. All she needed to do was STOP, look and listen, but she just steps out.
I know I learned the "Stop, look and listen" procedure
LIDAR (Score:5, Insightful)
Re: (Score:2)
This is a good example of why visual sensors are insufficient for autonomous driving.
Uber uses LIDAR. Of the major SDC companies, only Tesla does not. Tesla is camera-only.
I have no idea why the LIDAR didn't work to detect this woman. From the video, it looks like the car didn't brake at all.
Re: (Score:2)
Re: (Score:3)
Tesla has video (8 cameras), radar (up to 160M), and long-distance (8M) ultrasonic sensors.
Re: (Score:3)
Tesla’s system has also murdered two people.
Both of those deaths should have been preventable with just cameras. They were basically software failures.
This Uber death doesn't look like it was preventable with cameras only.
Re: (Score:2)
1) Tesla has far more miles on Autopilot than anyone else has on their systems. How many miles today isn't clear, but as of late 2016 there were 300 million miles in active mode and 1,3 billion in shadow mode. Today, those figures are probably more like 1B and 4B. Waymo, by contrast, has driven only 5m miles, and Uber a lot less than that.
2) Only one fatality has been confirmed to be under Autopilot. Concerning the case in China, they refused to let Tesla examine the logs, so all you have to go on is the
Re: (Score:2)
Why?
Humans use visual sensors exclusively for driving, humans would have had the exact same accident.
Trying to say that autonomous cars cannot have any accidents is silly and is an arbitrarily high bar.
I'll be plenty happy if they are just 2 - 3x less likely to have an accident compared to human drivers. That would be a massive improvement to society.
Re: (Score:2)
Humans use visual sensors exclusively for driving
Worldwide, human drivers kill 3500 people per day.
Trying to say that autonomous cars cannot have any accidents is silly
Saying that SDCs should do no better than humans is silly too.
I'll be plenty happy if they are just 2 - 3x less likely to have an accident compared to human drivers.
We should be aiming higher than that. This accident should have been preventable by a properly implemented Radar or Lidar system.
Re:LIDAR (Score:4, Insightful)
That's part of the crazy thing.... she was pushing a bike, radar should have seen her too.
LIDAR should always see pedestrians, easy. But when you're pushing a bike - large object made of interconnecting angular metal structures - across the road, it should be a glowing beacon to radar.
I don't know what sort of junk system Uber has implemented, but it clearly should not be allowed on the road without an audit.
Re: (Score:3)
Re: (Score:2)
Re: (Score:3)
Humans use visual sensors exclusively for driving,
Hmm, no. They use sound too, and you'd be amazed how much you use whatever the fuck the sense is that tells you whether you're moving or not - and not just in the direction of the car.
I sure as shit can't see that my rear tyres have no grip, I can, erm, sense it. I can't see that my brakes have locked, I sense it. I can't see that I just hit a woman with a bike, I was looking down at my phone and only sensed it.
humans would have had the exact same accident.
Possibly. The video footage shared is very inconclusive on that front. It doesn't sufficiently sh
Re: (Score:3)
I've been in this exact situation twice, where someone dressed in black decided to cross a darkened road directly in front of me. In both situations, I had to brake hard to prevent hitting them.
The tip-off was that I noticed lights blinking out ahead, due to something occluding them. It was an extremely subtle effect, one I would have missed if I hadn't been paying full attention, and one which I do not think AI is capable of recognizing.
Simply put, I doubt that computer based vision will meet the capabilit
Re: (Score:2)
The truth is, a good driver would not have hit that pedestrian.
If autonomous cars are not as good as a good driver (but only good as an average or below-average driver), the whole safety argument for autonomous cars goes out the window. Uber must pay the consequences for this accident, as the video clearly shows that Uber's self-driving car did not meet the safety standards that all self-driving cars ought to meet.
Re: (Score:3)
Re: (Score:2)
Except, of course, we are not talking about introducing autonomous pedestrians into the world. Burden is on the autonomous car.
Re: (Score:2)
Re: (Score:3, Insightful)
> What kind of sensors do you use to drive?
Better than that the shitty sensors Uber uses apparently.
If all they use is visible light, they are years behind human eye during night.
Re: (Score:3)
I can't find the setting. Can you tell me the page of TFM that tells me how to expand the iris?
Re: (Score:2)
Re: (Score:3)
More data is often not the solution, but the problem.
Correlating different data sources is NOT an easy task and as you add more data sources it becomes an N-squared problem, which takes geometrically more processing to sort out. Processing also takes time. In real time processing systems the required response time is set, so adding a new data source can often lead to response time problems, which software engineers often solve by moving parts of the process up to a less responsive priority level.
In this c
Wait, explain LIDAR again? (Score:5, Insightful)
Yes, this was hard to see using passive techniques with visible light (ie, your eyes), but WTF, the person wasn't sprinting or jumping off the curb, something active like LIDAR should have had no troubles spotting this.
Re: (Score:3)
Yes, this was hard to see using passive techniques with visible light (ie, your eyes), but WTF, the person wasn't sprinting or jumping off the curb, something active like LIDAR should have had no troubles spotting this.
I would think radar should have spotted her & the bike from a ways off, too
Re:Wait, explain LIDAR again? (Score:5, Informative)
The dynamic range of human eyes is much greater than a camera. It was not pitch black outside to a human. I have lived in Tempe and the ambient light of the city would be enough for at least minimal night vision to apply. This is the reason why you drive in a darkened vehicle without your dome lights at night, for your night vision to be effective. Texting on your phone in a part of town where there are a lot of people roaming the streets (such as south Scottsdale Rd) is simply a negligent thing to do.
About the rhetoric (Score:2)
Re: (Score:2)
Operator inattention did not cause this accident. Although he looked down several times, he was looking at the road when the woman appeared in the headlights. There was not enough time to react.
Re: (Score:2)
Re:About the rhetoric (Score:5, Insightful)
"Operator inattention" is absolutely the ultimate goal of automated cars.
If the human has to pay attention, then the human might as well drive, and the automation is pointless.
Scary that the pedestrian doesn't even look (Score:5, Insightful)
Re: (Score:2)
I drive through Berkeley all the time where pedestrians routinely step into the street when I have the green light.
But that doesn't mean I can actually run them over and not suffer any consequences as Uber is apparently about to.
P.S. And a real, human driver learns to drive more defensively in areas where they suspect pedestrians might behave more unpredictably. Apparently AI isn't "intelligent" enough to have that sense.
Re: (Score:2)
Having seen the video now (Score:2)
I stand by my post in the last topic :
https://slashdot.org/comments.... [slashdot.org]
(You people are CRAZY with your laws and cars, pedestrians should NOT have right of way at all times as it seems you do, or at least most of you behaved while I was there)
That being said, someone speculated this was right near a late night club in a boring area, only fun thing to do is drink there? So maybe drunk.
I'll tell you a few things, they crossed in the WORST spot, just IN the dark part after some light, holy crap does she come o
Re: (Score:2)
You are quite correct - the pedestrian should have been on the lookout, for their own safety at least even if the law might be on their side (though I am not sure of the state and local laws there). Moreover they could have been wearing a reflective vest or helmet, had reflective strips on the bike, not been wearing a black shirt, had a headlight on the bike, etc... things that may not have made a difference here, but are generally good ideas (and in some places legally required) when using a bicycle at nig
Re: (Score:2)
video ! = eyes
I think that's part of the issue. The unlit street portion in the video was too dark (zero information) and I'm guessing not likely what a typical driver would've seen. To be driving at that speed with zero information of what's in front of you at that short distance is a recipe for trouble. If the victim wasn't inebriated into total indifference it's possible she assumed she was visible to the driver who wouldn't dare hit her regardless of who had right of way.
How many people have Self-Driving cars Saved? (Score:2)
I am sure there are many instances that self-driving cars have braked and saved lives that otherwise would have been an injury or death. Such incidents never make to the front pages.
Expected (Score:5, Insightful)
Exactly what I expected to see....
Someone walking a bike.
At night.
No streetlights.
No backlighting at all.
Wearing black top and dark pants.
With no lights at all on the bike.
No lights on the person.
Not in a crosswalk.
Apparently not looking.
About 2 seconds of visibility.
The pedestrian is almost 100% wrong in every possible way. I don't see how this could be ANY human driver's fault, had a human been driving. As for autonomous, I guess it depends on what sensors. Could their system have had an infrared camera or other sensor that could have seen the wreckless pedestrian sooner than was evident in [human] visible light? That would have been nice. But does that make the pedestrian less at fault? I think not.
Re: (Score:2)
Re:Expected (Score:4, Insightful)
>"Yet you can see a person on a bike. So should have the car."
Yes, I saw the person for maybe 2 seconds in the video. Had it been that second I was checking speed or a mirror, it would have been less. And I might have had time to brake or swerve. And swerving might have made it worse. But just seeing the bike in 2 seconds doesn't make it the vehicle's fault.
Oh, you might think "well, if it were a car in front of you, and you followed the 2 second [following distance] rule, you should be able to stop in time". And I would agree... BUT the car would have tail lights AND probably brake lights and I would have already known it was there and from far, far away. AND it would be in a fairly predictable location with fairly predictable actions. In such a case, yes, I would be at fault as the rear-ender. And yet, same scenario- if that car in front at night had NO lights and NO brake lights, it would immediately shift to being their fault. And that is without that unlit car coming into view at the last few seconds FROM ACROSS A MEDIAN!
But I *do* agree that an autonomous car with lots of high-tech sensors should have been able to "see" what was happening [beyond human visible light] sooner and at least tried to brake. Still doesn't mean the car is at fault.
Re:Expected (Score:4, Interesting)
Re:Expected (Score:5, Insightful)
The difference, of course, being that an actual human driver would have actually been watching the road (imagine that) and would have, when finally seeing the pedestrian, (a) swerved; (b) slammed on the brakes, and/or (c) most likely, both, rather than plowing into her at full speed while mouthing "oh shit" after having finally looked up from staring at a smartphone in their lap. That difference might well have left her just seriously injured rather than dead.
It's not a perfect world. "SHE DIDN'T HAVE THE RIGHT OF WAY" doesn't come even close to excusing (a) an insufficiently designed guidance system paired with (b) an unbelievably irresponsible "safety driver."
Evolution at work ... (Score:5, Informative)
... however the test driver did not really pay attention.
Being test driver is obviously a fucked up job. 99% is killing time and 1% is killing time.
In Germany there is not one test driver but 3 ... one who would react if something goes wrong and 2 to write protocols about notable stuff.
In this case it is notable that the lights are configured incorrect. They barely shine 15 yards ahead, that is definitely wrong, and a driver or the automatic driving system should adjust speed to about 1/3rd of what it was driving.
The reality of the situation (Score:2)
Distracted operator (Score:2)
Yeah, he’s there baby sitting the new tech but, assuming the vehicle didn’t slow down, could he have intervened if he wasn’t distracted? He appears to be reading something since he smiles after looking at whatever he’s holding below the camera view. My money is on that guy getting an NTSB finger pointed at him.
Doesn't look good (Score:5, Insightful)
True she comes out of nowhere on the video, but that's a really crappy video. She was walking slowing and already in the car's lane when the headlights hit her, even if she had been stationary the result would have been the same.
Of course a human driver could have hit her as well, but I suspect that most often a human driver would have seen her far enough ahead to stop or at least swerve enough to avoid her (of course most Ubers might have as well).
I'm curious if that's the only video available since decent cameras are not that expensive, and I'd expect the car to have several cameras at different contrast levels.
Re: (Score:3)
True she comes out of nowhere on the video, but that's a really crappy video.
The video released was released in a way to make Uber look as good as possible.
Infrastructure is the problem (Score:2)
I maintain that a 2 mile stretch of road with a limit of 35 (previously 45mph) thatâ(TM)s eight lanes across should have more than one crosswalk, and probably shouldnâ(TM)t exist in this form at all. I hate Uber and the empire theyâ(TM)ve built on the backs of the working poor, but city planning has to modernise with our tech. The real wonder here is that people arenâ(TM)t killed on that road CONSTANTLY.
Re: (Score:3)
Yeah, they should build a pedestrian bridge across it, that way no one would ever get killed.
oh, wait.
Pretend it wasn't a self driving car... (Score:5, Interesting)
Suppose that this was not a self-driving car. You see a video of a driver spending 50% of their time looking down at a (phone, book, video game, etc.) and 50% looking ahead. They look ahead, and suddenly get an OH SH*T look and plow someone down. What would the law say?
1) The pedestrian was negligent.
2) The driver was negligent.
This is contributory negligence, and I don't think the driver would get off with no penalty just because the pedestrian was negligent. This cannot be allowed to continue.
So back to the self-driving part: either the driver thought "Oh, it's a self driving car, I'll play a video game" or Uber said "Monitor this status console here on your lap and just look up every now and then to make sure that you don't plow over someone." The police need to figure that out. If it is the former, the law should do whatever they normally do in cases of contributory negligence. But if it is the latter, then Uber needs to lose their license for testing these cars, and face a big fine.
Over a second of notice (Score:3)
That's at least enough time for a real human to begin applying the brakes.
Slowing down by just 5mph would have given the woman a 30% higher chance of surviving.
With real eyes looking and not a camera, you'd be able to see more detail in the shadows. There are street lights there and a human eye has a much greater dynamic range than a camera.
The person behind the wheel looked up and to their left, showing he saw the woman in the other lane before the impact. The camera couldn't see the woman until she was directly in front of the driver's side of the lane, proving a person could have seen her in the shadow where the camera, due to its limited dynamic range, couldn't.
Perhaps Uber should have forked out for HDR cameras.
Re: The Driver was Texting (Score:5, Insightful)
No human driver could have seen that woman in time to stop, but a car equipped with infrared lidar should be able to. Time to update the sensors on the test fleet.
Re: (Score:2)
Re: The Driver was Texting (Score:5, Insightful)
It is astounding that LIDAR failed to see that person. I can't even imagine how they went undetected. I really want answers as to why.
BTW, the concept of a "safety driver" on a Level 3+ autonomy system is just window dressing. Distraction is bad enough on Level 2 systems that mandate hands on the wheel and sometimes involve attention monitors. With a level 3+ system, where the person isn't driving at all, distraction is essentially guaranteed (this has been studied; it doesn't matter who you are, you will get distracted sitting behind the wheel for long periods without actually doing anything). A person simply cannot transition reliably from "not at all driving for hours on end" to "emergency driving" in a split second.
Re: The Driver was Texting (Score:4, Funny)
Maybe she was riding one of those Russian stealth bicycles.
Re: (Score:3, Insightful)
her name was Elaine Herzberg, not that person or the pedestrian , `Elaine was killed by ubers car.
https://www.standard.co.uk/new... [standard.co.uk]
Re: (Score:3)
Yes, it doe
Re: The Driver was Texting (Score:5, Insightful)
1) It's a shitty video, human eyes may have seen a better picture
2) A second and a half of breaking from 38 mph may not have been enough to stop the vehicle, but it would have been enough time to slow down and swerve. The pedestrian may have been hit regardless, but they also may have survived.
Re: (Score:3)
I agree. By the time the jaywalker was visible no one would be able to stop.
She may have been visible to a human much earlier than we can tell from that camera footage. That camera seem to have a relatively narrow exposure range. She was in the the other lane well before camera visible. In the camera picture, the left lane is totally blacked out past the direct field of the headlights. You can see well past that with your eyes.
That the car didn't even appear to slow down (although that is hard to tell as well) is another concern.
Re: The Driver was Texting (Score:5, Interesting)
The radar and lidar certainly SHOULD have seen the pedestrian and it certainly appears that the driver was NOT paying attention. I also will say that in the video I could see the pedestrian while still a way out where the car should have started braking and it could have avoided killing her. While not nearly as noticeable as they would have been had the bicycle had reflectors on the wheels I could still see it when pausing the video.
The pedestrian should have had reflectors on the bicycle wheels. Just the other night I barely saw a bicyclist crossing the street in front of me at a crosswalk until they were in my lights due to the lack of any reflectors and dark clothing. I don't know what the laws are in Arizona, but where I live bicycles are required to have reflectors, a headlight and a taillight at night.
If the driver were paying a lot more attention to the road than the phone then this also could have been prevented.
Re: The Driver was Texting (Score:5, Informative)
No human driver could have seen that woman in time to stop
Had the headlamps been aimed properly, they could have. In the video, when the car is traveling at 38 mph (56 feet/second), it takes about 1.5 seconds between the time the pedestrian came into view and when the collision occurred. That means that the headlamps are only lighting up an area 84 feet in front of the vehicle. If the vehicle's headlamps are about 2 feet off the ground, then when they're properly aiimed, they should be lighting up an area about 285 feet in front of the car (VOL headlamps where the left half of the horizontal beam cutoff is 2.1 inches below headlamp height at a distance of 25 feet from the front of the vehicle).
If the pedestrian was visible at 285 feet, it would have taken 5 seconds from the time the pedestrian came into view till when a collision could occur. That would have given the driver a second to react and 4 more seconds to slow down and/or change direction to avoid a collision.
Re: The Driver was Texting (Score:5, Interesting)
On the contrary, almost any human driver would have spotted that women from far away. Don't be fooled by the artificially darkened video, as others have noted in reality the lighting conditions on that road are pretty good, and she was also not jumping on the road but crossing it slowly and under perfect weather conditions. As some autonomous driving expert on HN as commented, the sensors should have had no problem picking her up from far apart and this looks a lot like a problem with the LIDAR software.
This accident was fully preventable.
Re: The Driver was Texting (Score:4, Interesting)
I do not think a human could have reacted fast enough to even think about slowing down.
They're probably going to find the pedestrian 100% responsible since they were illegally crossing an unlit section of a major highway in a totally reckless and inattentive manner, but it still looks like a failure of the Uber system -- even if they turn out not to be criminally liable, they're SUPPOSED to be better than a human at avoiding accidents, and should be held to that standard.
A Human's capabilities would be limited by the visible light, and a HUMAN cannot safely drive 38 Mph on a road at night if their visibility is not a sufficient footage down the road to safely stop in time upon an obstacle appearing at the edge of their visibility --- in a highway with no streetlights and no traffic, the driving conditions can be improved by turning on High-beams to allow a higher speed, otherwise the driver has a duty to slow down to a safe speed for the limited visibility under dark nighttime driving conditions, Therefore, the driver could be cited for hitting the pedestrian, because they were driving at an unsafely high speed that's not an allowable speed under the nighttime driving conditions without high-beam headlights enabled (By the time an obstacle appears in the illumination cone, it's already too late to react!).
Either that or the dashcam was misleading in terms of light levels, AND the driver had a longer time than 2 seconds when the pedestrian could be seen.
Re: (Score:3)
That safety driver was not even looking at the road, but the exterior camera showed the bicycle was visible to the car for about 2 seconds before the collision...
Thats the fallacy of "safety drivers" in these cars -- no human will stay alert and attentive for hours while letting the car do all of the driving, and when a bad situation does arise and they need to take over they don't have enough situational awareness to do the right thing. The same thing can happen to pilots.
Re: (Score:2)
Video appears to be digitally manipulated (Score:5, Insightful)
If you framegrab the images and then histogram the light curve it's hard edged at zero. Someone deliberately made the blacks blacker so it seems like no one could have seen her. Perhaps this is an artifact of the video compression algorithm or the camera itself.
Re:The Driver was Texting (Score:5, Interesting)
No freaking kidding. It doesn't matter that it was dark - LIDAR should have seen her from hundreds of meters away, watching her slowly step out into the road and watching her steady march across the road. Instead, it maintains speed without any braking whatsoever, straight into her. I mean, what the heck,Uber?
If you have kids anywhere near where Uber is doing automated driving tests, keep them inside. Seriously... This is just ridiculous. If it can't detect an adult slowly walking across the road, holding a bicycle, even when it's about to plow into her, what is it going to do when a little kid suddenly darts out in front?
Are they even using the LIDAR, or is it just a decoration to make passengers feel better?
Re: (Score:3)
The whole concept of 'this is good enough to do it all for you, except pay attention, just in case' is deeply flawed. If not normally engaged, a human's attention will of course drift.
Re: (Score:2)
Yeah, honestly, crossing a street with 35-40mph traffic at night in dark clothes and not at a crosswalk or intersection?
This isn't a self driving car problem, this is a Darwin award winner.
And this is *only* national news because it was a self-driving car. Had it been a normal motorist, it wouldn't have gone past the local scene.
With that said, I would expect that a self-driving car would have sensors beyond RGB. Any of IR / thermal, radar, or lidar should have been able to pick her and the bike up, even in
Re: (Score:2)
Moreover, she (the bike-walker) could have seen the car coming *way* before it (driver or computer) could see her. Why would she continue walking across when she should have seen the headlights coming many seconds away? And why *walking* a bike? Wouldn't *riding* it across have been faster? Or, you know, just riding with the traffic like I presume laws say she should have? (at least in my state, bikes on the road are supposed to follow most of the same rules as cars in terms of lanes, turning, etc)
Re: Pedestrian error = dead pedestrian. (Score:5, Insightful)
Sometimes people do things for which technology... has no way to compensate
Especially when that technology isn't fucking ready, yet.
Re: (Score:2)