Slashdot Mirror
Experts Say Video of Uber's Self-Driving Car Killing a Pedestrian Suggests Its Technology May Have Failed (4brad.com)
Ever since the Tempe police released a video of Uber's self-driving car hitting and killing a pedestrian, experts have been racing to analyze the footage and determine what exactly went wrong. (If you haven't watched the video, you can do so here. Warning: it's disturbing, though the actual impact is removed.) In a blog post, software architect and entrepreneur Brad Templeton highlights some of the big issues with the video:
1. On this empty road, the LIDAR is very capable of detecting her. If it was operating, there is no way that it did not detect her 3 to 4 seconds before the impact, if not earlier. She would have come into range just over 5 seconds before impact.
2.On the dash-cam style video, we only see her 1.5 seconds before impact. However, the human eye and quality cameras have a much better dynamic range than this video, and should have also been able to see her even before 5 seconds. From just the dash-cam video, no human could brake in time with just 1.5 seconds warning. The best humans react in just under a second, many take 1.5 to 2.5 seconds.
3. The human safety driver did not see her because she was not looking at the road. She seems to spend most of the time before the accident looking down to her right, in a style that suggests looking at a phone.
4.While a basic radar which filters out objects which are not moving towards the car would not necessarily see her, a more advanced radar also should have detected her and her bicycle (though triggered no braking) as soon as she entered the lane to the left, probably 4 seconds before impact at least. Braking could trigger 2 seconds before, in theory enough time.)
To be clear, while the car had the right-of-way and the victim was clearly unwise to cross there, especially without checking regularly in the direction of traffic, this is a situation where any properly operating robocar following "good practices," let alone "best practices," should have avoided the accident regardless of pedestrian error. That would not be true if the pedestrian were crossing the other way, moving immediately into the right lane from the right sidewalk. In that case no technique could have avoided the event. The overall consensus among experts is that one or several pieces of the driverless system may have failed, from the LIDAR system to the logic system that's supposed to identify road objects, to the communications channels that are supposed to apply the brakes, or the car's automatic braking system itself. According to Los Angeles Times, "Driverless car experts from law and academia called on Uber to release technical details of the accident so objective researchers can help figure out what went wrong and relay their findings to other driverless system makers and to the public."
1. On this empty road, the LIDAR is very capable of detecting her. If it was operating, there is no way that it did not detect her 3 to 4 seconds before the impact, if not earlier. She would have come into range just over 5 seconds before impact.
2.On the dash-cam style video, we only see her 1.5 seconds before impact. However, the human eye and quality cameras have a much better dynamic range than this video, and should have also been able to see her even before 5 seconds. From just the dash-cam video, no human could brake in time with just 1.5 seconds warning. The best humans react in just under a second, many take 1.5 to 2.5 seconds.
3. The human safety driver did not see her because she was not looking at the road. She seems to spend most of the time before the accident looking down to her right, in a style that suggests looking at a phone.
4.While a basic radar which filters out objects which are not moving towards the car would not necessarily see her, a more advanced radar also should have detected her and her bicycle (though triggered no braking) as soon as she entered the lane to the left, probably 4 seconds before impact at least. Braking could trigger 2 seconds before, in theory enough time.)
To be clear, while the car had the right-of-way and the victim was clearly unwise to cross there, especially without checking regularly in the direction of traffic, this is a situation where any properly operating robocar following "good practices," let alone "best practices," should have avoided the accident regardless of pedestrian error. That would not be true if the pedestrian were crossing the other way, moving immediately into the right lane from the right sidewalk. In that case no technique could have avoided the event. The overall consensus among experts is that one or several pieces of the driverless system may have failed, from the LIDAR system to the logic system that's supposed to identify road objects, to the communications channels that are supposed to apply the brakes, or the car's automatic braking system itself. According to Los Angeles Times, "Driverless car experts from law and academia called on Uber to release technical details of the accident so objective researchers can help figure out what went wrong and relay their findings to other driverless system makers and to the public."
Forget the Lidar, or lack of (they were testing cameras?). Forget the dude (heh, the first 12 hours thought he was a she. That's gotta hurt).
Had I been driving that car, full alert, I would have killed that chick. I'd have felt bad, even knowing it was her fault. But the fact is, this dumbass walked in front of a fast moving car, at night, when she had no illumination, and the car had headlights. Her best hope of survival was a 100% functioning self driving car, anything less and she's dead.
There were multiple failures all around which caused this death. If any one of those failures had not happened then the pedestrian would likely still be alive today.
I've summed it up here in a column which was written almost 24 hours ago so it's nice to see that others have come to similar conclusions.
I agree the LIDAR should have been able to see her before she entered the light.
However, what we still do not know is - when did she start moving?
If she was just standing in the left lane waiting to cross, the LIDAR may have seen her and just thought "well that lane is blocked, stick to this one". With the bike she might have looked like a barricade of some kind.
It could still be she started moving around the time we see her in the video, which means she essentially jumped in front of the car...
There could be a reason for her to do that - what if the car saw her, and slightly slowed out of caution? We know the car was going well under the limit when it hit, that could be a sign the car slowed down a bit prior.
The human, seeing a car slow light that might have assumed it saw her and was going to stop to let her cross. So it could easily be a case of mixed signals, with the cars cautious actions in the end being a bad thing, when driving an over-abundance of caution can often have bad consequences.
I'm still not sure the human safety driver would have seen her though, even though humans do have better dynamic range than cameras there still are times when you really can't see outside the headlights, and the woman crossing was all in dark clothing.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Well there you go. Clearly these cars should be kept on the road so Uber has every opportunity to make their technology better.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
So if it was daylight where the SNR of the LIDAR is worse with ambient light hitting the sensor and the car hit her where would you place blame? Because that is exactly what would have happened. The LIDAR system would still have missed her. The system failed period full stop. It is as bad as when Intel released the processor that did not always add right. I know the tech world is spending billions to make this work and it seems to be every techies wet dream for it to work, but it does not work yet. And deploying it in the wild is murder. It is not ready.
The rush is on to fire 10 million vehicle drivers, with all that potential money on the table who cares if it isn't safe! #uberlogic
When driving a car with normal headlights, can you genuinely not see what's the next lane over 150 feet ahead?
Lots of cars fudge the headlights a bit to the right to keep from blinding oncoming cars. Combine that with her stepping out of a very dark shadow of the tree, as well as her dark clothing (jeans and black jacket) and it's very likely that because of the dynamic range of the scene (bright headlights making darker anything street lights shone on, then her being in the shadow from even the street lights) a driver would have been too blinded by available light to see the pedestrian.
She didn't even have reflectors on the wheel of the bike, much less herself - even one might have saved her.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
The car had a single occupant who was not focused on the road. In a test vehicle, regardless of how autonomous the vehicle is if one person is required to monitor computer systems whilst the car is travelling, they need a second occupant to monitor the road.
The Lidar system should work with no light at all, its an infrared laser system, it emits its own light! In fact if anything it would probably work better in complete darkness!
This was obviously a major technology failure, however it needn't have resulted in someones death.
The greatest failure here is not in the technology but in Ubers testing procedures.
The lesson for the rest of us is that Uber's self driving technology is not ready for prime time, for whatever reason(s).
I dunno ... I mean even if we assume that a human driver could have avoided this particular accident, that doesn't mean the technology isn't still an improvement over human drivers in other cases. You'd need a lot more data to reach that conclusion. It could very well be that lives saved in other, more common types of preventable accidents massively outweigh the lives lost in these types of abnormal occurrences.
The timeline is thus: She looks up as she is going under the bridge, looks down again, then a few seconds later she looks up basically as the car is hitting the woman.
Yeah, the driver was totally negligent in any case, looking at her phone more than at the street. It's true a human can't pay perfect attention all the time, but any human can put their phone down.
"First they came for the slanderers and i said nothing."
I've got a forward facing car camera, and its very clear that eyes see things far bigger and brighter at the horizon than cameras. You would see her, even in that light.
This is why the moon seems bigger and brighter when its near the horizon. When actually its dimmer due to atmosphere and the same size as normal.
*But*, more than that.
I've also seen the other video of this stretch of road filmed with a normal smartphone camera and its clear the camera in the car had terrible dynamic range. There is absolutely no way a person looking at the road would not have seen her.
https://www.youtube.com/watch?v=CRW0q8i3u6E
Level 1 self drive is where the car acts as a safety feature for the driver.
Level 2 is where the driver acts as a safety feature for the car!
Which is it!! Is the car safer or the driver?? You cannot have Level 2 because its simply dumping liability on the driver, when he's not party to the decisions the car is making and does not know what it will do, till after its done it.
The car drives, the man observes, then deduces what decision its made from the changes to the car. He cannot be the safety backup for the car. It's not possible, its a legalese get out for self driving car makers.
Pull up google maps.
https://www.google.com/maps/@3...
Observe the sign that says not to cross there (one on the other side of the median too btw).
Observe the road widens from 2 to 4 lanes right before where the accident occurred.
Observe that trees and bushes on the edge of the median to the left would have blocked Lidar until the car was within 50' of where the accident occurred.
Confirm the scene against the video with the sign that reads
"
Begin right turn lane
-
Yield to Bikes.
"
Understand that it refers to bikes in the bike lane on the right.
Pull up a top down view and get a ruler and note that the street lights are about 120' apart here. They are much closer in my area ( 50-100' apart).
(Also, if you look up the light pollution map, the area to the right of the road is very dark).
Go 600' back down Mills road and observe the 45mph speed limit on this road.
It's a terrible accident. The road widening like that with bushes and small trees blocking lidar, the pedestrian walking across the road without even looking for oncoming traffic (because it was late at night on a sunday in what looks like a kinda empty area so there probably isn't much traffic).
Wait for people to really dig into this. Don't make a snap judgement.
I'm sorry someone died, but there has been a lot of bad information and a lot of people rushing to judgement before they have the facts.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
The street lights just before where the accident occurred may have blinded the camera too.
Also the bushes and small trees to the left on the median would have blocked lidar and visual L.O.S. until she entered the road *and* the car was within 50'.
The road literally widened from 2 to 4 lanes right where the accident occurred.
https://www.google.com/maps/@3...
I don't see how an approaching car or human could have seen her before she was on the road.
I regularly (2-3 times a year) drive up on people who I simply do not see at night until I'm on them. If they were crossing the road instead of walking right beside it, I dont' think I could avoid hitting them.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
When the car is deemed ready and put into production ans so on, yes absolutely.
But this was a *test driving* situation, during which unexpected and possibly dangerous situations may occur. It is irresponsible of Uber to not have two people in the car at all times, one to monitor the road and take evasive action if necessary, and one to monitor the systems. One person cannot do both at the same time.
Eat the rich.
https://www.youtube.com/watch?...
Take a look at 0:30 and beyond, the robocar obviously failed and the driver had lots of time to react if they were paying attention.