35 MPH is roughly 51 feet per second. 40 MPH is 58 feet per second. I think the interesting question here how far ahead you would normally expect a human to be able to see, via headlights. Interestingly, the answer appears to vary greatly with the car. Many cars' headlights work poorly, study finds, with big gap between best and worst
According to that article, even the crappiest headlights (on a new car) would illuminate a pedestrian 128 feet out, giving you 2-3 seconds to react at 35 MPH, which is a decent amount of time. The fact that the car didn’t seem to react at all implies that it didn’t pickup the pedestrian, whatsoever, when a human likely would have. Note that this is separate from a discussion of whether successful avoidance would have been possible by your normal human driver—I’m only talking about detecting and responding in some way.
While I think it might be fair to say that a reasonable human driver had only some chance (maybe a low chance) of avoiding the pedestrian, I think a reasonable human driver would have reacted prior to impact, in some way, given those speeds if the info in the article I linked is correct.
Is this accurate to time? Seems like you’d have about 1-1.75 seconds from stimulus to collision. Didn’t help that the observer appears preoccupied for the most part, but maybe that’s the point.
I suspect whether the video is real-time isn’t the only important question. The other question is whether the dash cam has the same low-light sensitivity as a human eye or the car’s own optical sensor suite. A lot of/most cameras aren’t as good in low light as a typical human eye. You can’t assume that what you see in the dashcam video is what you would have seen with your own eyes had you been driving.
As I posted above, charitably assuming crappy headlights would still give you at least 2+ seconds to see the pedestrian with typical human vision. Given that this was a Volvo SUV (I think), I suspect the headlights would have given the typical human 3-5 seconds of reaction time at 35 MPH. Maybe not enough time to completely avoid the pedestrian, but certainly some time to make some sort of reaction.
Street lights change the distance you can see to though. If you’re inside a cone of light from a street light it’s harder to see out of it than if it’s just dark.
Also, she stepped into the headlights’ cone of light from the side, and she was already pretty close to the car. It’s not like she was standing in the road stationary at a distance.
I’m surprised radar, lidar or whatever didn’t pick her up, but I don’t know what capabilities are built-in for that. I do think even an attentive human driver would quite likely have hit her, too. She was really careless.
She’s slowly walking her bike. I don’t think she was going much more than 2-3 ft/sec., that’s close to standing still in the timeframe that a driver would have to react. She’s wasn’t the equivalent of a fast-moving deer darting from the side of a road.
I don’t understand what you mean by the “she was already pretty close to the car” part. If you’re judging by how close she was when she appeared to become visible in that video, again, you can’t assume that the video is an accurate reflection of what a normal human would see under those conditions, due to the likely limitations of the camera capturing that video.
Most of us have driven at 35 MPH at night on dark-ish roads. While I wouldn’t claim that any driver is legally at fault for hitting a jaywalker under those conditions, I think most of us also know that you tend to have a handful of seconds to react (again, react in some way, not necessarily total avoidance) to things (e.g., cats, dogs, deer) that are standing in the middle of the road at those speeds. 35 MPH isn’t THAT fast.
Do people actually go through crosswalks when cars are coming and just assume the car will stop?
I am genuinely not saying the woman “deserved it” or anything like that. No one deserves to be hit or die, and it is a tragedy.
I’m just saying more from my perspective, if I can see headlights from a car that is that close, there is no fucking way I’m going across the street until the car is either: (i) past me; or (ii) so obviously slowing that I know it sees me and is going to stop.
I know you theoretically have the right of way as a pedestrian at certain points (crosswalks, with the green light and a walk signal), but there’s no fucking way I’m trusting cars to observe that, because the car wins the argument at the end of the day even if it’s in the wrong.
I think what this is going to clarify is whether autonomous vehicle companies are going to be able to use meatbags to avoid liability. That is, will a human “copilot” always be at fault for failing to avoid a collision.
With all due respect the woman, by her appearance and behavior, looks literally like the sort of person specifically chosen to be liability insurance, essentially a minimum wage warm body. Just throw a poor person in the driver’s seat and blame them for anything that goes wrong.
I’m sure Uber collaborated to some extent with the authorities in releasing that video because they were pleased to see the woman demonstrate very little interest in precautionary driving - she’s clearly just staring at her phone most of the time. That way the narrative that “look, this irresponsible woman staring at her phone and not even looking at the road is at fault!” can permeate social media and deflect some of the blame.
With all due respect, drawing any conclusions based on how they looked or their behavior in the limited context of that video is not giving them all due respect.
But i live the same world and we both know what we see. It isn’t polite, but it’s more like to be true than not. Uber isn’t hiring post-graduate Google-level employees.
My point was just that Uber is hiring poor people just to throw them under the bus when exactly these situations occur. And… she’s just more likely to be that by observation. I feel anything it will maker her even less sympathetic, the crass and lazy will be happy to crucify her even more because of it.
You really think in any lawsuit a decent lawyer wouldn’t go after the vehicle (and the big bag of money there) rather than the monkey sitting in the passenger seat who probably has a nominal insurance policy.
There is no way Uber avoids legal blame for this just because they have someone riding shotgun.
I am pretty sure even someone who is highly educated would be bored stiff sitting in a car doing nothing for hours. The only reason they’re not in that seat is Uber probably isn’t paying enough and they might have figured out that they would be sticking their neck out for a company and car that doesn’t car about them.
Since she’s not driving… but in the driver seat is it legal for her to review a phone like that? I’m not really judging her; heck I’d probably just fall asleep.
I think it really depends how they framed the job description and what they expect employees to do. If just “sit and watch” is all they’re normally expected to do it would be hard to do this for hours on end. But then if they’re hiring them as a failsafe they should make clear they are driving and be expected to act appropriately (hands on wheel the whole time).
However if they are driving the whole time they may be subject to regulations that would cost Uber more money. So i would expect to discover Uber wrote their contracts in such a way to avoid those.
Doing something boring and repetitive means you have people who become practically unable to respond the way you want. Just look at what happens with TSA agents. There’s a psychological reason why knives and fake guns pass right through those scanners. People are bad at exactly what’s required of these human drivers.
Which is why I agree with Adam_B on the other thread. Having human drivers who are probably untrained handling this task (especially without the sort of attention-maintaining technologies in place that were described elsewhere) is the dumbest part of this whole thing.
What is the practical alternative for the transition period when these vehicles need to be tested on real streets? Is the backup driver essentially security fiction? Yes, to a large degree. Though it is better than having zero backup for at least some circumstances (e.g., malfunction of sensors in a non-immediate emergency). But what else do you do, instead of having the drivers?
There is no practical alternative as far as I’m aware.
This is why I call it dumb.
But sure, I’ll immediately cede that the dumb thing is pretty much the only thing you can do. Although I would like it if they were trained for the task they really need to be good at, which is maintaining full awareness of their task even though it’s repetitive and boring, which needs real training and very competent people.
Like repetitive factory work. There’s a reason that all that heavy machinery has safety factors. After hours of doing the same thing over and over, the mind wanders. Then hands end up in a stamping press.
