Tempe Police Release Video of Uber Accident • Hacker News Board

Re: Tempe Police Release Video of Uber Accident

#171

Post by vkou » Thu, Mar 22, 2018, 12:19 AM UTC

chki wrote: post #151

osrec wrote: post #107

Earlier quoted context omitted.

I think it could have stopped safely. Assuming it has quicker reactions than a human, stopping distance should be around 18 meters. I think it just completely missed her. Edit: critical typo

that is a bold statement to make from the video information we got, at least in my opinion. I would estimate the speed to be at least 50 km/h which gives 25 m as a very rough estimate for the stopping distance without reaction time. It's probably better than that but not 8 m. But I do agree that the car doesn't seem to have stopped at all and missed the victim completely.

In 8m, though, it may have been able to brake enough to make the collision survivable.

Re: Tempe Police Release Video of Uber Accident

#172

Post by jmharvey » Thu, Mar 22, 2018, 12:19 AM UTC

As presented, this looks like a pretty classic "overdriving your headlights" situation.

Even though they're made of retroreflective material, only two lane divider dashes at a time can be seen in the video, indicating something like 50 feet of visibility. Stopping from 38 mph takes 70+ feet from the time the brakes are applied (and human reaction time adds quite a bit more). Things (people, animals, stopped cars, road hazards) appear in the road fairly routinely. If you can't stop inside the area you can see, you're operating your vehicle recklessly.

Re: Tempe Police Release Video of Uber Accident

#173

Post by mikeash » Thu, Mar 22, 2018, 12:19 AM UTC

Treblemaker wrote: post #32

The first question that comes to my mind watching this, is, why didn't the _pedestrian_ see the _car_? It looks as though she didn't even look up until the car was only a few feet away from here. I'm not just trying to "blame the victim". The car "should" have responded sooner. But see-and-avoid works both ways. If you're choosing to cross a road in the dark, you, too, have a responsibility to understand the drivers'…

I see a lot of people who cross the street without looking. I definitely don’t get it. I constantly check everywhere even when I clearly have the right of way. But that’s how people are.

Re: Tempe Police Release Video of Uber Accident

#174

Post by nickez » Thu, Mar 22, 2018, 12:19 AM UTC

osrec wrote: post #107

TrainedMonkey wrote: post #53

There was almost a second where the woman was clearly visible and yet the car did not attempt to emergency break. I would not expect the vehicle to safely stop, but it could have definitely slowed down and reduce the energy of the collision. This is not criticism against self-driving tech, I would not expect alert human driver to avoid collision either due to limited reaction time. With technology however, we should…

I think it could have stopped safely. Assuming it has quicker reactions than a human, stopping distance should be around 18 meters. I think it just completely missed her. Edit: critical typo

8m ?! That sounds completely unreasonable to me. The first google hit says it is more like 33m..

http://www.brakingdistances.com/38Mph

Re: Tempe Police Release Video of Uber Accident

#175

Post by jjeaff » Thu, Mar 22, 2018, 12:19 AM UTC

jackpirate wrote: post #78

atonse wrote: post #25

How did LIDAR and IR not catch that? That seems like a pretty serious problem. It's clear from the video that a human driver actually would've had more trouble since the pedestrian showed up in the field of view right before the collision, yet that's in the visible spectrum. When I argue for automated driving (as a casual observer), I tell people about exactly this sort of stuff (a computer can look in 20 places at t…

It's clear from the video that a human driver actually would've had more trouble since the pedestrian showed up in the field of view right before the collision, yet that's in the visible spectrum. That's not at all clear to me. I don't know too much about cameras, but it looks to me like the camera is making the scene appear much darker than it actually is. In the video, you can see many street lights projecting down…

Since in the video, we aren't seeing the original scene, but rather, the camera's interpretation of that scene, I think it would be hard to judge except to base it on what your average streetlight brightness is.

Like a camera, your eye also has only so much dynamic range. So if those street lights are bright enough, or your interior lights are too bright, you might have nearly zero visibility in those shadows.

But it is certain that a self driving car "should" be able to see. Even two cheap digital cameras one tuned to see the darker range and the other brighter should easily see in these type situations.

Re: Tempe Police Release Video of Uber Accident

#176

Post by alva » Thu, Mar 22, 2018, 12:20 AM UTC

F2 wrote: post #131

alva wrote: post #29

On the one hand, the cyclist became visible in the footage at a very late stage. If that is how the drivers eye saw it they would have very little time to react, however they may have been able to do a sharp swerve. On the other, the operator was clearly distracted regardless of if they would have been able to avoid this.

"however they may have been able to do a sharp swerve." As someone that enjoys racing cars and other types of high performance driving I feel like I can pretty confidently say this isn't even remotely a realistic statement. You're grossly overestimating human response times. Even if we're ignoring response time and imagining the driver instantly turning the steering wheel, large SUVs like the XC90 simply can't genera…

I accept this is almost certainly the case.

Re: Tempe Police Release Video of Uber Accident

#177

Post by cure » Thu, Mar 22, 2018, 12:20 AM UTC

Why was the car speeding? Doing 38 mph in a 35 mph zone. Are these things programmed to systematically speed?

Re: Tempe Police Release Video of Uber Accident

#178

Post by eyeing_see » Thu, Mar 22, 2018, 12:20 AM UTC

ashelmire wrote: post #49

She crossed at a spot that was pitch black. Sneakers come into view maybe 40 feet away. Clearly visible at 20. Is that enough time to stop at that speed? Not sure I would have stopped in time.

It seems pitch black on a video, but human vision is much better in low light. You may have had way more time to react.

Re: Tempe Police Release Video of Uber Accident

#179

Post by bigiain » Thu, Mar 22, 2018, 12:20 AM UTC

jackpirate wrote: post #78

atonse wrote: post #25

How did LIDAR and IR not catch that? That seems like a pretty serious problem. It's clear from the video that a human driver actually would've had more trouble since the pedestrian showed up in the field of view right before the collision, yet that's in the visible spectrum. When I argue for automated driving (as a casual observer), I tell people about exactly this sort of stuff (a computer can look in 20 places at t…

It's clear from the video that a human driver actually would've had more trouble since the pedestrian showed up in the field of view right before the collision, yet that's in the visible spectrum. That's not at all clear to me. I don't know too much about cameras, but it looks to me like the camera is making the scene appear much darker than it actually is. In the video, you can see many street lights projecting down…

I also suspect that human eyeballs would have a different view of the light/dark portions of what's depicted there, and especially eyeballs would have probably had a much higher chance of detecting movement in peripheral vision than that video gives any hint of.

We typically cant see much detail in the scene out of our small region of focus, but you can bet if a tiger appears from behind a tree our visual system will scream to the brain _look over there right now!_

Our eyes and our entire visual processing system is very much not "just like a webcam, but made out of meat".

Re: Tempe Police Release Video of Uber Accident

#180

Post by pdonis » Thu, Mar 22, 2018, 12:20 AM UTC

eanzenberg wrote: post #164

Although it seems from the dashcam footage that the scene is dark and the bike/pedestrian comes out of nowhere, humans are able to handle extreme low-light situations and I don't doubt that this pedestrian would have been visible to an attentive driver. If autonomous cars cannot handle low-light conditions given the current state of hardware, either from visible, ir, or lidar, then they need to run in the daytime onl…

> humans are able to handle extreme low-light situations

But this wasn't an extreme low light situation. It was a situation with street lights alternating with extreme low light regions. Under those conditions the human eye will not be dark adapted and your low light vision won't be very good.