Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona • Hacker News Board

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#221

Post by orangecat » Tue, Mar 20, 2018, 3:30 AM UTC

username223 wrote: post #133

danso wrote: post #94

FWIW, this seems to be where the accident took place, based on news photos: https://www.google.com/maps/place/N+Mill+Ave+%26+E+Curry+Rd,... It's not completely unlighted, though I'm not sure that's supposed to matter given the AV's sensors? Even if Uber nor the human driver is not legally at fault, I'd still be interested in knowing of Uber's LIDAR and other sensors even detected the woman, and whether it's part of t…

I think I'll wait for the dust to settle a bit here, since it's hard to tell what happened. This is the wreck photo from the local news, showing a damaged bike on the sidewalk, and an Uber robo-car with front passenger-side damage, sitting in the right lane: https://sharing.wcpo.com/shareknxv/photo/2018/03/19/poster_9... It's hard to square that with someone darting over from a brushy median on the driver's side of t…

This is the wreck photo from the local news, showing a damaged bike on the sidewalk, and an Uber robo-car with front passenger-side damage, sitting in the right lane

This looks like the corresponding street view: https://www.google.com/maps/place/N+Mill+Ave+%26+E+Curry+Rd,...

It's hard to square that with someone darting over from a brushy median on the driver's side of the car.

One possibility is that the pedestrian was trying to cross to the weird sidewalk/median thing and came out from behind the tree here: https://www.google.com/maps/place/N+Mill+Ave+%26+E+Curry+Rd,...

"Safety driver" of the robo-car may have a record for attempted armed robbery. Uber disputes the driver's identity.

Good grief. I would normally be surprised if a company denied a claim like that without having solid evidence, but Uber would be much less surprising.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#222

Post by bichiliad » Tue, Mar 20, 2018, 3:34 AM UTC

protomyth wrote: post #178

bichiliad wrote: post #117

Earlier quoted context omitted.

This is probably less the case with self-driving cars, as I imagine they're using pretty new or well-maintained vehicles, but there's usually a 5mph leeway with speed limits due to slightly mis-calibrated speedometers. I anecdotally remember hearing this being the reason for speed limits that end in "5" rather than being multiples of 10. https://travel.stackexchange.com/questions/40113/do-any-stat...

Wouldn’t that be mitigated by the GPS and the map? It would seem that an automated car using maps as an important input would no its exact position and therefore speed.

That could easily be the case, sure. Some other people mentioned that it could be just to more closely emulate a human driver. I don't think we'll have a good answer without knowing the implementation.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#223

Post by CydeWeys » Tue, Mar 20, 2018, 3:35 AM UTC

seanmcdirmid wrote: post #197

akira2501 wrote: post #190

Earlier quoted context omitted.

> They are the most dangeorus thing that most people encounter on a daily basis. It's actually yourself, and it's worse when you're at work. Deaths due to vehicles: 36k per year. Deaths due to suicides: 44k per year. Deaths due to accidents: 190k per year. > There are many age and other cohorts of the population for which the leading cause of death is vehicles. This is not true for any cohort. [1] And yes, I'm aware…

If we are considering deaths-per-journey, then cars look even worse when compared to planes: most people take many more journeys by car in comparison to planes. Let's say we take 500 journeys by car for every one journey we take by plane (not that outlandish), then cars are 1500 times worse than planes. Edit: it seems like it is the other way around. Planes are 3 times as worse than cars per journey, so flip that num…

Deaths per journey is also useless because a vehicle owner might make many trivial car trips per day, but when you're flying somewhere in a plane it's usually meaningful. For the same death rate, I'd much rather go on an awesome vacation across the world than on three errands.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#224

Post by avs733 » Tue, Mar 20, 2018, 3:35 AM UTC

prawn wrote: post #220

Pushing a bicycle laden with plastic shopping bags, a woman abruptly walked from a center median into a lane of traffic and was struck by a self-driving Uber operating in autonomous mode. From viewing the videos, “it’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway,” Moir said. The "driver"…

Apparently my stored links on this topic are dead but two worhtwhile pieces of information on that point

1) drivers who consistently drive below the average speed of other drivers on a road (i.e., who religiously drive the speed limit) are more likely to be involved in crashes

2) More accidents are caused by speed differential than by speed by itself.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#225

Post by r_klancer » Tue, Mar 20, 2018, 3:36 AM UTC

My takeaway from the comments in this thread: self-driving Ubers don't drive defensively, almost surely can't "read" pedestrian behavior, and even with lidar may never have seen Elaine Herzberg. Oh, and they speed.

This sounds like straight out of the "hell" scenario for driverless cars: "we cannot accept responsibility for what happens to pedestrians and cyclists foolish enough to stray out of their designated safe zone".

I'm referencing Robin Chase (the Zipcar founder)'s "heaven or hell" framing - https://www.citylab.com/transportation/2014/04/will-world-dr...

For a recent, thoughtful talk about the damage self-driving cars might do to the city if we are not careful, see this Congress for the New Urbanism talk about "Autonomous Vehicles & The Good City" at https://www.youtube.com/watch?v=utnPEbDNbrE

It's also worth mentioning that there's a bill (AV START, or S. 1885) in the current Senate that would actually preempt local and state regulation of autonomous car safety, but it has been put on hold.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#226

Post by brokenmachine » Tue, Mar 20, 2018, 3:37 AM UTC

danso wrote: post #134

showerst wrote: post #20

Earlier quoted context omitted.

A X,000lb car going 38mph takes a certain amount of time and distance to stop, even with a perfect reaction time. If she really did step out in front of the car with less than that distance to spare, than there wasn’t much to be done, sadly.

But isn't this the reason why most drivers, when seeing someone walking along a median (with a bike full of junk, no less) or even biking in a dedicated lane, will move into another lane to give some clearance? It doesn't sound like there were any other cars near the Uber AV, and it was relatively far from the intersection (in case it needed to make a left turn). So I'm wondering why it didn't make the decision to ch…

It didn't even attempt to brake, so likely it never had any idea the person was there.

Also I know I wouldn't be programming my self-driving car to jump to a different lane every time it saw a pedestrian in the distance.

But then again I wouldn't be programming my self-driving car to speed either...

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#227

Post by brokenmachine » Tue, Mar 20, 2018, 3:38 AM UTC

mattigames wrote: post #93

jakeogh wrote: post #67

Earlier quoted context omitted.

Personally, I am constantly scanning for people near the edge of the road and anticipating their movements. I bet most of us have had the experience where someone stepped out unexpectedly but you already took action because you could tell they were not paying attention. Good luck programming that.

Yeah there are a lot of subtleties that are hard to program; that's why it must be extremely effective on things that _can_ be programmed such as quick breaking, quick steering; prefer lanes that are farther away from pedestrians near-by, beep when entering a non-illuminated zone and hundreds of other details.

Quick braking is the one thing you would have thought they could get right for sure, and they didn't manage even that on this occasion.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#228

Post by digitalantfarm » Tue, Mar 20, 2018, 3:39 AM UTC

I read the person was homeless (and had a shopping cart). Wouldn't be surprised if she tried to get hit, seeing it was an Uber driverless car. I want to see the dash-cam footage.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#229

Post by noobermin » Tue, Mar 20, 2018, 3:41 AM UTC

taneq wrote: post #211

noobermin wrote: post #173

Earlier quoted context omitted.

Agree with part of it but no, pedestrians do not have an equal obligation. With power should come responsibility. In America, we seem to do the opposite, disassociate responsibility with those in power and blame victims.

With power should come responsibility, sure, but "power" means "ability to affect the situation", not "joules per second." In the scenario of a pedestrian suddenly running out in front of a car which is driving at the speed limit, the pedestrian has far more power to affect the outcome (by choosing not to run into the path of the car) than the driver (who could not reasonably predict the pedestrian would behave in th…

The power I'm talking about here isn't ability, it's literally the might and momentum of the car. And because of that power, drivers need to take more responsibility when they drive. A driver doesn't become responsible the milliseconds before they kill someone, they become responsible the moment they set foot in a car.

The ability that drivers do have is to drive defensively and keep a look out for pedestrians. That is what I'm talking about, not by frustrating Newton's 3rd law. And because of that extra ability to do damage, they must be subject to harsher scrutiny than those who can more easily suffer damage.

Re: Police Say Uber Is Likely Not at Fault for Self-Driving Car Fatality in Arizona

#230

Post by treis » Tue, Mar 20, 2018, 3:41 AM UTC

portofcall wrote: post #27

showerst wrote: post #20

Earlier quoted context omitted.

A X,000lb car going 38mph takes a certain amount of time and distance to stop, even with a perfect reaction time. If she really did step out in front of the car with less than that distance to spare, than there wasn’t much to be done, sadly.

https://www.theverge.com/2018/3/19/17140936/uber-self-drivin... The brakes didn’t even start to engage. Just how fast do you think this woman moved exactly?

There's miles of difference between the brakes not engaging and not showing significant signs of slowing down.

It's enormously unlikely that the Uber car failed this badly. The most plausible scenario is that the car did brake but just didn't leave skid marks. The car not recognizing a human in its path would be a shocking failure.