Live data from Hacker News

Tempe Police Release Video of Uber Accident

twitter.com

931–940 of 1001 posts

Re: Tempe Police Release Video of Uber Accident

#931

Lots of mention about LIDAR here. As well as IR sensors, etc. Unless Uber comes out with all the technical specifications that is mounted on the car to detect motion and obstruction on the road, all are mere speculations. The vehicle in question here would be in the custody of the authorities (impounded), so they should be able to investigate the details on it. Uber should provide as much technical specifications ava…

It doesn't really matter what the car was actually capable of, so much as what it should have been capable of.

Let's say for argument's sake that the car wasn't equipped with whatever it needed in order to see a person crossing the street in the dark. That doesn't excuse Uber: it just means they shouldn't be testing in the dark on public roads.

Re: Tempe Police Release Video of Uber Accident

#932
post #660

A lot of people seem to mention how LiDAR should have seen her coming: Generally speaking, that's true. But it's difficult. In a single frame the system would hopefully detect something on the left of the lane, but wouldn't know what it is (if they use AI it's probably not trained for "person shoving bike packed with plastic bags"). If you have the object detection in place, you still need a trajectory for that objec…

I don't know how you can watch this video and then say that they should "maybe do more testing in controlled environments". This was a complete failure and Uber's cars should be immediately banned from testing on public roads and there should be investigation into their practices to ensure public safety while testing their vehicles. If they rushed it trying to play catch up with Waymo they should be punished. Persona…

It's a "maybe" because I don't have any more data points than what's present in the video - only assumptions.

But I strongly assume this could have been avoided by more thorough testing (e.g. use a contraption to shove various inanimate objects in the path of the car at various speeds [both object and car] in various light conditions). And iff a proper investigation finds out people responsible for this accident (engineers doing dirty fixes, managers neglecting safety, the driver just being an unaware dummy; you name it), they should carry the consequences same as a human driver would.

And yes, their self driving cars should stay of the road until the fault has been found and fixed. I never claimed the opposite. (Aren't they? IMHO it's the obvious thing to do; but then we're talking about Uber and their poor sense of responsibility...)

Eventually "maybe" has too weak of a connotation in this context (as a non-native speaker I sometimes find it difficult to pick the correct variant for the exact thing I want to implicate) - for which I apologise if I mislead you there.

Re: Tempe Police Release Video of Uber Accident

#933
The video is very damning. Either a problem of camera contrast or a big bug. And having a backup human driver is useless.

There is a terrible ethical dilemma here. Humans are currently at a a death rate of about 1 per 10^8 (hundred million) miles travelled. Waymo, which is best in class, is at 1 human engagement every 5,000 miles as per California data. Of course, most of those were likely minor. But there are more than 4 orders of magnitude there to allow for plenty of surplus dead people. These things shouldn't be on the streets!

Re: Tempe Police Release Video of Uber Accident

#934
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…

I don't work on self driving cars or even vision, but I have heard when you are on a highway you start with filtering out small stationary reflections, since they are almost certainly not cars. (Which is maybe why the tesla didn't see the big red firetruck - it was not visible). It's not a big leap to imagine that Uber's LIDAR ignored the bike because it was not recognized as a person, was moving perpendicularly at a low speed, so got pre-filtered as a puddle. We can only guess and Uber will report whatever they want.

Re: Tempe Police Release Video of Uber Accident

#935

Earlier quoted context omitted.

And your point is what exactly?

Accident aside, if Uber is so "strict" on their background checks for drivers, how do they let a convicted robber be apart of their high profile pilot?

I'd say if you haven't done something criminal in 17 years you'd probably learned your lesson and should be allowed to fully function in society.

Re: Tempe Police Release Video of Uber Accident

#936
post #869
post #818

Someone on Reddit posted a comparison of the same spot at night but with a good camera: https://m.imgur.com/a/PM7uu from a video at https://www.youtube.com/watch?v=1XOVxSCG8u0 That paints the whole accident in a very different light from the video from Twitter. Would any human not notice someone crossing in the good camera screenshot?

I wonder if the moon phase or clouds had anything to do with the lighting conditions.

Well, obviously it was dark. It was new moon on March 17, one day before accident. However, dark is normal conditions for driving.

Re: Tempe Police Release Video of Uber Accident

#937

A straight road, in dry conditions, obstructed by an entire person and a bicycle with REFLECTORS on it, and the car doesn't see her? I think whoever wrote this code should be in prison, quite honestly, along with their entire management chain. I suspect many people will feel the same way. This is a really really bad look for automated vehicles in general. It also backs up my contention that anything less than full 10…

I don't see any reflectors in the video. The wheels and something in the vicinity of the handlebars are the brightest objects I see.

Not the wheels I think, but the shoes of the woman (plus something near handlebars). That bike would not be street legal here (north Europe) due to not enough reflectors - although a self-driving car still should be able to spot it.

Re: Tempe Police Release Video of Uber Accident

#938

Why are safety drivers not: - working in pairs, so there is social pressure, conversation, and two pairs of eyes to increase alertness and safety? - doing shifts of 30 - 45 minutes at most [1] (although they could potentially swap back and forth with a co-driver) - issued a dumb-phone for emergencies and searched for entertainment devices (it's good enough for Amazon warehouse staff) - being monitored by the driver-f…

All of those suggestions will increases cost and decrease profit, there's no law mandating any of those so why would they go above and beyond for 'safety' reasons? I understand it's the ethical decision but at the end of the day for-profit companies are only interested in one thing: maximizing profits.

Profit maximisation is not the only factor of running a business, even a public company.

But even given that, what are the PR and legal costs of this predictable [1] accident being made worse by a poor safety system? Plus the downtime on doing training/testing? And the probably harsher requirements that might now be enforced?

Maximizing profits doesn't mean doing everything on the cheap.

[1] in the sense there was likely to be an accident at some point.

Re: Tempe Police Release Video of Uber Accident

#939
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…

Indeed! To LIDAR, she was basically standing in the lane, with a bulky bicycle. To visible light, including the driver, who was apparently half asleep or watching the dashboard, she was in shadow until just before the collision. So yes, LIDAR should have caught this. Easily. So something was clearly misconfigured. And even if the driver had been carefully watching the road, he probably wouldn't have seen her in time.…

Probably the LIDAR did catch it. Probably the algorithm (neural network) that takes in LIDAR data and outputs whether there is a object in front failed or gave a really low probability which was less than the threshold specified. This happens all the time with deep neural networks. In any case technology is to blame and self driving should be banned until these issues are resolved.

Re: Tempe Police Release Video of Uber Accident

#940

Earlier quoted context omitted.

How did LIDAR and IR (?) not catch that? That seems like a pretty serious problem. Something is badly wrong there. That should have been detected by LIDAR, radar, and vision. Yes, they need a wide dynamic range camera for night driving, but such things exist.[1][2] They're available as low-end dashcams; it's not expensive military night vision technology. Radar should pick up a bicycle at that range. The old Eaton VO…

I am glad to hear the NTSB is investigating this and not just the local police (who would lack the technical resources to make a useful judgement). Have there been any statements at all from them so far?

The NTSB tends to not make even initial statements until upwards of a several weeks in, and final statements as much as 6 months to a year later. They're nothing if not thorough.
Post reply on HN