Uber vehicle reportedly saw but ignored woman it struck • Hacker News Board

Re: Uber vehicle reportedly saw but ignored woman it struck

#81

Post by Sangermaine » Mon, May 07, 2018, 7:30 PM UTC

dahart wrote: post #75

Rotdhizon wrote: post #39

Earlier quoted context omitted.

Not related to the article but I want to chime in on your point. Having spent some time in VA, aggressive driving and dangerous scenarios were infrequent from my experience. Then having spent some time in FL, aggressive driving was the norm and driving laws did not exist for most people. So I wonder, do self-driving cars needs something like driving profiles for certain regions, where defensive/aggressive response be…

Personally, I’d rather see driving behavior normalize nationally than cater to regional differences in enforcement. The driving laws in each city & state the US are less different than the locally accepted behaviors, but it would be safer all around if drivers and pedestrians all expected the same things, and if we took traffic & safety laws more seriously. More enforcement of existing traffic laws will save many ord…

The "reality should conform to my ideal, not my ideal to the reality" approach is and has always been a guaranteed recipe for failure. "Let's change everyone's minds and behavior everywhere" is much harder than changing software.

Re: Uber vehicle reportedly saw but ignored woman it struck

#82

Post by digikata » Mon, May 07, 2018, 7:31 PM UTC

tlb wrote: post #4

The disturbing thing about this crash is that it's such a basic thing to test on a test track. A pedestrian walking in front at night, with nothing else unusual happening, should be on the list of scenarios to get 5 9s reliability on before taking it out in public. The things you have to test in public are the complex real-world interactions, like these: https://medium.com/kylevogt/why-testing-self-driving-cars-in...

I think the disturbing thing to me is that there wasn't any sort of layered reliability approach to the testing process itself. When you know you're going to be testing changes to a very complex, multi-sensor real-world system, it would be a standard approach to lay in a more foundational set of technology and practices to the system tests to serve as a conservative safety net.

For example, an overriding set of heuristics to brake if needed could be one precaution, and in terms of practices, the human monitor, should have been swapped out at high intervals to alleviate attention fatigue. One can go back and forth on the details, but the need for a separate and independent treatment of safety reviews from the primary product development activity in this kind of endeavour should have been drawn from any of a number of existing practices in other industries.

And this is something that a lot of the casual tech articles completely miss - criticizing the primary autonomy tech which of course will have faults, but if you know you will have faults, then where is the serious plan to operate safely knowing the dev system is anticipated to have faults of many types.

Re: Uber vehicle reportedly saw but ignored woman it struck

#83

Post by stefan_ » Mon, May 07, 2018, 7:32 PM UTC

DaiPlusPlus wrote: post #54

assblaster wrote: post #7

The question I want to know is: Did Uber doctor the camera video to make the road seem darker than it actually was? If you recall, there was initially a doubt that Uber was at fault because the streets appeared extremely dark, but I think they were buying time to delay the outage at incompetent Uber. The police initially didn't fault Uber based on this video.

The cameras used in applications like these are not like your typical 24-bit RGB cameras in phones or webcams that need gain adjustment: their sensors support a very high range of brightness values (like "high dynamic range") - but the video released to the public and the media would be MPEG/H.26x/etc which do not support HDR - or at least most video players wouldn't. The police probably lacked the expertise to know…

The video was very obviously obtained from a dashcam solution that isn't actually used to do any of the autonomous driving. It's possible they don't even record the raw footage from the AI cameras, that's a huge bandwidth hog.

Re: Uber vehicle reportedly saw but ignored woman it struck

#84

Post by Udik » Mon, May 07, 2018, 7:33 PM UTC

SilasX wrote: post #47

Sorry for the snark, but there's a joke in here about: "Uber business development saw but ignored laws it broke."

There is without doubt a grotesque irony in this story. Uber stands as a perfect symbol of the reckless capitalism of the 21st century: invests hundreds of millions in building a self-driving monster in the shape of a 2 ton luxury SUV, then puts behind its wheel- as a "safety driver"- a former convict for attempted robbery. And the capitalism-fueled metal golem immediately crushes a homeless woman pushing a humble bicycle. I mean, it could have been written as a satire of our times. Except, it really happened.

Re: Uber vehicle reportedly saw but ignored woman it struck

#85

Post by TheCoelacanth » Mon, May 07, 2018, 7:34 PM UTC

freditup wrote: post #18

I'm curious what degree of "cultural context" is built into automated cars - I feel that there has to be a lot of locale-specific techniques for driving. For example, in NYC, there's nothing at all uncommon about a pedestrian beginning to cross the street in front of you and approaching the area of your lane where you will shortly be. Pedestrians typically then stop about a foot away from your lane, let you drive pas…

There is an obvious solution: don't stop, just slow down to <25 mph. On the streets where it is normal for pedestrians to approach that close, the speed limit is almost always 25 mph or lower and often the typical speed is much lower because of traffic. On streets with higher speeds, if a pedestrian is approaching that close to traffic, they either don't see the car or they have a death wish.

Re: Uber vehicle reportedly saw but ignored woman it struck

#86

Post by 0x00000000 » Mon, May 07, 2018, 7:35 PM UTC

dahart wrote: post #75

Rotdhizon wrote: post #39

Earlier quoted context omitted.

Not related to the article but I want to chime in on your point. Having spent some time in VA, aggressive driving and dangerous scenarios were infrequent from my experience. Then having spent some time in FL, aggressive driving was the norm and driving laws did not exist for most people. So I wonder, do self-driving cars needs something like driving profiles for certain regions, where defensive/aggressive response be…

Personally, I’d rather see driving behavior normalize nationally than cater to regional differences in enforcement. The driving laws in each city & state the US are less different than the locally accepted behaviors, but it would be safer all around if drivers and pedestrians all expected the same things, and if we took traffic & safety laws more seriously. More enforcement of existing traffic laws will save many ord…

It can't and won't happen without significant increase in police and traffic enforcement spending which is not always a popular position and there is little motivation. Some states just can't afford it. But in Maryland for example, 55 almost universally means 75 and if an autonomous car didn't know that they will cause many accidents due to unsafe and aggressive driving even if they are technically not "at fault"

Re: Uber vehicle reportedly saw but ignored woman it struck

#87

Post by stale2002 » Mon, May 07, 2018, 7:35 PM UTC

joncrane wrote: post #28

GenerocUsername wrote: post #22

Earlier quoted context omitted.

And welcome to why self-driving cars will always be 3 years away. I think we should be targeting self driving for long-haul freeway scenarios where the rules are relatively simple and predictable. Trying to control for the infinite number of scenarios in cities is a nightmare which will ultimately bog down the development of any company which makes that its goal.

> And welcome to why self-driving cars will always be 3 years away. It's also possible that human/pedestrian behavior will adapt to the consistent behavior of driverless cars.

On a similar note, the innovation of Serri, and other "AI" programs was not to get computers to understand how humans talk. The innovation was getting humans to speak in a way that computers understand.

Re: Uber vehicle reportedly saw but ignored woman it struck

#88

Post by volkl48 » Mon, May 07, 2018, 7:35 PM UTC

Johnny555 wrote: post #68

freditup wrote: post #18

I'm curious what degree of "cultural context" is built into automated cars - I feel that there has to be a lot of locale-specific techniques for driving. For example, in NYC, there's nothing at all uncommon about a pedestrian beginning to cross the street in front of you and approaching the area of your lane where you will shortly be. Pedestrians typically then stop about a foot away from your lane, let you drive pas…

I think legally, the car has to stop for an approaching pedestrian if it's physically able to stop, so if the pedestrian doesn't "stop about a foot away from your lane" as expected, you don't run into him. Though I don't know the NYC traffic code. In California, I believe 21950 paragraph C applies... the pedestrian that stops just before walking into the path of the car might be violating paragraph B since the driver…

The law as written and the law in practice, are different matters.

NYC would not function with actual adherence to the law.

There is never a break in pedestrians long enough for a car to turn in a fully legal fashion at most intersections, which is why human drivers wait for the largest mass of the built-up crowd to go and then inch their way through the stream of those just arriving at the intersection (with pedestrians seeing what the car is going to do and then flowing around the car). Technically, the pedestrians have the full right of way.

On the pedestrian side, blocks are short, and if you don't jaywalk it would take forever to walk up/downtown.

And if you were to fully split pedestrian and vehicle phases (and get New Yorkers to start adhering to it), you'd just grind the city to a halt because you're moving less people/cars than the "maximum throughput" flows than we have now.

Re: Uber vehicle reportedly saw but ignored woman it struck

#89

Post by Symmetry » Mon, May 07, 2018, 7:36 PM UTC

Maybestring wrote: post #72

Symmetry wrote: post #56

Earlier quoted context omitted.

I think you're overselling the extent of the difficulty here. People from out of state usually manage to get through New York, pedestrians and all, without getting into an accident. A car won't get through a block as quickly as it could if it merely obeys the laws makes sure it doesn't run into anybody. I'm sure this will cause great joy among people crossing the street and consternation among other drivers but the p…

>I think you're overselling the extent of the difficulty here. People from out of state usually manage to get through New York, pedestrians and all, without getting into an accident. At this point, "it's easy for a person to do it" is probably better evidence that it's a hard problem in AI. In this case, I believe it's an inverse learning problem. We observe other agents acting in an environment, deduce their policy,…

That sounds much harder than just stopping if a pedestrian is on a trajectory to enter your lane. I mean, obviously it would be better to deduce their intentions and act according to that but I don't see any reason to say that that's necessary to have useful self driving cars.

Re: Uber vehicle reportedly saw but ignored woman it struck

#90

Post by vkou » Mon, May 07, 2018, 7:37 PM UTC

carlosdp wrote: post #27

leggomylibro wrote: post #13

Earlier quoted context omitted.

As long as Uber's system logged which automatic brake was triggered - or simply failed to log that its own braking system stepped in - it could have saved a life without impacting the data gathered. They didn't put a tarp over the windshield; why get rid of other proven safety features?

You don't generally want multiple autonomous control systems in a single vehicle competing with each other for control. There are scenarios where the less-sophisticated collision avoidance radar could do the unsafe thing where the self-driving system has a more holistic plan. Basically, there's no reason the self-driving tech shouldn't have been way more effective than the collision avoidance radar other than Uber ro…

> You don't generally want multiple autonomous control systems in a single vehicle competing with each other for control.

And yet, every modern car has two such systems, competing for control, all the time.

1. The car's safety systems.

2. The human driver.