- This topic has 81 replies, 16 voices, and was last updated 6 years, 7 months ago by
Rich_cb.
-
CreatorTopic
-
November 7, 2019 at 4:36 pm #30254
CygnusX1
The Register has an article on the findings of the NTSB investigation into the death of Elaine Herzberg who was killed by an Uber test driver who was more interested in looking at her phone than watching the road and being prepared to take control if the AI made a mistake.
https://www.theregister.co.uk/2019/11/06/uber_self_driving_car_death/
-
CreatorTopic
-
AuthorReplies
-
Argus Tuft
Nice to know it’s only
Nice to know it’s only “legal reasons”that are keeping these things off the road. What would legislators know about anything?
FluffyKittenofTindalos
Rich_cb wrote:
Rich_cb wrote:The Uber fatality was the result of the failure of two systems.1. The AI system.
2. The human driver.Using this example to demand the end to the development of system 1 in order to continue using system 2 indefinitely seems a bit short sighted.
The market leader in driverless cars has already launched a completely driverless taxi service in Phoenix, Arizona.
They have also clocked up 10 million driverless miles without, AFAIK, any deaths or serious injuries.
The better solution is to have far fewer cars and keep them away from vulnerable road-users.
And Phoenix Arizona? Really? A very-low-population-density, low latitude city, with high levels of sunlight and a desert climate, wide straight roads laid out on a grid pattern, relatively few pedestrians and clear sightlines? And jaywalking laws. In what possible way is that representative of most of the world’s cities? Get back to me when it works in, say, Moscow. Or even London.
AlsoSomniloquism wrote:
AlsoSomniloquism wrote:hawkinspeter wrote:Also, here’s the info showing the difference between lighting conditions on the road in question and the Uber video: https://arstechnica.com/cars/2018/03/police-chief-said-uber-victim-came-from-the-shadows-dont-believe-it/Unless the original Uber video was deliberately doctored, I would argue that the lights were off along that stretch on that day as only one in the distance at the junction is showing and none of the others that should have appeared are showing at all. I wonder if the NTSB full report will mention that or not. And as for the headlight distance, unfortunately we now have cars with adjustable dipping, supposedly to stop blinding of people if you put too much weight in the back to lift the front. Unfortunately most people set and forget so might have been on lowest setting. That woiuld have been on Uber / Vehicle operator.
However it still doesn’t stop the fact that the ped / cyclist stepped in front of a car to cross a road with the car in view (and probably computer controlled to do the legal speed limit on that stretch) and expected the driver to move out of her way by either braking or switching lanes.
I don’t see where you get that from. It hardly has to be specifically ‘doctored’ – everyone who has ever taken a photo or a video knows that such things are crucially-dependent on the exposure settings (or ASA speed equivalent).
The victim-blaming at the end isn’t worthy of comment.
hawkinspeter
AlsoSomniloquism wrote:hawkinspeter wrote:Also, here’s the info showing the difference between lighting conditions on the road in question and the Uber video: https://arstechnica.com/cars/2018/03/police-chief-said-uber-victim-came-from-the-shadows-dont-believe-it/Unless the original Uber video was deliberately doctored, I would argue that the lights were off along that stretch on that day as only one in the distance at the junction is showing and none of the others that should have appeared are showing at all. I wonder if the NTSB full report will mention that or not. And as for the headlight distance, unfortunately we now have cars with adjustable dipping, supposedly to stop blinding of people if you put too much weight in the back to lift the front. Unfortunately most people set and forget so might have been on lowest setting. That woiuld have been on Uber / Vehicle operator.
However it still doesn’t stop the fact that the ped / cyclist stepped in front of a car to cross a road with the car in view (and probably computer controlled to do the legal speed limit on that stretch) and expected the driver to move out of her way by either braking or switching lanes.
The video doesn’t have to be deliberately doctored – it’s just got a different sensitivity/shutter speed and thus doesn’t match what human eyes would perceive. It’s quite common for cameras to record scenes quite differently than you would expect as human eyes/brain do a lot of image processing so that colours look the same under low light etc.
The headlight issue also confirms that the dashcam was poorly calibrated for night-time footage – it’s discussed in the ArsTechnica link I supplied.
I find your attitude regarding the pedestrian to be quite disturbing – basically it seems like you want a death sentence for anyone who makes a mistake about where they’re walking.
Mungecrundle wrote:
Mungecrundle wrote:Argus Tuft wrote:The collision avoidance systems were turned off because they weren’t working.After all the time and money that’s been poured into autonomous cars we’re nowhere near a vehicle that can travel on any road at any time under any conditions without human input.
And if a human has to be standing by ready to take over at any time he/she may as well be driving the bloody thing!
Unrealistic expectation. No technology has ever been delivered in a perfectly safe form. Think of the history of commercial flight, or closer to home, the introduction of carbon fibre bicycle frames. Far fewer catastrophic failures these days.
My prediction is that some sections of British motorway will be opened to autonomous vehicles for large scale trials within 5 years.
Well, OK , it’s an “unrealistic expectation” – which is why I suspect self-driving cars will never work and why it will turn out to be one of those tech-fads that over-promise and under-deliver and turns out to be a waste of money. Unless they turn all roads into motorways, with everyone not in a car banned from them.
You can’t compare autonomous vehicles to carbon-fibre bike frames! The crucial point, as illustrated by this very case, is that the half-way house of self-driving cars that need a driver ready to jump in, is intrinsically dangerous and can’t be allowed on public roads. So they have to be either ‘in perfectly safe form’ or not allowed on the road.
Rich_cb wrote:
Rich_cb wrote:The Uber fatality was the result of the failure of two systems. 1. The AI system. 2. The human driver. Using this example to demand the end to the development of system 1 in order to continue using system 2 indefinitely seems a bit short sighted. The market leader in driverless cars has already launched a completely driverless taxi service in Phoenix, Arizona. They have also clocked up 10 million driverless miles without, AFAIK, any deaths or serious injuries.One Taxi Service in one city-After how many years -And what busy metropolis were the 10 milion miles clocked up in? Don’t hold your breath waiting for this miracle.
AlsoSomniloquism
hawkinspeter wrote:Also, here’s the info showing the difference between lighting conditions on the road in question and the Uber video: https://arstechnica.com/cars/2018/03/police-chief-said-uber-victim-came-from-the-shadows-dont-believe-it/Unless the original Uber video was deliberately doctored, I would argue that the lights were off along that stretch on that day as only one in the distance at the junction is showing and none of the others that should have appeared are showing at all. I wonder if the NTSB full report will mention that or not. And as for the headlight distance, unfortunately we now have cars with adjustable dipping, supposedly to stop blinding of people if you put too much weight in the back to lift the front. Unfortunately most people set and forget so might have been on lowest setting. That woiuld have been on Uber / Vehicle operator.
However it still doesn’t stop the fact that the ped / cyclist stepped in front of a car to cross a road with the car in view (and probably computer controlled to do the legal speed limit on that stretch) and expected the driver to move out of her way by either braking or switching lanes.
The Uber fatality was the
The Uber fatality was the result of the failure of two systems.1. The AI system.
2. The human driver.Using this example to demand the end to the development of system 1 in order to continue using system 2 indefinitely seems a bit short sighted.
The market leader in driverless cars has already launched a completely driverless taxi service in Phoenix, Arizona.
They have also clocked up 10 million driverless miles without, AFAIK, any deaths or serious injuries.
ktache
Never happen mate, I do like
Never happen kil0ran, I do like to watch Police Interceptors (catching wrong’uns, mostly in cars) and when they do catch banned or licenceless drivers they tend to get a paltry fine and a few extra points on their non existent driving licences. Same with no insurance, the car is seized, but that just means a few extra quid in costs. I do feel for these highly trained and proffessional Interceptors that their ‘result’ is often so derisory.
Mungecrundle]
Unrealistic expectations. You seem to think that a complete and infallible technology is going to be delivered to the public at the first iteration. It isn’t, no new technology ever is. Personally I’ll be more than happy to drive my car to the A1, switch to autopilot for the journey north and be alerted to take back control whenever the system is less than 100% sure of its own decisions or reaches a section of the road system not designated as suitable for autonomous control. As for capability. I suggest you book a Tesla road test. The system available today is more than capable of basic driving in traffic, but you have to keep your hands on the wheel for legal reasons. With regards to snow, maybe you should check out some youtube videos (mostly featuring a boxy Lada for some reason) to see how shit humans are at driving in it.[/quoteArgus</strong></p> <p>[quote=Mungecrundle wrote:[quote=Argus Tuft]i’m not talking about safety -it’s capability that’s the issue. Even now at this late stage of development these things can’t deal with snow and need lane markings to operate. What percentage of the world’s roads have lane markings?
Imagine you’re driving in a rural town and you notice a massive hailstorm brewing up.A human might mount the kerb and shelter under a shop awning until it blows over. Would an autonomous vehicle? 5 years seems optimistic to me.
We’ve different definitions of autonomy.What you’re describing I’d define as glorified cruise control.
I’d only be interested in what we’ve been promised for so long-No controls or instruments,just a keyboard to type your destination into.
What’s on offer now is pretty lame.-You’re sitting there with your hands on the wheel-but you’re not “Driving”? And how much do you pay for that rewarding experience?
kil0ran
All driving bans to carry a
All driving bans to carry a suspended prison sentence equivalent to the length of the ban. Caught driving whilst banned, banged up to enforce the ban.
Simon E
Mungecrundle wrote:
We all know that you don’t need snow to show shit driving. Here’s one that occurred this morning:Mungecrundle wrote:With regards to snow, maybe you should check out some youtube videos (mostly featuring a boxy Lada for some reason) to see how shit humans are at driving in it.https://twitter.com/samuriinbred/status/1192707129121165313
I’d prefer to see the money spent on:
– compulsory eye tests for all drivers and random roadside checks.
– driving retests every 5 years and required when 6 points accrued on licence.
– enforcement of speed limits and police to use dashcam/cycle cam footage. Increase the fines & sell impounded vehicles to pay for more police resources, both on the road and behind the desk.
– automatic ban on 12 points, no stupid excuses like ‘financial hardship’ (including for selfish gits like Alan Partridge). Suck it up, lowlife.
– caught driving when banned: vehicle crushed or impounded and sold, lifetime ban.
(I could probably think of a few more given a bit of time but aware that I’m straying a bit OT so I’ll leave it at that).
brooksby
Argus Tuft wrote:And if a human has to be standing by ready to take over at any time he/she may as well be driving the bloody thing!This!
At present, the machine learning systems don’t seem to be up to everything they’ll encounter in a typical urban (or rural, for that matter) road environment.
How many people are we willing to sacrifice so that the machines can learn, so that the later models are safer and can deal with the Real World?
Mungecrundle
Argus Tuft wrote:
Argus Tuft wrote:i’m not talking about safety -it’s capability that’s the issue. Even now at this late stage of development these things can’t deal with snow and need lane markings to operate. What percentage of the world’s roads have lane markings?
Imagine you’re driving in a rural town and you notice a massive hailstorm brewing up.A human might mount the kerb and shelter under a shop awning until it blows over. Would an autonomous vehicle? 5 years seems optimistic to me.
Unrealistic expectations. You seem to think that a complete and infallible technology is going to be delivered to the public at the first iteration. It isn’t, no new technology ever is. Personally I’ll be more than happy to drive my car to the A1, switch to autopilot for the journey north and be alerted to take back control whenever the system is less than 100% sure of its own decisions or reaches a section of the road system not designated as suitable for autonomous control.
As for capability. I suggest you book a Tesla road test. The system available today is more than capable of basic driving in traffic, but you have to keep your hands on the wheel for legal reasons.
With regards to snow, maybe you should check out some youtube videos (mostly featuring a boxy Lada for some reason) to see how shit humans are at driving in it.
i’m not talking about safety
i’m not talking about safety -it’s capability that’s the issue. Even now at this late stage of development these things can’t deal with snow and need lane markings to operate. What percentage of the world’s roads have lane markings?
Imagine you’re driving in a rural town and you notice a massive hailstorm brewing up.A human might mount the kerb and shelter under a shop awning until it blows over. Would an autonomous vehicle? 5 years seems optimistic to me.
-
AuthorReplies
- You must be logged in to reply to this topic.