Armed Polite Society
Main Forums => The Roundtable => Topic started by: K Frame on March 19, 2018, 01:17:59 PM
-
A self-driving Uber ran down and killed an Arizona woman.
I don't think it was an accident.
Of course the story (blurb really) came up on my phone app, but they've not yet updated the home page on the internet to include it.
-
https://www.nytimes.com/2018/03/19/technology/uber-driverless-fatality.html
http://money.cnn.com/2018/03/19/technology/uber-autonomous-car-fatal-crash/index.html
https://www.npr.org/sections/thetwo-way/2018/03/19/594950197/uber-suspends-self-driving-tests-after-pedestrian-is-killed-in-arizona
It is in the news.
-
Instead of walking, she should have taken an Uber.
-
It looks like they had a human in the driver's seat as a safeguard. I guess that didn't work out well. Wonder if the person was even paying attention.
-
It looks like they had a human in the driver's seat as a safeguard. I guess that didn't work out well. Wonder if the person was even paying attention.
Probably not. It's really tough to have people on watch "just in case" where they don't normally need to do anything. Witness the Tesla fatality where the human involved was actively not paying any attention.
Personally, it is looking like Uber is running their self-driving car program like they do the rest of their company - pushing the edges, slightly incompetent.
They've had more reported incidents, and more serious ones, than the rest of their competitors combined.
-
https://www.usatoday.com/story/tech/science/2018/03/15/walmart-robot-drone-bees-patent/428935002/
If the robot uprising has bee-gun, imagine what these little things could do.
-
From the above article:
Walmart files hundreds of patents each year, many of which never go beyond the "what if?" stage,
I found this kind of interesting, maybe they are trying to corner the market on future stuff, or at least make someone pay if they want to go forward with some technology.
bob
-
Instead of walking, she should have taken an Uber.
That is quite the business plan...if you are walking then you are not using our service so you must die.
-
Dash Cam video has been released.
https://thenextweb.com/artificial-intelligence/2018/03/22/dashcam-footage-fatal-uber-self-driving-car-accident/
Looking at the video, I don't know if I'd have been able to stop in time, even as I'm specifically looking for someone with a bike to pop up in front of me.
I'd have thought that the self-driving cars can actually see better than I can, but apparently, they also are unable to see thing outside of headlight range.
Also, who walks a bike in the middle of the night SLOWLY across a highway?
-
Reading a bit on the background of the "driver", I was actually quite surprised to see that Uber appears to be using the equivalent of "burger flippers" as their test drivers. I would have assumed they would be using people with expertise in the technology to assist in realtime diagnosis, and also have some minimum requirements on the "paying attention" front - especially at night.
I can totally see the "driver" taking their eyes off the road for extended periods in daylight on a desert highway to monitor diagnostic equipment, but not in the city or at night. In that case it seems like they should have two people in the vehicle - one paying 100% attention to the road, and one paying 100% attention to diagnostic stuff.
-
That would cost money.
Of course a person is now dead but hey, that's just collateral damage on the road of progress.
Kind of sucks to be that person's family but I'm sure the paltry couple of grand they eventually get after multiple years of litigation will assuage the minuscule (if any at all) amount of guilt felt by the officers of the company.
-
Dash Cam video has been released.
https://thenextweb.com/artificial-intelligence/2018/03/22/dashcam-footage-fatal-uber-self-driving-car-accident/
Looking at the video, I don't know if I'd have been able to stop in time, even as I'm specifically looking for someone with a bike to pop up in front of me.
I'd have thought that the self-driving cars can actually see better than I can, but apparently, they also are unable to see thing outside of headlight range.
Also, who walks a bike in the middle of the night SLOWLY across a highway?
When I saw that dash cam video this morning my first thought was there was no way I’d have been able to stop in time. It also looked like they were crossing the street in a strange place; not a normal spot someone would cross a street.
The “driver” was obviously not paying attention to the road, almost looked like texting or checking a phone; but could have been something they do while monitoring the car too. It would probably be difficult to pay attention in the same way as when driving though just sitting in the car not doing anything.
-
There is probably one guy in the driver's seat playing on his phone or reading his kindle and another diagnostic guy at a remote site to monitor multiple vehicles who is also playing with his phone or reading his kindle.
-
When I saw that dash cam video this morning my first thought was there was no way I’d have been able to stop in time. It also looked like they were crossing the street in a strange place; not a normal spot someone would cross a street.
The “driver” was obviously not paying attention to the road, almost looked like texting or checking a phone; but could have been something they do while monitoring the car too. It would probably be difficult to pay attention in the same way as when driving though just sitting in the car not doing anything.
Okay, I hadn't seen the video. That is a fast road, not extremely well lit, and the pedestrian didn't seem to be paying a lot of attention either.
-
That would cost money.
The "burger flipper" alone would be fine if they had some minimal training and their sole job was acting as if they were driving the vehicle, just as if it was not automated. All the diagnostic stuff is being recorded and maybe realtime transmitted anyway.
I think the self-driving experimenters have just become too complacent too early. They probably figured they have a 97% success rate so they can slack off. IMO, you don't do that until 99.5% for something like this.
-
I've heard a couple of times now that manned vehicle crashes have a fatality rate of one per 100 million miles driven. They're saying the unmanned (self-crashing =) ) cars have been driven only about 8 million miles (combined). So I guess they have to make it to 108 million miles without a fatality to be as safe as a manned vehicle?
I did find a source on fatality rates.
https://www.rand.org/news/press/2016/04/12.html
Although the total number of crashes, injuries and fatalities from human drivers is high, the rate of these failures is low in comparison with the number of miles that people drive. Americans drive nearly 3 trillion miles every year, according to the Bureau of Transportation Statistics. In 2013, there were 2.3 million injuries reported, which is a failure rate of 77 injuries per 100 million miles driven. The related 32,719 fatalities correspond to a failure rate of about 1 fatality per 100 million miles driven.
“The most autonomous miles any developer has logged are about 1.3 million, and that took several years. This is important data, but it does not come close to the level of driving that is needed to calculate safety rates,” said Susan M. Paddock, co-author of the study and senior statistician at RAND. “Even if autonomous vehicle fleets are driven 10 million miles, one still would not be able to draw statistical conclusions about safety and reliability.”
-
Reading a bit on the background of the "driver", I was actually quite surprised to see that Uber appears to be using the equivalent of "burger flippers" as their test drivers. I would have assumed they would be using people with expertise in the technology to assist in realtime diagnosis, and also have some minimum requirements on the "paying attention" front - especially at night.
Burger-flipper drivers - hmmmm. A self-driving food truck. Now that has potential. Is the cooking autonomous?
-
Burger-flipper drivers - hmmmm. A self-driving food truck. Now that has potential. Is the cooking autonomous?
Pre-cooked, a rolling Automat (https://en.wikipedia.org/wiki/Automat). Drive to predetermined points like constructions sites, hospitals, etc, spend 20 minutes or less if no purchases in 8 minutes or so and then off to the next spot.
bob
-
It looks like they had a human in the driver's seat as a safeguard. I guess that didn't work out well. Wonder if the person was even paying attention.
In-car video shows that he wasn't. However, it was at night and the forward-facing video shows that she wasn't visible until an instant before impact. It's unlikely the safety driver could have seen her in time to react anyway.
https://www.youtube.com/watch?v=XtTB8hTgHbM
-
I assume that driver-less technology relies on GPS for choosing a route? If so, it can't work for me. For some obscure reason, my TomTom GPS thinks the entrance to my property isn't on the street where my address, mailbox and driveway are located, but on a parallel street behind my house. Any time I fire up the GPS in the driveway, it starts off telling me to proceed straight for 100 yards. That would require driving through the rear wall of the garage, down a steep hill, through a section of woods, across a stone wall, and through a neighbor's yard.
I think I'll hold off on the self-driving thing.
-
There is an obvious solution to this problem; if we all had some form of rfid/gps trackers installed the autonomous vehicles could track us. What could possible go wrong. Besides it’s for the children!
-
I figure that autonomous vehicles will at some point be designed to communicate between each other within a certain distance in order to be most efficient. If my car can communicate with the car in front of me, we can achieve a safe following distance approaching zero. Traffic lights can become much more efficient, adjusting the speed of approaching vehicles so that they mesh with cross traffic as they pass through intersections, possibly requiring much less frequent full stops. I suppose this communication might also extend to alerting the pack if a particular car in the pack is about to malfunction or crash. I also assume the cars may decide among themselves whether a particular car might need to "sacrifice" itself to achieve the lowest possible number of casualties of an incident. Your car, which is only carrying you (maybe a chain smoking 80 year old conservative gun owner), may be chosen to run itself off the cliff so that the passenger van full of nuns on the way to a Planned Parenthood seminar may be spared.
Are you okay with this new technology that's coming our way?
-
I assume that driver-less technology relies on GPS for choosing a route? If so, it can't work for me. For some obscure reason, my TomTom GPS thinks the entrance to my property isn't on the street where my address, mailbox and driveway are located, but on a parallel street behind my house. Any time I fire up the GPS in the driveway, it starts off telling me to proceed straight for 100 yards. That would require driving through the rear wall of the garage, down a steep hill, through a section of woods, across a stone wall, and through a neighbor's yard.
I think I'll hold off on the self-driving thing.
That's really more of a TomTom thing than problem with all of GPS tech. I rank TomTom devices slightly above an enthusiastic mentally handicapped person riding shotgun.
-
We need four people monitoring these cars at all times! One person for each side. And maybe put a fifth up top in some sort of cupola with something to warn/chase people out of the path of the vehicle.
And maybe a sixth to be constantly on the radio back to base with updates.
And of course a car full of technicians and mechanics following it to make repairs if needed.
With that many people you probably need some folks from HR to make sure the interpersonal stuff goes smoothly so they'll need their own car.
-
Does the car have that radar stuff the new car commercials advertise that is supposed to apply the breaks if it sees something in your path? It seems to me something like that would have helped if it wasn't there.
-
We need four people monitoring these cars at all times! One person for each side. And maybe put a fifth up top in some sort of cupola with something to warn/chase people out of the path of the vehicle.
And maybe a sixth to be constantly on the radio back to base with updates.
And of course a car full of technicians and mechanics following it to make repairs if needed.
With that many people you probably need some folks from HR to make sure the interpersonal stuff goes smoothly so they'll need their own car.
I feel like I've seen this in a political cartoon once somehow.
-
Does the car have that radar stuff the new car commercials advertise that is supposed to apply the breaks if it sees something in your path? It seems to me something like that would have helped if it wasn't there.
I believe it did have that technology, and they're trying to figure out why it didn't work.