So how would a self drive car deal with a cat?
Discussion
Dracoro said:
What if a robot (self drive car or whatever) was in the position whereby it needed to take action to avoid killing human A, however that action killed human B. Either way a human dies but how does the robot choose which?
Humans can make that decision, probably on emotional grounds (e.g. You save your child instead of the stranger). What if the robot is owned by you, should/could it be programmed to save your child over others?
Humans don't make those decisions in a reactive situation. They just react. These ethical conundrums are a red herring IMO, the car will try to avoid if there is clear space to do so, otherwise it will try to stop. If it can do neither it will slow as much as it can to minimise the damage.Humans can make that decision, probably on emotional grounds (e.g. You save your child instead of the stranger). What if the robot is owned by you, should/could it be programmed to save your child over others?
They also miss the point that they are incredibly rare. Combinsations of crowd of childrens next to cliffs with cats and/or mothers with prams running into the road do not happen often, compared with the many normal shunts which driverless cars will be able to prevent.
Sump said:
Why are old people constantly against advances in technology? Does it scare them?
Age has nothing to do with it. I'm fairly young - but my work history is more or less chock-full of working with expensive computer systems that were really nice ideas on paper which would save us all time, money, even save lives in certain cases. Some have worked - some have been woeful catastrophes. Hence a certain degree of healthy pragmatism- just because something is new/high tech - doesn't mean it'll actually work all that well in practice. Driver-less cars offer fantastic scenarios of better road safety, less congestion, better management of traffic, fuel savings, etc. BUT, there are lots of variables and scenarios that engineers and legal people need to get right before that happens. The first time someone is inevitably killed by one - it'll be subject to FAR more scrutiny (rightly or wrongly) than if a human were behind the wheel - it has to be beyond reproach. The technology has to be able to undertake split-second problem-solving as well as a human if not better (e.g. bag blown into round - is it solid, etc.), it has to make moral/ethical judgements that will stand up in a court of law or inquest - heck.... the actual mechanics of guiding a car autonomously are the easy bits.
Not just that, they need to get them damn-near perfect before us pragmatists ever use them out of choice - a complex computer modelling system crashing on me is plain irritating, wastes my time, time of my boss's etc. - my car crashing on me is potentially terminal..... =)
CrutyRammers said:
Humans don't make those decisions in a reactive situation. They just react. These ethical conundrums are a red herring IMO, the car will try to avoid if there is clear space to do so, otherwise it will try to stop. If it can do neither it will slow as much as it can to minimise the damage.
They also miss the point that they are incredibly rare. Combinsations of crowd of childrens next to cliffs with cats and/or mothers with prams running into the road do not happen often, compared with the many normal shunts which driverless cars will be able to prevent.
There is also the consideration that the car will foresee and calculate events such that it doesn't end up in the position where it has to 'choose'.They also miss the point that they are incredibly rare. Combinsations of crowd of childrens next to cliffs with cats and/or mothers with prams running into the road do not happen often, compared with the many normal shunts which driverless cars will be able to prevent.
Look at the example of the Tesla in the Netherlands recently. It avoided the accident before it had even happened, so a 'choice' of whether to pile into the vehicle in front or head for the crash barrier was a decision it didn't have to make. A human would probably have ended up having to make that reactive choice
(incident in question, note the emergency brake alarm triggering about half a second before the impact of the vehicle in front)
https://www.youtube.com/watch?v=FadR7ETT_1k
For many folk I speak to regarding some pretty straight forward aspects of AI, machine learning and suchlike, there's a perception of "I don't understand how it can be possible so I assume it cannot be possible"
Edited by feef on Tuesday 28th March 15:12
mikeyr said:
Winning NASCAR races would be much easier I'd have thought so let's work out how long it would take for an automated car to win one of them. I'd wager 5 years max if all the current effort was diverted that way?
If the rules allowed it, an automated car would win the F1 world championship next year. 100% consistent lap times, millisecond reaction times, never gets tired or emotional, the ability to sense the slip levels of each individual wheel. All it would take is a few practice laps for it learn the circuit and no human driver would get near.Mr Will said:
If the rules allowed it, an automated car would win the F1 world championship next year. 100% consistent lap times, millisecond reaction times, never gets tired or emotional, the ability to sense the slip levels of each individual wheel. All it would take is a few practice laps for it learn the circuit and no human driver would get near.
It's coming. There already is an autonomous vehicle series in parallel with Formula E, called RoboRace:http://www.theverge.com/2017/2/27/14749740/roborac...
They raced two of the prototypes for the first time in Buenos Aires the other week:
http://www.theverge.com/2017/2/21/14669028/roborac...
Note that one of the vehicles "successfully dodged a dog who wandered onto the track" and the other went beyond the limits of it's software and had a bit of an off, but it shows where things have already got to, and this is VERY early days yet.
Zigster said:
Driverless cars don't have to be flawless, just safer than having a human at the wheel.
I'm not so sure that's true. I heard someone on tv say recently that driverless cars will be twice as safe as humans. Well twice as safe is a polite way of saying half as dangerous. I think Joe public accepts 1800 deaths a year due to human error. But I don't think he'll accept 900 deaths a year due to computer shortcomings.TwigtheWonderkid said:
I'm not so sure that's true. I heard someone on tv say recently that driverless cars will be twice as safe as humans. Well twice as safe is a polite way of saying half as dangerous. I think Joe public accepts 1800 deaths a year due to human error. But I don't think he'll accept 900 deaths a year due to computer shortcomings.
Anyone who grew up with Bill Gates' tech will be perfectly used to their car crashing twice a day for no reason. It's certainly going to be an interesting time on the roads with driverless cars. I'm all for it but still there will be limitations.
As has been said how will an automated car deal with an empty McDonalds bag either stationary or blown along? Full on brakes and an emergency manoeuvre into the opposing lane with oncoming traffic.
Rain. A huge dirty sheet of water is washed up by a passing HGV. How will a autocar react.
Differences in road surfaces. We've all seen badly repaired roads in contrasting materials (concrete repaired with tarmac and vice versa) and sometimes mistaken contrasts which may or may not be objects. How does autocar cope?
b
d pothole from hell. At low speeds autocar may be OK. How will autocar cope at high speeds?
Roadworks. A few autocars have had issues with lane closures. You have the fully overconed areas and Jimbob with 3 cones cutting down trees.
Parking. tt parkers who create a chicane and leave minimal space to get around. What if tt blocks the road? What does autocar do?
For all the limitations most average drivers with their eyes and limbs deal OK with incidents.
As has been said how will an automated car deal with an empty McDonalds bag either stationary or blown along? Full on brakes and an emergency manoeuvre into the opposing lane with oncoming traffic.
Rain. A huge dirty sheet of water is washed up by a passing HGV. How will a autocar react.
Differences in road surfaces. We've all seen badly repaired roads in contrasting materials (concrete repaired with tarmac and vice versa) and sometimes mistaken contrasts which may or may not be objects. How does autocar cope?
b
d pothole from hell. At low speeds autocar may be OK. How will autocar cope at high speeds?Roadworks. A few autocars have had issues with lane closures. You have the fully overconed areas and Jimbob with 3 cones cutting down trees.
Parking. t
t parkers who create a chicane and leave minimal space to get around. What if tt blocks the road? What does autocar do?For all the limitations most average drivers with their eyes and limbs deal OK with incidents.
Sten. said:
I tend to liken it to air travel. I'm pretty sure back in the early days lots of people thought there was too much risk involved, it wasn't safe and they couldn't ever see it taking off (pun intended). Today, millions of us travel by air every year even though it's not perfect and people do occasionally die.
I bet not many people would get in a self flying plane though. I sure wouldn't, the only way people would trust self driving cars without someone sitting there with a hand on the controls would be something like the shuttle thing at Birmingham airport to the nec. The technology isn't there yet at all, not by a long way, at the moment it's just geeks playing.If you need to have someone there to be in control just in case then you might as well just have a driver. Frankly I just don't get the point of it.
Zigster said:
Driverless cars don't have to be flawless, just safer than having a human at the wheel. I doubt that's very far in the future.
WHEN it does happen - and it seems inevitable given the big players seem obsessed with it - how many fatal accidents will the likes of UBER stomach before they pull them?Currently, if an UBER driver mills someone its unlikely to make much more than local press. The first fatal accident with a driverless UBER car (or google or Apple) will make worldwide news, as will the second.... third then people will start to avoid using them, then profits will be down....
DonkeyApple said:
The Vambo said:
vikingaero said:
Rain. A huge dirty sheet of water is washed up by a passing HGV. How will a autocar react.
Autocar would do what they always do...... publish a crap artists rendering of a McLaren/Lotus then award Jaguar car of the year.
As an Autocar reader since the age of about 7 to a few years ago this is so true.
For me Jag were always Big Billy Bulls
tters for decades. They used to launch concept cars galore with the motoring press masturbating in delight and fail to deliver in every way. For that reason alone I've never liked Jags.Have all self driving cars issued with an additional manual:
But cats are the tip of the iceberg. What about a self drive car dealing with a charging rhino?
But cats are the tip of the iceberg. What about a self drive car dealing with a charging rhino?
Is this even a thing? I haven't noticed a cat holocaust on the roads and have never seen a cat come anywhere near getting run over by a car, cats are going to be no more prone to being flattened by Google than they are Mrs Miggins driving home from Sainsburys. Darwinism will take care of the rest.
EazyDuz said:
Pointless thread tbh, it wont happen until the 2200's. It needs intelligent AI to become fool proof, which is at least 100 years off
200 years?You think it will take the same duration from when the telegraph was invented until now, as from now until the the first full autonomous vehicle?
Crazy talk I tellya
Gassing Station | General Gassing | Top of Page | What's New | My Stuff