Volvo will claim full liability in case of self-driving car crash

[Curufinwe_wins]

Really, you'd pick something with no sense or right or wrong, no sense of good or bad and no conscience, compassion or emotion over a human being? WOW dude. A machine would unplug your life support without even considering the humanity side of the argument, a simple calculation of the chances of you making a recovery and its wham, bam, thank you mam, your no longer alive, the machine has killed you off because you entrusted it with your life. Have movies taught you nothing at all? Computers = good at calculation, really really bad at emulating humans.

 

Why are aeroplanes safer? Well for a start you FLY a plane, not drive one but English lesson aside its simple mathematics. When there's 100x more people driving around every day than there is flying around every day then your simply going to see and hear about 100x more car crashes than plane crashes. Pilot Error still attributes for a huge margin of plane crashes around the world and you better believe that when a plane goes it takes a lot of people along for the ride, something which is not a factor in a car crash.

 

Besides that, have you not noticed your self contradicting statement? Pilots are humans too so why is it somehow better to have a human fly a plane than it is driving a car? I mean its not like anybody can just drive a car without training first so I'm just wondering what exactly is the difference here?

 

 

I mentioned the number game in 1. but they are safer per person by numerous orders of magnitude. 2 has to deal with the massive requirement differences in driving license and flying. Even semi-drivers are worse off because the people around them don't have nearly the amount of training/skills on the road as they do. Pilots don't have that issue.

 

The point isn't to emulate human action. Humans are bad at this job. The goal is to get humans off the road all together because 99.999% of all accidents (BS number that basically is unity. We cause all of the accidents that basically don't include a tree/object falling on top of us on the road.) are completely avoidable if a human wasn't behind the wheel fucking up (most weather related accidents even are avoidable with proper technique, and reaction times for a computer are infinitely faster than for a human).

[Trik'Stari]

You people can whine about my preferences all you want, it isn't going to change my stance on the matter so there's no point in complaining or trying to persuade me to think something else. I had no idea this topic had so many big babies surrounding it, sorry for insulting you all.

 

@Matu20 I'll do as I please, which isn't watching what you're telling me to watch. You aren't my mother, buddy.

All of the people playing the "humans can't drive well" card probably have no idea how to drive. They probably have no idea how hard they can brake and turn before their vehicle loses traction.

[Alurit]

You people can whine about my preferences all you want

we don't have problem with your preference, we have problem with you calling self driving cars "only bad thing that leads to bad things", you want to drive yourself fine, I don't think anyone will be forced out of driving in 50 years, but don't say it is only bad, when every data suggests it is not

[Curufinwe_wins]

All of the people playing the "humans can't drive well" card probably have no idea how to drive. They probably have no idea how hard they can brake and turn before their vehicle loses traction.

Yep I have no idea how to drive... Thanks for pointing it out.

 

Also I love how in certain states a dusting of snow causes 100s of accidents, yet where I am from feet of slush and ice on the roads is not only expected but dealt with very well.

 

Anything a human can do when it comes to driving a computer can do better, safer. Esp considering they don't have response times on the order of a second.

[Trik'Stari]

Yep I have no idea how to drive... Thanks for pointing it out.

 

Also I love how in certain states a dusting of snow causes 100s of accidents, yet where I am from feet of slush and ice on the roads is not only expected but dealt with very well.

 

Anything a human can do when it comes to driving a computer can do better, safer. Esp considering they don't have response times on the order of a second.

Except for deciding whether or not to hit that pedestrian, or run into something and kill the passengers in the vehicle. Machines can't make moral choices beyond how we program them. An automated vehicle will only be capable of driving as well as it's programmed to. The same thing applies to humans, if you have very little experience, you will have less skill, and thus be a more dangerous driver. The main problem is that in the US, our drivers ed programs are a complete joke. My "in car" drivers ed teacher drove like an old woman, way too slow, and didn't teach us ANYTHING about evasive/emergency driving (how to run off the road and not flip the car, how to correct during a slide without over-correcting, etc).

 

And as I said, the problem isn't that drivers are human, it's that there are a lot of moronic humans that have no business driving, on the roads. If people want to switch to automated cars that's fine with me, but I'm going to keep driving myself, because I'm more then comfortable behind the wheel. I'm generally not comfortable with other people driving.

[Master Disaster]

I mentioned the number game in 1. but they are safer per person by numerous orders of magnitude. 2 has to deal with the massive requirement differences in driving license and flying. Even semi-drivers are worse off because the people around them don't have nearly the amount of training/skills on the road as they do. Pilots don't have that issue.

 

The point isn't to emulate human action. Humans are bad at this job. The goal is to get humans off the road all together because 99.999% of all accidents (BS number that basically is unity. We cause all of the accidents that basically don't include a tree/object falling on top of us on the road.) are completely avoidable if a human wasn't behind the wheel fucking up (most weather related accidents even are avoidable with proper technique, and reaction times for a computer are infinitely faster than for a human).

 

OK bud, can I pose a hypothetical question, one which (AFAIK) was first asked by non other than Jeremy Clarkson from Top Gear.

 

If you have a self driving car with a driver and passengers inside it and it gets itself into a crash situation, could be any situation you care to pose but Clarkson used the following...

 

 

 

If the car were to get into a situation where it had to decide between the lives of the passengers or the life of a civilian outside of the vehicle, what exactly would it do?

 

The answer? Simple, it would only be able to deal with it it if had SPECIFIC programming of that SPECIFIC situation otherwise it would kill the pedestrian. Now that means the passengers would have to sit there and watch their car kill a pedestrian with zero chance of being able to stop it from happening.

 

Ignoring the legalities for a second, self driving cars essentially mean your vehicle can kill someone without you being able to stop it from happening. Its exactly because humans can evaluate a situation and come up with a way around it on the fly that humans make the best pilots of vehicles you can ever get. You say human training is the deciding factor, consider that a computer cannot make a decision its not pre programmed to make. You can devise a 100 trillion line of code set of rules for a car, its still doesn't guarantee the computer isn't going to encounter a situation it cannot deal with.

 

On to the legality, whose to blame when your car kills a ped? I really need to say nothing more than that really. Sure Volvo say they'll take the blame but until it happens there is no answer to this question.

 

Finally the human factor, imagine the trauma of watching helpless as your car kills a pedestrian with you sat inside it as it happens.

 

You mentioned Aeroplanes earlier, I'm sure your aware (as you seem like an intelligent dude/ette) that its 100% possible for a plane to take off, fly to its destination and then land when it arrives by autopilot however they still make humans take off, do any turns and other manoeuvrers and land, even when the computer could do it better its still mandatory for a human to do these things. Why is that? Because if something goes wrong the computer does the wrong thing and everybody dies. I once saw a simulation of a Boeing 777 where they demonstrated the Autopilot, they then showed what happens when the most advanced auto pilot system in the world encounters a mechanical error. At first it compensated and did the right thing however it wasn't able to throttle control the engines in the way needed to make the plane fly straight again, instead it did the wrong thing entirely and stalled the engines.  They repeated the test again with the human at the helm and by using the yoke and the throttle he was able to regain control of the plane and get it back to the airport. Now I'm aware this agrees with your pilot training statement however it also demonstrates perfectly my point too, a human can adapt on the fly to anything it encounters, it might not be correct but its at least something, if a computer does the same it will almost always do the wrong thing.

 

Sorry for the wall

[qwertywarrior]

cant wait for this

maybe 3 years ?

[Trik'Stari]

OK bud, can I pose a hypothetical question, one which (AFAIK) was first asked by non other than Jeremy Clarkson from Top Gear.

 

If you have a self driving car with a driver and passengers inside it and it gets itself into a crash situation, could be any situation you care to pose but Clarkson used the following...

 

 

 

 

The answer? Simple, it would only be able to deal with it it if had SPECIFIC programming of that SPECIFIC situation otherwise it would kill the pedestrian. Now that means the passengers would have to sit there and watch their car kill a pedestrian with zero chance of being able to stop it from happening.

 

Ignoring the legalities for a second, self driving cars essentially mean your vehicle can kill someone without you being able to stop it from happening. Its exactly because humans can evaluate a situation and come up with a way around it on the fly that humans make the best pilots of vehicles you can ever get. You say human training is the deciding factor, consider that a computer cannot make a decision its not pre programmed to make. You can devise a 100 trillion line of code set of rules for a car, its still doesn't guarantee the computer isn't going to encounter a situation it cannot deal with.

 

On to the legality, whose to blame when your car kills a ped? I really need to say nothing more than that really. Sure Volvo say they'll take the blame but until it happens there is no answer to this question.

 

Finally the human factor, imagine the trauma of watching helpless as your car kills a pedestrian with you sat inside it as it happens.

 

You mentioned Aeroplanes earlier, I'm sure your aware (as you seem like an intelligent dude/ette) that its 100% possible for a plane to take off, fly to its destination and then land when it arrives by autopilot however they still make humans take off, do any turns and other manoeuvrers and land, even when the computer could do it better its still mandatory for a human to do these things. Why is that? Because if something goes wrong the computer does the wrong thing and everybody dies. I once saw a simulation of a Boeing 777 where they demonstrated the Autopilot, they then showed what happens when the most advanced auto pilot system in the world encounters a mechanical error. At first it compensated and did the right thing however it was able to throttle control the engines in the way needed to make the plane fly straight again, instead it did the wrong thing entirely and stalled the engines.  They repeated the test again with the human at the helm and by using the yoke and the throttle he was able to regain control of the plane and get it back to the airport. Now I'm aware this agrees with your pilot training statement however it also demonstrates perfectly my point too, a human can adapt on the fly to anything it encounters, it might not be correct but its at least something, if a computer does the same it will almost always do the wrong thing.

 

Sorry for the wall

^This.

 

The human mind is more adaptive in emergency situations than any computer. The problem we have with driving is that most people barely know how to drive by the time they've gotten their drivers license. Along with old people who drive way too slow.

[TetraSky]

I wonder how well will self driving cars perform in terrible weather... Like icy roads, hail, snowstorms, etc...

[Curufinwe_wins]

Except for deciding whether or not to hit that pedestrian, or run into something and kill the passengers in the vehicle. Machines can't make moral choices beyond how we program them. An automated vehicle will only be capable of driving as well as it's programmed to. The same thing applies to humans, if you have very little experience, you will have less skill, and thus be a more dangerous driver. The main problem is that in the US, our drivers ed programs are a complete joke. My "in car" drivers ed teacher drove like an old woman, way too slow, and didn't teach us ANYTHING about evasive/emergency driving (how to run off the road and not flip the car, how to correct during a slide without over-correcting, etc).

 

And as I said, the problem isn't that drivers are human, it's that there are a lot of moronic humans that have no business driving, on the roads. If people want to switch to automated cars that's fine with me, but I'm going to keep driving myself, because I'm more then comfortable behind the wheel. I'm generally not comfortable with other people driving.

By the very fact of being machines who do the same thing every time, always do things on purpose as you put it, they will get into dramatically fewer of those moral dilemma's in the first place. Furthermore the entire system can be reworked to take advantage of machine's tighter tolerances.

 

Example: When traffic gets heavier people slow down. This is the WRONG ACTION. Slowing down lowers the volume of cars that can get through the section which further increases traffic. We create a self-reinforcing issue (which is why traffic gets so bad.) The proper action would be to maintain speed or actually INCREASE speed, to allow a greater volume of traffic to flow. Why don't we do this? Well first, we are particularily bad at making the right decision under stress, and more than making the right decision, we are inconsistent, and indeed not knowing what the people around you driving are doing/going to do is one of the biggest reasons people slow down in the first place (it gives them more time to react to a changing situation). A network of driver-less vehicles could communicate all proposed behaviors/actions (or even just use a consistent behavior set) and remove that dilemma.

 

Another good example of human's consistently being WRONG is decreasing lanes or passing obsticles. You know those times when you are going 65 MPH (for example) in a three lane, as you get to the drop to two lanes your speed decreases to next to nothing, then once you pass it the speed goes back up to basically where it was n the first place? Those are caused by human errors piling up (or human actions that detrimentally affect the flow but were made on purpose aka selfishness). A network of driverless computers (again primarily due to a known expected behavior set) could convert that merger miles in advance of the event (indeed over a course of miles if needed as well, not waiting till the last second) completely eliminating the flow issue.

 

Car traffic is like a flowing fluid, we are clotting mechanisms (most of the time as individuals we are fine, but mistakes happen and clots form or constricted pathways lead to clot buildups), machines don't have that issue. They don't make "mistakes".

[mattebad]

OK bud, can I pose a hypothetical question, one which (AFAIK) was first asked by non other than Jeremy Clarkson from Top Gear.

 

If you have a self driving car with a driver and passengers inside it and it gets itself into a crash situation, could be any situation you care to pose but Clarkson used the following...

 

 

 

 

The answer? Simple, it would only be able to deal with it it if had SPECIFIC programming of that SPECIFIC situation otherwise it would kill the pedestrian. Now that means the passengers would have to sit there and watch their car kill a pedestrian with zero chance of being able to stop it from happening.

 

Ignoring the legalities for a second, self driving cars essentially mean your vehicle can kill someone without you being able to stop it from happening. Its exactly because humans can evaluate a situation and come up with a way around it on the fly that humans make the best pilots of vehicles you can ever get. You say human training is the deciding factor, consider that a computer cannot make a decision its not pre programmed to make. You can devise a 100 trillion line of code set of rules for a car, its still doesn't guarantee the computer isn't going to encounter a situation it cannot deal with.

 

On to the legality, whose to blame when your car kills a ped? I really need to say nothing more than that really. Sure Volvo say they'll take the blame but until it happens there is no answer to this question.

 

Finally the human factor, imagine the trauma of watching helpless as your car kills a pedestrian with you sat inside it as it happens.

 

You mentioned Aeroplanes earlier, I'm sure your aware (as you seem like an intelligent dude/ette) that its 100% possible for a plane to take off, fly to its destination and then land when it arrives by autopilot however they still make humans take off, do any turns and other manoeuvrers and land, even when the computer could do it better its still mandatory for a human to do these things. Why is that? Because if something goes wrong the computer does the wrong thing and everybody dies. I once saw a simulation of a Boeing 777 where they demonstrated the Autopilot, they then showed what happens when the most advanced auto pilot system in the world encounters a mechanical error. At first it compensated and did the right thing however it was able to throttle control the engines in the way needed to make the plane fly straight again, instead it did the wrong thing entirely and stalled the engines.  They repeated the test again with the human at the helm and by using the yoke and the throttle he was able to regain control of the plane and get it back to the airport. Now I'm aware this agrees with your pilot training statement however it also demonstrates perfectly my point too, a human can adapt on the fly to anything it encounters, it might not be correct but its at least something, if a computer does the same it will almost always do the wrong thing.

 

Sorry for the wall

This is exactly why I believe there needs to be a human element to self driving cars. If a sensor on my car stops working and the failsafe on it fails to stop the car, what do I do? Turning your car off isn't an option at 40mph. If the car slams it's breaks because an animal jumps in front of the car instead of swerving you are dead. If it isn't inputting my destination for some reason, how am I supposed to use it until it can get fixed, you can't. These are valid concerns that can easily be addressed by installing a sort of manual takeover mode like from that scene in iRobot.

 

 

Or people who are into car culture and genuinely enjoy driving.

And while automated cars interest me somewhat, I do happen to enjoy driving and I am very good at it.

Also, keep in mind the same companies making these automated cars have had many, many issues with regular cars. Automation is no guarantee of safety when they can't even get the basics right. Air bags filled with shrapnel, throttles sticking, ignition issues, my car had a recall last year for possible fires due to oil leaks, the list goes on. I mean we can't even trust VW to not lie about emissions let alone safety. Look at the GM ignition deaths. A simple ignition switch wasn't even done right and many died for it.

In the US, driving is not a right, only a privilege so we may see it taken away one day. But hopefully by then they will be able to build a machine I trust fully. That time isn't here yet.

And I am not sure SCCA events would be as fun.

That is exactly how I feel. I'm a gearhead, I love cars, I love watching them on the track and car shows in parking lots. I love the sound of a hard acceleration when getting on a highway or making those tight turns on swervy roads. I'm not selfish for saying that I don't want to watch my hobby be taken away because car manufacturers don't want to put an option for a manual mode.

 

If you think that this will keep autonomous vehicles out of the hands of the masses, then your blind to how technological innovations work and who they cater to. The minority doesn't matter, the masses do, and the masses want autonomous cars @Matu20

[Master Disaster]

I wonder how well will self driving cars perform in terrible weather... Like icy roads, hail, snowstorms, etc...

 

Infinitely better than humans, that kind of quick calculation and fine control is exactly what a computer is good at. Most cars these days use TCS to do exactly that anyway.

 

Self driving cars are really a good thing in almost every way but it only takes one scenario to play out and its a huge catastrophe.

[Matu20]

This is exactly why I believe there needs to be a human element to self driving cars. If a sensor on my car stops working and the failsafe on it fails to stop the car, what do I do? Turning your car off isn't an option at 40mph. If the car slams it's breaks because an animal jumps in front of the car instead of swerving you are dead. If it isn't inputting my destination for some reason, how am I supposed to use it until it can get fixed, you can't. These are valid concerns that can easily be addressed by installing a sort of manual takeover mode like from that scene in iRobot.

 

 

That is exactly how I feel. I'm a gearhead, I love cars, I love watching them on the track and car shows in parking lots. I love the sound of a hard acceleration when getting on a highway or making those tight turns on swervy roads. I'm not selfish for saying that I don't want to watch my hobby be taken away because car manufacturers don't want to put an option for a manual mode.

 

If you think that this will keep autonomous vehicles out of the hands of the masses, then your blind to how technological innovations work and who they cater to. The minority doesn't matter, the masses do, and the masses want autonomous cars @Matu20

The thing is that as far as I can see, the people who believe in self driving cars is the minority.

[Master Disaster]

I'd like to go on record now and say FULLY AUTONOMOUS VEHICLES, defined by a vehicle which requires no user interaction outside of the destination and cannot be overridden by the driver, will never happen. Yeah I'm sure all cars will eventually be able to do self driving in a way which will be great for us all but the driver will ALWAYS be able to take control if needed.

[Curufinwe_wins]

OK bud, can I pose a hypothetical question, one which (AFAIK) was first asked by non other than Jeremy Clarkson from Top Gear.

 

If you have a self driving car with a driver and passengers inside it and it gets itself into a crash situation, could be any situation you care to pose but Clarkson used the following...

 

 

 

 

The answer? Simple, it would only be able to deal with it it if had SPECIFIC programming of that SPECIFIC situation otherwise it would kill the pedestrian. Now that means the passengers would have to sit there and watch their car kill a pedestrian with zero chance of being able to stop it from happening.

 

Ignoring the legalities for a second, self driving cars essentially mean your vehicle can kill someone without you being able to stop it from happening. Its exactly because humans can evaluate a situation and come up with a way around it on the fly that humans make the best pilots of vehicles you can ever get. You say human training is the deciding factor, consider that a computer cannot make a decision its not pre programmed to make. You can devise a 100 trillion line of code set of rules for a car, its still doesn't guarantee the computer isn't going to encounter a situation it cannot deal with.

 

On to the legality, whose to blame when your car kills a ped? I really need to say nothing more than that really. Sure Volvo say they'll take the blame but until it happens there is no answer to this question.

 

Finally the human factor, imagine the trauma of watching helpless as your car kills a pedestrian with you sat inside it as it happens.

 

You mentioned Aeroplanes earlier, I'm sure your aware (as you seem like an intelligent dude/ette) that its 100% possible for a plane to take off, fly to its destination and then land when it arrives by autopilot however they still make humans take off, do any turns and other manoeuvrers and land, even when the computer could do it better its still mandatory for a human to do these things. Why is that? Because if something goes wrong the computer does the wrong thing and everybody dies. I once saw a simulation of a Boeing 777 where they demonstrated the Autopilot, they then showed what happens when the most advanced auto pilot system in the world encounters a mechanical error. At first it compensated and did the right thing however it wasn't able to throttle control the engines in the way needed to make the plane fly straight again, instead it did the wrong thing entirely and stalled the engines.  They repeated the test again with the human at the helm and by using the yoke and the throttle he was able to regain control of the plane and get it back to the airport. Now I'm aware this agrees with your pilot training statement however it also demonstrates perfectly my point too, a human can adapt on the fly to anything it encounters, it might not be correct but its at least something, if a computer does the same it will almost always do the wrong thing.

 

I'd like to go on record now and say FULLY AUTONOMOUS VEHICLES, defined by a vehicle which requires no user interaction outside of the destination and cannot be overridden by the driver, will never happen. Yeah I'm sure all cars will eventually be able to do self driving in a way which will be great for us all but the driver will ALWAYS be able to take control if needed.

 

Sorry for the wall

Humans also cause all of the other crashes so in the REAL overall CONTEXT the moral and situation qualms are not only unfounded but completely wrong.

 

The equivalent to your argument is refusing to kill a baby to save millions of others. (when without a shadow of a doubt many orders of magnitude more potentially fatal event would be avoided than caused by the removal of HUMAN ERROR.)

 

Also proper situation programming could put those same behavioral traits to solve your selected issue, but since we don't honestly care very much about it in the first place (thinking WRONGLY that it is safer to have a human do it) the programming is woefully behind it's true potential.

 

We can argue about the value of a specific implementation all day long, but there CANNOT be even a SLIVER of doubt that on an absolute scale computational (objective) based systems are safer than human (subjective) based systems.

[Curufinwe_wins]

This is exactly why I believe there needs to be a human element to self driving cars. If a sensor on my car stops working and the failsafe on it fails to stop the car, what do I do? Turning your car off isn't an option at 40mph. If the car slams it's breaks because an animal jumps in front of the car instead of swerving you are dead. If it isn't inputting my destination for some reason, how am I supposed to use it until it can get fixed, you can't. These are valid concerns that can easily be addressed by installing a sort of manual takeover mode like from that scene in iRobot.

 

 

That is exactly how I feel. I'm a gearhead, I love cars, I love watching them on the track and car shows in parking lots. I love the sound of a hard acceleration when getting on a highway or making those tight turns on swervy roads. I'm not selfish for saying that I don't want to watch my hobby be taken away because car manufacturers don't want to put an option for a manual mode.

 

If you think that this will keep autonomous vehicles out of the hands of the masses, then your blind to how technological innovations work and who they cater to. The minority doesn't matter, the masses do, and the masses want autonomous cars @Matu20

 

 

And what would happen if you had a brain hemorrhage while driving? There are massive amounts of human incapacitation mechanisms with higher failure rates than a computerized system (which furthermore could potentially isolate the failed car entirely preventing it from causing a massive escalation like what happens in pileup's after fender-benders.)

 

The most dangerous time is while we still have a few humans on the road with mostly computers. Adding any humans into the mix (as primary control) literally only makes the system more prone to failure in the first place. I'm fine for emergency manual control, but giving people the OPTION of controlling manually or using a computer takes away almost all of the advantages of computerized systems in the first place.

[mattebad]

The thing is that as far as I can see, the people who believe in self driving cars is the minority.

Really? I have the opposite view to that, I was under the impression that the majority wanted autonomous cars while the loud minority don't. This is just as comparable as the advent of electric cars, at first the majority didn't want them. Then they became knowledgable of how much cheaper they were in the long run and that they were more practical than gasoline cars. Now we have every major manufacturer pushing either hybrids or fully electric cars.

 

The same will happen with autonomous vehicles. First we have semi-autonomous vehicles already, the likes coming from Mercedes, Lexus, Acura, BMW. Soon after, I predict around 2-3 years, we will start seeing autonomous cars being pushed into the main market. The first autonomous vehicle was actually approved for commercial use in California recently I believe. Manufacturers won't push a technology that they don't believe the masses would be interested in, it's marketing 101

[Trik'Stari]

By the very fact of being machines who do the same thing every time, always do things on purpose as you put it, they will get into dramatically fewer of those moral dilemma's in the first place. Furthermore the entire system can be reworked to take advantage of machine's tighter tolerances.

 

Example: When traffic gets heavier people slow down. This is the WRONG ACTION. Slowing down lowers the volume of cars that can get through the section which further increases traffic. We create a self-reinforcing issue (which is why traffic gets so bad.) The proper action would be to maintain speed or actually INCREASE speed, to allow a greater volume of traffic to flow. Why don't we do this? Well first, we are particularily bad at making the right decision under stress, and more than making the right decision, we are inconsistent, and indeed not knowing what the people around you driving are doing/going to do is one of the biggest reasons people slow down in the first place (it gives them more time to react to a changing situation). A network of driver-less vehicles could communicate all proposed behaviors/actions (or even just use a consistent behavior set) and remove that dilemma.

 

Another good example of human's consistently being WRONG is decreasing lanes or passing obsticles. You know those times when you are going 65 MPH (for example) in a three lane, as you get to the drop to two lanes your speed decreases to next to nothing, then once you pass it the speed goes back up to basically where it was n the first place? Those are caused by human errors piling up (or human actions that detrimentally affect the flow but were made on purpose aka selfishness). A network of driverless computers (again primarily due to a known expected behavior set) could convert that merger miles in advance of the event (indeed over a course of miles if needed as well, not waiting till the last second) completely eliminating the flow issue.

 

Car traffic is like a flowing fluid, we are clotting mechanisms (most of the time as individuals we are fine, but mistakes happen and clots form or constricted pathways lead to clot buildups), machines don't have that issue. They don't make "mistakes".

Yeah....no. Machines do make mistakes. As someone pointed out earlier, the reason we still have human pilots instead of just autopilot, is because no machine is capable of handling every possible situation. How is the driverless car going to react when a deer runs into the side of the car? (not the car hitting the deer, but the deer hitting the car. It happens all the time). How is the car going to react in the middle of the city, when a pedestrian steps out in front of the car, and there's no way to swerve because there's another car coming in the opposite direction in the next lane? Granted that last one is still the result of human error, a human not looking before the step out into the road, but it's still going to happen, and what is the car going to do?

[Suika]

It's called driving defensively. You leave plenty of room between yourself and the other driver so you have time to react. You also leave for where you need to be early, so you can avoid rush hour traffic.

Very difficult thing to do with heavy traffic, the innocent driver who was hit could have done nothing to prepare for or had even seen a car hitting him from the front at 60 miles per hour.

[Curufinwe_wins]

Really? I have the opposite view to that, I was under the impression that the majority wanted autonomous cars while the loud minority don't. This is just as comparable as the advent of electric cars, at first the majority didn't want them. Then they became knowledgable of how much cheaper they were in the long run and that they were more practical than gasoline cars. Now we have every major manufacturer pushing either hybrids or fully electric cars.

 

The same will happen with autonomous vehicles. First we have semi-autonomous vehicles already, the likes coming from Mercedes, Lexus, Acura, BMW. Soon after, I predict around 2-3 years, we will start seeing autonomous cars being pushed into the main market. The first autonomous vehicle was actually approved for commercial use in California recently I believe. Manufacturers won't push a technology that they don't believe the masses would be interested in, it's marketing 101

Technically speaking with our current energy portfolio they are neither cheaper nor more practical than gasoline cars, but they do have the potential to be better based again on energy generation portfolio's.

 

Another option however that may end up beating both is non-electrochemical carbon neutral fuel sources (such as hydrogen and/or hydrocarbon generation from waste heat).

[Trik'Stari]

Very difficult thing to do with heavy traffic, the innocent driver who was hit could have done nothing to prepare for or had even seen a car hitting him from the front at 60 miles per hour.

Well that's just a problem of road building. Why was there not a divider of some sort?

[mattebad]

And what would happen if you had a brain hemorrhage while driving? There are massive amounts of human incapacitation mechanisms with higher failure rates than a computerized system (which furthermore could potentially isolate the failed car entirely preventing it from causing a massive escalation like what happens in pileup's after fender-benders.)

 

The most dangerous time is while we still have a few humans on the road with mostly computers. Adding any humans into the mix (as primary control) literally only makes the system more prone to failure in the first place. I'm fine for emergency manual control, but giving people the OPTION of controlling manually or using a computer takes away almost all of the advantages of computerized systems in the first place.

I don't think you read my comment thoroughly enough, I don't believe humans should be the primary controller. We are prone to mistakes and ignorance. I said word for word, "These are valid concerns that can easily be addressed by installing a sort of manual takeover mode like from that scene in iRobot." 

 

Now if you haven't seen the movie, there is a scene where Will Smith is surrounded by two trucks in front and behind him, travelling with him at the same speed so he couldn't escape. He enabled manual control by pressing a button and having a steering wheel pop out and allowing him to control the vehicle instead of the computer. This is what I mean, there should be an optional takeover mode for people in very specific situations the computer wasn't programmed for (a certain slick on the road, deer jumping in front of the car, debris falling off a truck on the highway, hurricane like weather, etc.) or for the small minority of people that want to take their car to a race track to have some fun.

[mattebad]

Technically speaking with our current energy portfolio they are neither cheaper nor more practical than gasoline cars, but they do have the potential to be better based again on energy generation portfolio's.

 

Another option however that may end up beating both is non-electrochemical carbon neutral fuel sources (such as hydrogen and/or hydrocarbon generation from waste heat).

Here's a good source that shows the difference in long term savings between electric cars and gasoline cars. I don't agree with including the battery replacement in there due to the fact that A. It is covered by insurance and B. The batteries are rated to last 6 years before they start to lose their efficiency, and even then they don't really start to degrade physically until around the 8-9 year mark.

 

Hydrogen or Hydrogen Fuel Cells are not a viable option due to our lack of hydrogen production that is available to us and how long it takes for hydrogen to replenish itself naturally. I've never heard of Hydrocarbon generation from waste heat for cars. Can you explain that to me? I heard about it being used in a few factories to regain lost energy from their emissions, but that's about it, nothing on a small scale level.

[Master Disaster]

Humans also cause all of the other crashes so in the REAL overall CONTEXT the moral and situation qualms are not only unfounded but completely wrong.

 

The equivalent to your argument is refusing to kill a baby to save millions of others. (when without a shadow of a doubt many orders of magnitude more potentially fatal event would be avoided than caused by the removal of HUMAN ERROR.)

 

Also proper situation programming could put those same behavioral traits to solve your selected issue, but since we don't honestly care very much about it in the first place (thinking WRONGLY that it is safer to have a human do it) the programming is woefully behind it's true potential.

 

We can argue about the value of a specific implementation all day long, but there CANNOT be even a SLIVER of doubt that on an absolute scale computational (objective) based systems are safer than human (subjective) based systems.

 

I cannot argue with you because you're right. Computers are massively better at us in almost every aspect of piloting any vehicle you can name, doesn't change the fact that without SPECIFIC PROGRAMMING it cannot deal with a situation where a human can.

 

As I said, fully autonomous vehicles will never happen, humans will always be able to override, that said though I fully expect self driving vehicles to almost entirely eliminate road accidents.

 

Don't get me wrong, I'm not against them at all, the problem right now is the legal mess and the moral questions they pose. If these things can be worked out then they will change the world. But eventually it will also happen that somewhere a SDC will kill a pedestrian, and when that happens the shit will hit the fan big time.

[Curufinwe_wins]

Very difficult thing to do with heavy traffic, the innocent driver who was hit could have done nothing to prepare for or had even seen a car hitting him from the front at 60 miles per hour.

Not only is it difficult, but it is the WRONG ACTION to take. By slowing down and allowing more space, you perpetuate the issue in the first place. We do it only because we are dealing with other humans who also make mistakes and the margin for error decreases in those situations.

 

Comparison. Take a pipe with flowing water, decrease the diameter while keeping the pressure constant, the linear flow rate increases to preserve the system balance. Humans decide to decrease the linear flow rate making the pressure larger, further decreasing the linear flow rate, until either the volumetric flow rate drops or the pipe bursts.

 

@Trik'Stari except their actions are not mistakes, and knowing before hand how the individual performer is going to act in a situation (even if it is detrimental to the individual) is massively helpful in mitigating the spread of that action. Example. A person swerves to avoid an animal in real time in traffic. Other people then have to act in sync with each other to prevent that relatively small disruption from self-promoting (which most of the time they don't, because they can't know how other people are going to react.) In a diver-less situation (supposing the exact same initial action is taken), the entire system around it knows exactly how each other will react, so coordination to dissipate the disruption can be taken (like seamlessly merging in traffic) not to mention the overall latency of the system decreases dramatically (because human reaction time is pretty bad in comparison to computational response time).

Sure there might be situations where an INDIVIDUAL computer may fail to act in a manner with it's own personal outcome being optimized (not a mistake per-say but not the most optimal action according to our own logic parameters), but as a system not only are errors nearly eliminated (the only ones being present are those HUMANS failed to address in the first place), but the ability to dissipate local errors without causing systematic failure is incomparably improved (again mainly because behavior is systematic not erratic, and to a lesser extent because overall system knowledge can be massively increased with fully computational flows)