Yes, because they’re completely different concepts.
What is fundamentally different between a puddle and a hole created by a construction worker or a sinkhole? My tires already tell me if they're low on air pressure. My car flashes a signal which indicates the roads are icy, the 4WD is kicking in and I should reduce my speed. I'm not sure why anyone thinks a computer can't recognize driving conditions that they already are.
Catalog of existence, refractive index, measurement of threat to vehicle, maximum possible size, potential of location…
Virtually everything, really.
LOL. I mean to the point of argument of those saying, "driverless cars can never happen, because... puddles and random tornados." Computer technology already deals with most of these factors quite well. Hard to believe – and no one has explained – why it's realistic that a driverless car could avoid a pothole but not a large puddle.
the 4th video you posted, demonstrates that Google cars already can anticipate construction detours, what looks like an intoxicated cyclist, necessary lane changes etc. So it can see construction workers and drunk bicyclists and other obstacles in the road, but not giant puddles or tornados. Oooookkkaaayyy.
How can it see a tornado? It could be a mile away. The car was looking for traffic signs on the construction site. A human driver could see weather conditions and know not to drive that way. Same with a mudslide, huge wave or even a deer/rabbit running from the side of the road. The automated car will focus on the events on the road not knowing about nor how to interpret off-road events.
The depth of puddles would be very hard to determine by camera as they are flat. If it could detect a puddle, taking the correct course of action is still a problem. If it notices it just as the car is about to hit at 30mph, does it slam on the brakes, does it steer into potentially oncoming traffic, does it drive right through, possibly putting the car out of control, possibly splashing water over pedestrians?
Humans can see other cars or people hitting ice in the road ahead and know to anticipate it. A car has no way of detecting ice on a road. It can detect certain conditions in general from temperature, traction, humidity etc but it can't identify individual dangerous scenarios accurately.
LOL. I mean to the point of argument of those saying, "driverless cars can never happen, because... puddles and random tornados." Computer technology already deals with most of these factors quite well.
I’ve not seen one that can.
Hard to believe – and no one has explained – why it's realistic that a driverless car could avoid a pothole but not a large puddle.
How can it see a tornado? It could be a mile away. The car was looking for traffic signs on the construction site. A human driver could see weather conditions and know not to drive that way. Same with a mudslide, huge wave or even a deer/rabbit running from the side of the road. The automated car will focus on the events on the road not knowing about nor how to interpret off-road events.
The depth of puddles would be very hard to determine by camera as they are flat. If it could detect a puddle, taking the correct course of action is still a problem. If it notices it just as the car is about to hit at 30mph, does it slam on the brakes, does it steer into potentially oncoming traffic, does it drive right through, possibly putting the car out of control, possibly splashing water over pedestrians?
Humans can see other cars or people hitting ice in the road ahead and know to anticipate it. A car has no way of detecting ice on a road.
Naive presumptions IMO. You're asking us to believe that a computer can interpret an erratic cyclist but not mud coming down a hill.
Tornados: So weather reports are published with Tornado warnings. Bob foolishly decides it would be "fun" to attempt driving anyway, or perhaps prudently believes he can escape the tornado by moving himself to a different location. How is Bob holding the steering wheel an advantage over a driverless car with sensors, or a driverless train for that matter? (Driverless trains do exist already).
Sonar to detect depth through water has existed since WW1. I have no doubt the engineers are aware of that issue and are working on it.
If a car can already detect an erratic cyclist or pedestrian, it can most certainly detect erratic cars caused by ice or any other factor. As I said, on icy roads I get a warning light flashing on my dashboard and the 4WD kicks in.
As for "what to do" when an unexpected obstacle arises, a computer's reaction time will be faster than mine. They are already averting construction zones, pedestrians and cyclists. Just like the human driving a car, the computer needs to assess and make a decision. Cars are already doing a lot of this, so it's ludicrous to assume they "cannot" do more of it as the tech progresses.
EDIT: I think it's reasonable to assume that before driverless cars become legal, there will be locks in place that prevent them from leaving port in cases of mudslides, avalanches, sinkholes or tornado warnings, the same way human drivers SHOULD do but quite often don't. The same way convicted drunk drivers have to get an Interlock installed on their cars, you can prevent cars from driving into mudslides easier via technology than you can hope to rely on the common sense of some humans.
Before I go on a road drip through the Rocky Mountains where I live in Western Canada, I check my iPhone for road conditions and whether reports. There are sensors on our highways here that include measurements of snow, ice, temperature, accidents and so on, which a prudent driver checks before getting in the car. There is no reason this data can't be funnelled directly into the onboard computer.
SIRI: "ICY ROADS AHEAD. I'm sorry, but we'll need to pull over and have a rest until the situation clears. Would you like me to make a reservation at the nearest restaurant?" Or in some cases you can override it and take the steering wheel. It's going to happen, driverless cars are inevitable, not one of the "obstacles" raised by Luddites seems permanently insurmountable.
TS – Umm, I wasn't addressing the ethical/should component about this, only the technical obstacles that some seem to think are insurmountable. I'm not clear whether you think these issues with random puddles and tornadoes coming down the centre line movie-style are permanent issues, or merely present issues?
As for the ethics of it, I think it's going to be great once the bugs are worked out. I have not heard any arguments for why this would be "bad" or "wrong". **** Google, I don't want them tracking my activities any further than they already do, so that they can pump ads in my face. But I would trust an Apple driverless car or a Tesla driverless car.
I see the technology unfolding a little differently than some seem to imagine. Basically modified HOV lanes, "tracks" without the physical tracks. I remember as a kid in the 1970s riding the monorail with my dad at Disney World and peppering him with questions about how that was safe, and all kinds of "what if" scenarios. I couldn't get over the fact that there was no human at the controls. I found it a little frightening until I rode it about 35 times and stopped being afraid. My dad essentially told me that the engineers had thought of everything, and the key is they keep crap (obstacles) off the track and out of the way, and if something does get on to the track, the train "knows it" and it stops until someone can come out and fix the issue, maybe arrest someone like a troublemaker who might try to throw debris on to the track.
We will still need a "track" for a very long time, but with cameras and sensors it is – to me at least – easily conceivable that cars can be driven on to a special lane and set to arrive at a destination, safe enough that a person could go to sleep, watch television or do work on a computer. Trains have been transporting humans for centuries, and while they usually have a human engineer at the helm, they don't always, especially in the last few decades. And train accidents happen with humans at the control, just like car accidents do. So while safety is the paramount concern, I am not convinced that humans are permanently better than computers at this task, especially when you can provide a special lane (a virtual "track"), equipped with sensors and all kinds of wonderful technology.
You're asking us to believe that a computer can interpret an erratic cyclist but not mud coming down a hill.
It would also need to know if the mud posed a threat of collision. The bigger issue is getting all the data rather than interpreting it. If the sensors in the vehicle are designed to track things within a 1/4 mile radius then it would have some chance at detecting upcoming problems. Picking out important parts from that amount of data would be very difficult. Even at their primitive stage just now, they are processing 1GB of data per second.
Sonar to detect depth through water has existed since WW1.
That's completely different sampling. That's underwater sending waves directly at surfaces and getting waves bounced back. As you drive along, a puddle is at a very shallow angle. It would need to be able to sample both the top and bottom of the puddle using two different signals but these signals would bounce off the surface of the water.
If a car can already detect an erratic cyclist or pedestrian, it can most certainly detect erratic cars caused by ice or any other factor. As I said, on icy roads I get a warning light flashing on my dashboard and the 4WD kicks in.
Your car can detect loss in traction but it can't see a sheet of ice before speeding into it. It needs to see the ice well before hitting it.
The link I posted above says that "Google's autonomous cars can't navigate 99% of the US, an insider said... Cars aren't driven in the rain or snow as detection technology can't distinguish between droplets or flakes and obstacles on the road... Google’s self-driving cars can’t currently cope in heavy rain or snow... the current prototype cars are very reliant on maps to navigate and can’t react like a human driver, dodging potholes and other hazards."
If they have an override to allow for worst case scenarios (which limits their design) and make enough advances to be able to deal with the majority of obstacles, maybe enforce automated cars inside cities so that all vehicles are automated then maybe it can be practical but until these issues are overcome, it's not practical. If they never overcome these issues then it's not safe enough.
What do you see as the obstacle to a driverless car being able to detect a sheet of ice? As it stands, human beings aren't so great at that today.
Would you be comfortable with driverless cars in above-freezing temperatures and/or dry weather? My belief is this is incremental. Banning driverless cars in icy conditions is not the same thing as saying "driverless cars can't work/aren't safe".
I think the tornado thing is a red herring. People die in tornados. Just because a computer may not save a person doesn't mean a human could do any better. I think a driverless car could ascertain the next move just as well as a human could, the same way it can break, change lanes, stop and accelerate when an erratic cyclist swerves over. In fact I would argue that you've got (or your computer has) a little more time to assess a tornado situation than it does to assess what to do when a cyclist swerves in front.
Self driving cars would be great for productivity. Spending 1-3 hours navigating roads every day is a huge loss of time. Public transport doesn't solve this problem because people can see your screens and you can't do anything really important, or make calls without pissing everyone off. If you're in the privacy and comfort of your own car, things turn different very very quickly.
What do you see as the obstacle to a driverless car being able to detect a sheet of ice? As it stands, human beings aren't so great at that today.
It's mainly the sensors that would need to be improved. Ideally they'd be able to detect any kind of material on the road. There are scanners that can measure what materials things are made from, I don't know if they'd be able to work at a distance:
Some signals won't be able to bounce back so they'd have to rely on measuring incoming light. These scanners have been used on movies like Iron Man to match the physical suit to the digital one:
With a special scanner, they'd be able to detect all forms of rain, snow and ice and so wouldn't mistake raindrops or snowdrops as obstacles. Depth scanners would help with that too. It still needs to see through the weather though. If signals are bounced off heavy rainfall, snow or fog, the car is essentially blind to everything.
Would you be comfortable with driverless cars in above-freezing temperatures and/or dry weather? My belief is this is incremental. Banning driverless cars in icy conditions is not the same thing as saying "driverless cars can't work/aren't safe".
I think it can be an incremental process. Google is trying to teach the cars how to react to certain events based on human input. Weather can change quickly though. You don't want to be driving in traffic or at high speed and a sudden heavy downpour appears and the car is unable to detect anything more than 10cm in front of the sensors.
I think the tornado thing is a red herring. People die in tornados. Just because a computer may not save a person doesn't mean a human could do any better.
The scenario that I'm describing is one where we have a driverless car without manual override and the car doesn't have sensors to detect a far away tornado. It then continues driving closer to it not aware of the danger. If it has the ability to detect this and track its movement then it could do a better job than a human but all of the decisions a machine is making come from a human programmer. It's not a human vs machine, it's a human with immediate information making specific decisions vs a human programmer making general decisions about what to do in general circumstances. The human in danger will make a difference set of rationalizations for their own survival than a programmer would e.g there's a dead person blocking the vehicle, driver says run over them to get away, computer says no.
This was the problem with the 3 laws of robotics, the following is a scene in the movie iRobot where a robot is saving 2 humans involved in a car crash that are underwater, one is an adult, the other a child trapped in the car:
"NS-4 Robot: You are in danger!
Detective Del Spooner: Save her! Save the girl!
Detective Del Spooner: But it didn't. It saved me.
Dr. Susan Calvin: A robot's brain is a difference engine, it must have calculated...
Detective Del Spooner: It did. I was the *logical* choice. It calculated I had a forty-five percent chance of survival. Sarah only had an eleven percent chance. That was somebody's baby. Eleven percent is more than enough. A human being would have known that. But robots, nothing here. (points at heart). They're just lights, and clockwork. But you go ahead and trust them if you wanna."
If a child runs out into the road and the car has only two choices: drive into the child, killing them or swerve off the road potentially killing a family of 5 passengers and pedestrians, what's the right decision? The computer will decide that killing the person walking onto the road is a high probability while killing the family and others is unknown. The driver can make a better judgement.
Nobody wants to engage very much with the policeman waving me down in London. Or the legal challenges.
So let's throw another spanner in the works.
Why do some cars cost so much? The answer isn't just looks, brand, fit and finish. It's horsepower. 0-60 in 2.x seconds.
To buy those those cars is to buy something which would be illegal to run at 40% or more of its maximum speed. People who pay the big bucks like driving.
A safe pod with no steering wheel. That's a small bus.
Nobody wants to engage very much with the policeman waving me down in London. Or the legal challenges.
So let's throw another spanner in the works.
Why do some cars cost so much? The answer isn't just looks, brand, fit and finish. It's horsepower. 0-60 in 2.x seconds.
To buy those those cars is to buy something which would be illegal to run at 40% or more of its maximum speed. People who pay the big bucks like driving.
A safe pod with no steering wheel. That's a small bus.
I agree.
And your policeman is but the tip of the iceberg. There are so many showstopping problems to driverless cars. Like: in a city, you simply couldn't have pedestrians anywhere near them, or you'd have gridlock. The cars would, of course, give way to pedestrians, and this would result in cars going nowhere fast. You'd have to ban people from roads.
If Microsoft made cars you'd have Autocar Basic, which just gives you the standard move, turn, brake system. Autocar Premium would add the pedestrian and accident avoidance system. Autocar Enterprise would allow optional pedestrian collision depending on how late you were to work.
That journalist isn't going to be trusting driverless cars anytime soon. There will be different classes of autonomy like automated parking so they'll have to make it clear which vehicles have the features to avoid people. I can see this coming with regulation stickers on cars so that people will know the car will try and avoid them.
Comments
Yes, because they’re completely different concepts.
What is fundamentally different between a puddle and a hole created by a construction worker or a sinkhole? My tires already tell me if they're low on air pressure. My car flashes a signal which indicates the roads are icy, the 4WD is kicking in and I should reduce my speed. I'm not sure why anyone thinks a computer can't recognize driving conditions that they already are.
What is fundamentally different between a puddle and a hole created by a construction worker or a sinkhole?
Catalog of existence, refractive index, measurement of threat to vehicle, maximum possible size, potential of location…
Virtually everything, really.
Catalog of existence, refractive index, measurement of threat to vehicle, maximum possible size, potential of location…
Virtually everything, really.
LOL. I mean to the point of argument of those saying, "driverless cars can never happen, because... puddles and random tornados." Computer technology already deals with most of these factors quite well. Hard to believe – and no one has explained – why it's realistic that a driverless car could avoid a pothole but not a large puddle.
How can it see a tornado? It could be a mile away. The car was looking for traffic signs on the construction site. A human driver could see weather conditions and know not to drive that way. Same with a mudslide, huge wave or even a deer/rabbit running from the side of the road. The automated car will focus on the events on the road not knowing about nor how to interpret off-road events.
The depth of puddles would be very hard to determine by camera as they are flat. If it could detect a puddle, taking the correct course of action is still a problem. If it notices it just as the car is about to hit at 30mph, does it slam on the brakes, does it steer into potentially oncoming traffic, does it drive right through, possibly putting the car out of control, possibly splashing water over pedestrians?
Humans can see other cars or people hitting ice in the road ahead and know to anticipate it. A car has no way of detecting ice on a road. It can detect certain conditions in general from temperature, traction, humidity etc but it can't identify individual dangerous scenarios accurately.
http://www.dailymail.co.uk/sciencetech/article-2739765/Is-Google-s-driverless-car-fair-weather-friend-Vehicle-WON-T-work-heavy-rain-snow-insider-reveals.html
I’ve not seen one that can.
Just did…
How can it see a tornado? It could be a mile away. The car was looking for traffic signs on the construction site. A human driver could see weather conditions and know not to drive that way. Same with a mudslide, huge wave or even a deer/rabbit running from the side of the road. The automated car will focus on the events on the road not knowing about nor how to interpret off-road events.
Humans can see other cars or people hitting ice in the road ahead and know to anticipate it. A car has no way of detecting ice on a road.
Naive presumptions IMO. You're asking us to believe that a computer can interpret an erratic cyclist but not mud coming down a hill.
Tornados: So weather reports are published with Tornado warnings. Bob foolishly decides it would be "fun" to attempt driving anyway, or perhaps prudently believes he can escape the tornado by moving himself to a different location. How is Bob holding the steering wheel an advantage over a driverless car with sensors, or a driverless train for that matter? (Driverless trains do exist already).
Sonar to detect depth through water has existed since WW1. I have no doubt the engineers are aware of that issue and are working on it.
If a car can already detect an erratic cyclist or pedestrian, it can most certainly detect erratic cars caused by ice or any other factor. As I said, on icy roads I get a warning light flashing on my dashboard and the 4WD kicks in.
As for "what to do" when an unexpected obstacle arises, a computer's reaction time will be faster than mine. They are already averting construction zones, pedestrians and cyclists. Just like the human driving a car, the computer needs to assess and make a decision. Cars are already doing a lot of this, so it's ludicrous to assume they "cannot" do more of it as the tech progresses.
EDIT: I think it's reasonable to assume that before driverless cars become legal, there will be locks in place that prevent them from leaving port in cases of mudslides, avalanches, sinkholes or tornado warnings, the same way human drivers SHOULD do but quite often don't. The same way convicted drunk drivers have to get an Interlock installed on their cars, you can prevent cars from driving into mudslides easier via technology than you can hope to rely on the common sense of some humans.
I’ve not seen one that can.
Just did…
Before I go on a road drip through the Rocky Mountains where I live in Western Canada, I check my iPhone for road conditions and whether reports. There are sensors on our highways here that include measurements of snow, ice, temperature, accidents and so on, which a prudent driver checks before getting in the car. There is no reason this data can't be funnelled directly into the onboard computer.
SIRI: "ICY ROADS AHEAD. I'm sorry, but we'll need to pull over and have a rest until the situation clears. Would you like me to make a reservation at the nearest restaurant?" Or in some cases you can override it and take the steering wheel. It's going to happen, driverless cars are inevitable, not one of the "obstacles" raised by Luddites seems permanently insurmountable.
I’m sorry…
Canadian Siri sure is different…
Some would have you not be able to do that.
Yeah, that’s an instant “disregard” in my book. “Going to” is never as meaningful as “should”.
Again, you don’t seem to know who the luddites are here.
TS – Umm, I wasn't addressing the ethical/should component about this, only the technical obstacles that some seem to think are insurmountable. I'm not clear whether you think these issues with random puddles and tornadoes coming down the centre line movie-style are permanent issues, or merely present issues?
As for the ethics of it, I think it's going to be great once the bugs are worked out. I have not heard any arguments for why this would be "bad" or "wrong". **** Google, I don't want them tracking my activities any further than they already do, so that they can pump ads in my face. But I would trust an Apple driverless car or a Tesla driverless car.
I see the technology unfolding a little differently than some seem to imagine. Basically modified HOV lanes, "tracks" without the physical tracks. I remember as a kid in the 1970s riding the monorail with my dad at Disney World and peppering him with questions about how that was safe, and all kinds of "what if" scenarios. I couldn't get over the fact that there was no human at the controls. I found it a little frightening until I rode it about 35 times and stopped being afraid. My dad essentially told me that the engineers had thought of everything, and the key is they keep crap (obstacles) off the track and out of the way, and if something does get on to the track, the train "knows it" and it stops until someone can come out and fix the issue, maybe arrest someone like a troublemaker who might try to throw debris on to the track.
We will still need a "track" for a very long time, but with cameras and sensors it is – to me at least – easily conceivable that cars can be driven on to a special lane and set to arrive at a destination, safe enough that a person could go to sleep, watch television or do work on a computer. Trains have been transporting humans for centuries, and while they usually have a human engineer at the helm, they don't always, especially in the last few decades. And train accidents happen with humans at the control, just like car accidents do. So while safety is the paramount concern, I am not convinced that humans are permanently better than computers at this task, especially when you can provide a special lane (a virtual "track"), equipped with sensors and all kinds of wonderful technology.
It would also need to know if the mud posed a threat of collision. The bigger issue is getting all the data rather than interpreting it. If the sensors in the vehicle are designed to track things within a 1/4 mile radius then it would have some chance at detecting upcoming problems. Picking out important parts from that amount of data would be very difficult. Even at their primitive stage just now, they are processing 1GB of data per second.
Not everyone listens to weather reports and a tornado warning is not the same as a tornado. A person can see the tornado (3:40):
[VIDEO]
A car might be able to detect that threat on film but a human would know best how to avoid it and it may even require driving off-road.
That's completely different sampling. That's underwater sending waves directly at surfaces and getting waves bounced back. As you drive along, a puddle is at a very shallow angle. It would need to be able to sample both the top and bottom of the puddle using two different signals but these signals would bounce off the surface of the water.
Your car can detect loss in traction but it can't see a sheet of ice before speeding into it. It needs to see the ice well before hitting it.
The link I posted above says that "Google's autonomous cars can't navigate 99% of the US, an insider said... Cars aren't driven in the rain or snow as detection technology can't distinguish between droplets or flakes and obstacles on the road... Google’s self-driving cars can’t currently cope in heavy rain or snow... the current prototype cars are very reliant on maps to navigate and can’t react like a human driver, dodging potholes and other hazards."
http://www.technologyreview.com/news/530276/hidden-obstacles-for-googles-self-driving-cars/
If they have an override to allow for worst case scenarios (which limits their design) and make enough advances to be able to deal with the majority of obstacles, maybe enforce automated cars inside cities so that all vehicles are automated then maybe it can be practical but until these issues are overcome, it's not practical. If they never overcome these issues then it's not safe enough.
What do you see as the obstacle to a driverless car being able to detect a sheet of ice? As it stands, human beings aren't so great at that today.
Would you be comfortable with driverless cars in above-freezing temperatures and/or dry weather? My belief is this is incremental. Banning driverless cars in icy conditions is not the same thing as saying "driverless cars can't work/aren't safe".
I think the tornado thing is a red herring. People die in tornados. Just because a computer may not save a person doesn't mean a human could do any better. I think a driverless car could ascertain the next move just as well as a human could, the same way it can break, change lanes, stop and accelerate when an erratic cyclist swerves over. In fact I would argue that you've got (or your computer has) a little more time to assess a tornado situation than it does to assess what to do when a cyclist swerves in front.
Even if Apple isn't developing an (electric) (autonomous) car, I'm sure that there is research underway. MacObserver however, is convinced:
http://www.macobserver.com/tmo/article/apple-is-working-on-a-car?utm_content=buffer4684c&utm_medium=social&utm_source=app.net&utm_campaign=buffer
May we all continue to live in interesting times lol.
It's mainly the sensors that would need to be improved. Ideally they'd be able to detect any kind of material on the road. There are scanners that can measure what materials things are made from, I don't know if they'd be able to work at a distance:
http://www.optis-world.com/products/hardware/OMS4.html
Some signals won't be able to bounce back so they'd have to rely on measuring incoming light. These scanners have been used on movies like Iron Man to match the physical suit to the digital one:
http://www.fxguide.com/featured/ben-snow-the-evolution-of-ilm-lighting-tools/
With a special scanner, they'd be able to detect all forms of rain, snow and ice and so wouldn't mistake raindrops or snowdrops as obstacles. Depth scanners would help with that too. It still needs to see through the weather though. If signals are bounced off heavy rainfall, snow or fog, the car is essentially blind to everything.
I think it can be an incremental process. Google is trying to teach the cars how to react to certain events based on human input. Weather can change quickly though. You don't want to be driving in traffic or at high speed and a sudden heavy downpour appears and the car is unable to detect anything more than 10cm in front of the sensors.
The scenario that I'm describing is one where we have a driverless car without manual override and the car doesn't have sensors to detect a far away tornado. It then continues driving closer to it not aware of the danger. If it has the ability to detect this and track its movement then it could do a better job than a human but all of the decisions a machine is making come from a human programmer. It's not a human vs machine, it's a human with immediate information making specific decisions vs a human programmer making general decisions about what to do in general circumstances. The human in danger will make a difference set of rationalizations for their own survival than a programmer would e.g there's a dead person blocking the vehicle, driver says run over them to get away, computer says no.
This was the problem with the 3 laws of robotics, the following is a scene in the movie iRobot where a robot is saving 2 humans involved in a car crash that are underwater, one is an adult, the other a child trapped in the car:
"NS-4 Robot: You are in danger!
Detective Del Spooner: Save her! Save the girl!
Detective Del Spooner: But it didn't. It saved me.
Dr. Susan Calvin: A robot's brain is a difference engine, it must have calculated...
Detective Del Spooner: It did. I was the *logical* choice. It calculated I had a forty-five percent chance of survival. Sarah only had an eleven percent chance. That was somebody's baby. Eleven percent is more than enough. A human being would have known that. But robots, nothing here. (points at heart). They're just lights, and clockwork. But you go ahead and trust them if you wanna."
If a child runs out into the road and the car has only two choices: drive into the child, killing them or swerve off the road potentially killing a family of 5 passengers and pedestrians, what's the right decision? The computer will decide that killing the person walking onto the road is a high probability while killing the family and others is unknown. The driver can make a better judgement.
So let's throw another spanner in the works.
Why do some cars cost so much? The answer isn't just looks, brand, fit and finish. It's horsepower. 0-60 in 2.x seconds.
To buy those those cars is to buy something which would be illegal to run at 40% or more of its maximum speed. People who pay the big bucks like driving.
A safe pod with no steering wheel. That's a small bus.
I agree.
And your policeman is but the tip of the iceberg. There are so many showstopping problems to driverless cars. Like: in a city, you simply couldn't have pedestrians anywhere near them, or you'd have gridlock. The cars would, of course, give way to pedestrians, and this would result in cars going nowhere fast. You'd have to ban people from roads.
[VIDEO]
You have to make sure and get all the upgrades like the 'pedestrian detection' system:
http://www.independent.co.uk/life-style/gadgets-and-tech/news/selfparking-volvo-plows-into-journalists-after-owner-neglects-to-pay-for-extra-feature-that-stops-cars-crashing-into-people-10277203.html
If Microsoft made cars you'd have Autocar Basic, which just gives you the standard move, turn, brake system. Autocar Premium would add the pedestrian and accident avoidance system. Autocar Enterprise would allow optional pedestrian collision depending on how late you were to work.
That journalist isn't going to be trusting driverless cars anytime soon. There will be different classes of autonomy like automated parking so they'll have to make it clear which vehicles have the features to avoid people. I can see this coming with regulation stickers on cars so that people will know the car will try and avoid them.
[VIDEO]