Cannot wait to buy one!
I’d only be OK with that if everyone was on auto-pilot. I dislike the idea of having to rely on the automated system to cope with sudden threats from “drivers” on manual control doing idiotic things, since I live in an area where a large percentage of the “drivers” apparently got their licenses (if any at all) out of Crackerjack boxes. Not fond of the idea of being a helpless target for the stupid, and they’d be the ones who’d insist on manual control.
I’d much rather have the automated system handle sudden unexpected events.
It’s got a faster reaction time. Its attention doesn’t wander. It knows and can sense the actual limits of the vehicle, meaning it will be able to take the best evasive action rather than just slamming on the brakes and/or swerving off the road, as humans often do when presented with an unexpected emergency.
I’m not as confident as you are about the reaction time question, since things can happen that wouldn’t fall into its programming that a human brain has the flexibility to deal with. If it can be proven that an automated system can handle unexpected and unusual events (programmers can’t necessarily think of EVERYTHING, you know), then I’d be a lot more willing to trust it if other cars around me were on manual. Otherwise, put EVERYONE on automatic, but don’t take away the ability of some to exercise human judgement while allowing others to do random weird stuff.
Yeah, but nitwit drivers are hardly unexpected events. A tree falling on you, maybe.
I’ve had two traffic tickets in over 40 years of driving, none in the last 17 years, and I’d so get one. And I hope other people will also. Not for safety, but traffic would move a lot faster with computers in control. No one leaving tons of room and having others weave around them, no cars trying to change four lanes in 500 feet, and I assume that when there is the need to change lanes the car will communicate with other cars and they would slow to make room - so no chain reaction braking which is very disruptive to traffic flow.
I bet when we get them there will be a new business opportunity in opening public speedways to allow those who love to drive themselves do obstacles without annoying/endangering those in automated cars.
Just this morning I was behind someone who was leaving tons of room and weaving almost into the next lane. I thought he was on a phone, but I was wrong. He was shaving. Please, put that clown in an automated car - I’m sure Google can drive better than he can.
I long for the day it is common. As much as I love the open road, I hate commuting. To be able to snooze, or read or talk on phone would be awesome. I wouldn’t even mind if it took longer.
When insurance companies start giving rates lower to drivers that use such systems is when I will opt in for one.
But I think it is a long, long way off.
I like the idea–but only if I can turn off the robot when I want to.Sometimes driving is FUN .
But I don’t see it happening any time soon–there are too many problems, both technical and legal.
Technical problems:
Suppose a contruction worker (not a policeman) flags you down, expecting you to stop in the middle of your lane. Then he tells you to drive in the opposite lane , after checking by walkie-talkie with another construction worker a mile down the road who has blocked the oncoming traffic for you. Will a robot-car understand?
Suppose you are the only robo-car in a funeral procession, and a policeman guides the line of cars to move steadily through a red light, The cop is sitting on his motorcycle 10 yards to your left, blocking traffic for you, but your car sees the red traffic light, and stops.
Suppose there are icy spots on the road, or potholes to steer around?
Legal problems: see Shodan’s post above.
There may be one way to avoid the legal problems: make the driver responsible, not the car. (Like google’s experimental car is today----there is a licensed driver sitting behind the wheel, alert at all times.) But that would defeat the purpose of a robot-car.
I think it will be 25 years or more before we have robo-cars, and even then, they will be limited in their capabilities or usefulness. (maybe restricted to driving only on certain roads)
Whether its ice proficient AI or transmitting markers to safely reroute traffic, they are just the details to make it possible. They have to be assumed to somehow be successful or the premise doesn’t exist to be considered.
Otara
Note that self-driving is likely to be limited to highways at first. So some of these issues will be ironed out in a simpler environment first. What I mean by that is it will be routine for workmen and the emergency services to upload information on when lanes or whole roads need to be blocked (which I think it is already, just not in a neat joined-up way).
So when rolling out to urban environments, it will simply (in a relative sense that is) be a matter of extending the system.
However, if the construction worker needed to just block a road and couldn’t upload such information I’m sure an autodrive system would understand to do a U-turn and pick a new route. Or let the humans inside know there’s been a problem they need to make a decision on.
Again, by the time such a scenario is likely, autodrive will have been around a while (for highway driving) and people would be quite used to it. It would be seen as obvious that the selfdrive cars need to be told they’re in a convoy and/or they have right of way.
The former definitely better than human drivers, the latter possibly better, possibly just the same.
Agree that there are a lot of legal issues to work out.
In an abstract sense the legal issues are simpler for autodrive than they are for human drivers, but of course in reality we have all the legal structures for human drivers and it’s going to be painful to change, particularly while the two kinds of driver share the road.
I would, as long as it was the manufacturer who is legally responsible for accidents, not me (kind of how I’m just a passenger in a plane or train)
I think if I could opt to only use it on the long interstate segments and do the short local driving myself.
Substituting for that of the OP the hypothetical that self-driving cars will be substantially safer than the average driver, I don’t see why liability issues would be much different. You’d buy insurance, just like you do now, only the insurance company would charge less if you had a self-driving car, since it was safer.
If the self-driving car’s programming broke down and caused an accident, your insurance company might in turn file a claim against the manufacturer, but that wouldn’t be your problem.
For some reason, a lot of people - not you, AFAICT, but a lot of people - think insurance issues would be some insurmountable barrier to self-driving cars. I don’t know what they’re smoking.
You are much more optimistic about human ability to react to unexpected situations than I am 
I think your optimism does not match reality. Realistically, what many many humans do when presented with unexpected situations is totally incorrect and results in a worse outcome than if they’d simply done nothing. Faced with a surprise obstacle, humans generally either slam on the brakes or swerve to avoid an obstacle. This is very often the wrong thing to do, and makes things much worse. To give one example, people quite commonly swerve to avoid, say, an animal that runs into the road and end up colliding with oncoming traffic.
But it’s not even the uncommon cases that you should be worried about. It’s the common ones. Most crashes aren’t caused by something uncommon occurring. They’re caused by inattention, tiredness, impaired mental states, and unfortunate cognitive weaknesses. Even at our best, our cognitive capabilities (which evolved dealing with human-on-foot speeds) are pretty badly suited to controlling fast-moving heavy chunks of metal. We are terrible at judging the speed of something coming directly toward us until it gets very close. That’s why we need big bright red lights on the backs of cars to indicate that they’re slowing down. Without those, we’d have even more rear-end collisions (even with them, we have lots). We’re really bad at judging how long it will take to stop a vehicle. We’re incredibly bad at predicting the path of anything moving in a non-linear motion (either curving, or accelerating). Rationally, you should be more than willing to trade the small chance that your ingenuity will save you in a novel situation for the much larger chance that you’ll crash for a much more mundane reason.
But no one really wants to think about their limitations and weaknesses. They want to be in control, regardless of what the math and studies say. It’s why more people are afraid of flying, even though it’s phenomenally safer than driving. Loss of control is scary.
The other reason I’m very confident that self-driving cars will very quickly outpace human capabilities is that all those uncommon cases only need to be encountered once, by one car, and the software for all other cars will be updated. There will come a day when we investigate each car crash as meticulously as we investigate airplane crashes, and ensure that we will never again fail in the same way.
Unexpected events will still happen – driverless systems won’t control pedestrians, animals or freak of nature occurrences (falling limbs, etc). But the biggest unknown, what the other car is going to do, will be taken out of the equation. Your car will know what Car X is going to do and Car X is going to know what your car is going to do. And they’ll both know it before they “see” one another.
It’ll be boring as hell from one perspective. All four lanes on the beltway going a uniform 55 miles per hour. Which will actually get us to our destinations faster than when everyone is trying to go 15 miles over the speed limit.
I share the opinion of others who say that insurance issues will be of little consequence to the development of this technology. Insurance is a reactor to events, not a driver of them.
I certainly could have used a good collision avoidance system when I hit the side of a broken down truck last week trying to get around it. I see a lot of automated stuff to better help drivers coming online within the next 5 years, and some of it is here already, like cameras helping you to back up.
I think the next decades will show a lot of movement forward, but changing over to a completely automated system will involve a lot of changes to the infrastructure and cars as well as to people’s attitudes about driving, and I don’t see a complete changeover for maybe 100 years.
This video from Discovery talks about a Romanian teen who has invented a self-driving car or method for navigation that’s much cheaper than Google’s self-driving car. The parts cost about $4000 to add to the car, if I understood the video right.
Yes, this is coming about quickly. Human driving will be relegated to recreational purposes/areas. There will be a vocal minority who won’t want to give up the steering wheel; they will adjust to playing Crazy Taxi on their in-car Xbox instead, while being commuted to work by the autonomous vehicle.
Nissan Pledges ‘Affordable’ Self-Driving Cars By 2020
Man, if it happened and was safe and really affordable, I would really consider buying a Nissan (or whoever has them available at the time) just to have a self-driving car
That’s what I have issues with. They would have to be a helluva lot safer than what we have now. If that were the case, then sign me up!!
The Firebug will be old enough to drive in 2023. If this works out, he’ll be ‘driving’ a self-driving car.