Will you like a "driverless car"?

From my Post #8
The thought of how The typical driver deals with maintenance makes me shudder about cars which drive themselves.

They might be wonderful for the first 10-15 thousand miles, but I worry about the next 10 thousand miles.

SideShowBob Post #9
They'll know when they need maintenance and they'll drive themselves to the shop when the time comes.
I expect that driverless cars will have driver override & there are many idiots who would exercise it to avoid maintenance.

We are a long way from the technology required to eliminate drive override.

BTW: I am glad that my father was correct circa 60 years ago when he said
We will be fortunate when the predictions of common use of flying vehicles do not come true in our life time. The average driver is hardly capable of dealing with 2 dimensions: 3D would be disastrous.
 
I expect that driverless cars will have driver override & there are many idiots who would exercise it to avoid maintenance.
We've had the technology for years to prevent a car from starting when it needs maintenance.
 
Yes. Indeed, it can do a better determination of the odds of the resulting accident than a human could, since it can accurately measure closing speeds.

If safe to do so - yes, it could.

I agree that in time this could be a reality. How far off are we do you think?
 
I agree that in time this could be a reality. How far off are we do you think?
Not far at all. Autonomous cars already estimate closing speeds in other lanes to decide if it is safe to move over, even though neither of the cars closing on each other is the autonomous car. Likewise, they calculate whether it's safer to speed up or slow down to avoid an imminent threat (like someone about to T-bone them.)
 
Not far at all. Autonomous cars already estimate closing speeds in other lanes to decide if it is safe to move over, even though neither of the cars closing on each other is the autonomous car. Likewise, they calculate whether it's safer to speed up or slow down to avoid an imminent threat (like someone about to T-bone them.)
I understand that, but in order to have many of these things on the road, they all need to communicate with one another. There is alot of info to process if we want safety.
 
I understand that, but in order to have many of these things on the road, they all need to communicate with one another. There is alot of info to process if we want safety.
Why? Why would they need to communicate any more than drivers do now?
 
I don't want a car, can we engineer a society to avoid these stupid, expensive, and dangerous products? At one point decades ago, the car companies lobbied to force us to buy cars, it's a scandalous situation that needs to be corrected.
 
Not far at all. Autonomous cars already estimate closing speeds in other lanes to decide if it is safe to move over, even though neither of the cars closing on each other is the autonomous car. Likewise, they calculate whether it's safer to speed up or slow down to avoid an imminent threat (like someone about to T-bone them.)

Fully autonomous cars are sneaking up on everyone through the incremental introduction of semi-autonomous features like autopilots and collision-avoidance. There are already production cars being sold capable of that.

I'm told that some of the high-end production cars from Tesla and Mercedes-Benz are already autonomous cars except for the software. They already have the necessary sensors, computers and the ability for the computer to operate the controls. When fully-autonomous cars finally arrive, it might initially be in the form of software downloads to those cars able to accept it.

Which suggests that one of the biggest hold-ups at this point might be writing new vehicle laws for these things and figuring out stuff like accident liability. (Imagine the court cases that are coming!) I'm told that California's hellish Department of Motor Vehicles is currently in the process of writing a new 'drivers'/owners handbook for autonomous cars and figuring out what new rules should be in it.
 
Why? Why would they need to communicate any more than drivers do now?

In order for them to be aware of every possible dangerous situation on the road, every car would need to at least know location, speed, and destination of every other vehicle within a certain distance.
 
In order for them to be aware of every possible dangerous situation on the road, every car would need to at least know location, speed, and destination of every other vehicle within a certain distance.
Right - but why can't the methods they use now (lidar, radar, cameras) work for that? And why do they have to know the car's destination? Does accident avoidance work differently if the car two lanes over is going to Eloy instead of Tucson?
 
In order for them to be aware of every possible dangerous situation on the road, every car would need to at least know location, speed, and destination of every other vehicle within a certain distance.
This can only happen if ALL the cars on the road are automated.

An advantage of automated cars that is rarely discussed is the ability to travel much closer together, perhaps actually coupled like railroad cars. By being in constant communication, this will allow a much larger number of cars to use the road at the same time, in addition to being able to travel much faster than cars with human drivers.

In other words, traffic jams will be much less common, and much less dire when they do occur.

And of course the next step is for the cars to be interchangeable, making it unnecessary to devote such an enormous amount of land to parking. When you get out of the car that took you to work, the car trundles off to pick up somebody else and take him to some other place. Only in the dead of night would there be a significant number of cars not moving, but since they're all computer controlled, they can be parked IN THE STREET, which now has empty lanes to accommodate them.

But if there's just ONE non-automated car on the road, it destroys the system. Nobody can guess what that driver is going to do, and his reaction time is not good enough to be tailgating another car at 100mph.
 
Right - but why can't the methods they use now (lidar, radar, cameras) work for that? And why do they have to know the car's destination? Does accident avoidance work differently if the car two lanes over is going to Eloy instead of Tucson?
I think destination would be helpful. The other cars would know the route that your car is taking. They could anticipate a merge for exampke.
 
This can only happen if ALL the cars on the road are automated.

An advantage of automated cars that is rarely discussed is the ability to travel much closer together, perhaps actually coupled like railroad cars. By being in constant communication, this will allow a much larger number of cars to use the road at the same time, in addition to being able to travel much faster than cars with human drivers.

In other words, traffic jams will be much less common, and much less dire when they do occur.

And of course the next step is for the cars to be interchangeable, making it unnecessary to devote such an enormous amount of land to parking. When you get out of the car that took you to work, the car trundles off to pick up somebody else and take him to some other place. Only in the dead of night would there be a significant number of cars not moving, but since they're all computer controlled, they can be parked IN THE STREET, which now has empty lanes to accommodate them.

But if there's just ONE non-automated car on the road, it destroys the system. Nobody can guess what that driver is going to do, and his reaction time is not good enough to be tailgating another car at 100mph.
Yes. That is what im getting at. I am skepical of this technology to be any safer with human drivers sharing the roads.
 
Yes. Indeed, it can do a better determination of the odds of the resulting accident than a human could, since it can accurately measure closing speeds.

If safe to do so - yes, it could.

Ok
One little thing:
Signs demand "trucks use right lanes" so my action may not have computed well. I didn't just move over one lane, but took the far left lane, and did it safely.
Hitting the brakes would have had the effect of causing the drivers behind me to switch lanes which would have enlarged the accident.
Staying in the same lane at the speeds of the vehicles in front of me would have involved me in the accident. One needs to also understand the normal driving behaviors of the people sharing the road, assess the gaps between cars, nd act quickly in a determined manner.
I doubt that any current computer could take all of that into consideration and act accordingly.
(admittedly------had the loonie driving the big buick 225 been replaced by an ai, the initiating accident might not have happened.)
Loonies will always remain a random variable.
Just looking at the computer models for predicting climate should indicate the limitations of a computer.
...........
comparing chess and driving is much worse than comparing apples and oranges, more like comparing lignite and kittens.
 
Ok
One little thing:
Signs demand "trucks use right lanes" so my action may not have computed well. I didn't just move over one lane, but took the far left lane, and did it safely.
Right - which AI's will do as well.
Hitting the brakes would have had the effect of causing the drivers behind me to switch lanes which would have enlarged the accident.
Staying in the same lane at the speeds of the vehicles in front of me would have involved me in the accident. One needs to also understand the normal driving behaviors of the people sharing the road, assess the gaps between cars, nd act quickly in a determined manner.
Right. An AI will realize that accelerating can move the car to a safer location within the traffic flow.
Just looking at the computer models for predicting climate should indicate the limitations of a computer.
Traffic is far more deterministic than weather - or climate.
comparing chess and driving is much worse than comparing apples and oranges, more like comparing lignite and kittens.
Uh - OK. I wasn't comparing the two.
 
apples and oranges----------response to post 37
..........
billvon, i suspect that you have far too much faith in ai
ergo "I'll believe it when I see it"
 
Back
Top