AI Controlled Cars, or Your Right to Drive

[Borachio]

Schumacher.

I do see what you mean. But poor guy. Car, fine. No car, not so fine.

 

[Luckymoose]

All you had to say was "I cannot find any examples." You don't have to dodge it and try to spin my question into a different area that I wasn't asking about.

You think you're better so you want to continue the upward trend of automobile deaths in America? Thirty-four thousand deaths in 2012. Do you want those people to die? They obviously weren't better than an A.I. They died due to human error. Easily preventable human error. I don't trust you to not kill me while driving around, but I do trust two A.I to not hit each other, speed, run stop signs, etc.

Hopefully you get hit by a drunk, texting, fatigued driver who is speeding on an icy road. That'd be a really sweet deal, bro. Wouldn't it?

Moderator Action: Wishing violence on someone here is never allowed, wether in jest or not. This is a serious violation of our rules. There is always a civil way to get your idea across.
Please read the forum rules: http://forums.civfanatics.com/showthread.php?t=422889

 

[Angst]

@bhsup dude it's fine if you don't want to address my points- but i'm really curious about the pic you posted on page one. why won't you tell me about it

 

[Formaldehyde]

I don't see how that logically follows at all.

You think it "logically follows" that chronic traffic jams in exactly the same spots every single day are caused by a few bad drivers instead?

This is why highways are widened to allow more traffic instead of more cops being assigned, which ironically typically causes even more congestion when they merely pull someone over.

I mean, it's true that higher traffic densities are associated with more accidents. But why computer control couldn't eliminate them, especially under these conditions, is beyond me.

If you don't follow that a given diameter pipe can only move so much water per hour, or that a particular bottling plant can only fill so many bottles per day, I don't see how I'm going to convince you otherwise.

The great hunk of metal means there's a hell of a lot more momentum at the point of impact. This increases the force impacted upon the body of the person being crashed into. A skier doing 30mph is going to do a lot less damage than a car doing 30mph. As a result, I'd strongly suspect that while skiing accidents might be more common, the severity of an auto accident is generally much higher.

Auto accidents do indeed cause damage to the vehicles. But they don't typically injure or kill the passengers.

 

[Borachio]

No. It's not a few bad drivers at all. (Though, they can certainly have an effect by needlessly crashing into each other. )

It's the overwhelming driving habits of nearly all drivers, who simply cannot match their speeds properly to the prevailing conditions. Who do not look ahead, whose appreciation of hazard is poor, and who are busy dreaming about other things. I could go on.

 

[azzaman333]

So... that's no legislation on seat-belts or motor-cycle helmets, then?

Lack of seatbelts, definitely not. Having your body turn into a missile at the point of collision definitely increases the potential harm to others.

Not sure about mandatory helmets for motorcyclists (or cyclists for that matter). I'd need to see data regarding the relative costs of crashes with and without helmets.

 

[Silurian]

Why do you think automating cars is going to make most of them disappear?

I gave an example of how some cars would no longer be needed.

A two or three etc car familly could have one robot car.

X and Y drive to Xs place of work then onto Ys where the car is parked.

When X is due to finish work the car drives its self to Ys place of work then they go to the mall.

Whilst X is shopping the car drives itself to Y and picks them up. The car goes back to the mall to pick up X and they both go home and they live happly ever after.

 

[azzaman333]

Auto accidents do indeed cause damage to the vehicles. But they don't typically injure or kill the passengers.

And what about road users who aren't inside a car, such as a cyclist? And what about pedestrians? Do they not matter at all?

 

[Farm Boy]

It'll just be built into the cars at production. It might add a little to the sticker price. But the big story in cars like most consumer durables is falling real prices.

Well, just because you build something for inclusion on the line rather than after-market does not mean that you don't get to build it. In fact, it makes it more profitable since you get to build it in enormous scale. Then you get to build replacement parts, and service parts. Just because you don't see it as clearly at point of sale does not mean that it's no longer big capitalism. Much of big capitalism isn't in products you actually think all that much about buying. Cable ties, just as an example, make a lot of money from the auto industry. When was the last time you wondered about Panduit's profit margins on the vehicle you were buying(or whatever company that happened to make the cable ties for the car you bought)? And you have to be kidding if you don't think this will get worked into big data. There's probably even more money in that.

 

[Borachio]

Lack of seatbelts, definitely not. Having your body turn into a missile at the point of collision definitely increases the potential harm to others.

Not sure about mandatory helmets for motorcyclists (or cyclists for that matter). I'd need to see data regarding the relative costs of crashes with and without helmets.

I think the figures on head injuries for motorcyclists before and after mandatory helmets speak for themselves.

(Though, cheekily, I've not researched the matter.)

 

[Zelig]

The problem isn't the drivers with lack of skills (generally), it's the drivers that are situationally distracted and thus their skill level hardly matters. For that purpose, driving proficiency tests are insufficient.

I suspect there's a fairly large inverse correlation between being at top 10% driver skillwise and not talking on your phone, driving drunk, driving tired, etc.

I've thought about it too though, it shouldn't be too hard to stick some sensors in a car that disallow drunk driving, phone use, etc.

Have you heard of ZipCars? The "robo" part is obviously lacking at this point, but they've got everything else fairly well sorted.

Yeah, we've got car2go here. They're a subsidiary of Daimler, so some car manufacturers are already getting involved.

But that's like using public transport. What I like about my car is that it smells of me, and all the dirt and crap in it is mine.

More importantly, the crap isn't someone else's.

My car doesn't smell or have crap in it.

You're basically just complaining that you don't want a clean car.

And again someone comes up with the "saving lives" trumps everything argument. Then ban all swimming pools. Ban bungee jumping. Ban skiing. You cannot use that argument and ignore the people that harm themselves, so all of those other activities must also be banned for you to be consistent.

You got in a wreck and were harmed. I am sorry, I really am, but life is not a guaranteed bubble existence where you are ensured safe passage everywhere. So stop using that as an excuse to try to take away my privilege of enjoying a nice drive with my doggie.

I can't tell if you're being serious.

Robotic cars are basically like robotic surgeons. Currently, they're expensive and no safer than humans, but in the future they'll be cheaper and safer. Should be allow human surgeons to continue to operate because they enjoy the privilege of performing a nice appendectomy?

Why do you think automating cars is going to make most of them disappear?

The idea is that the robocars are all linked together - if you have a computer running the system it's not too difficult to regulate speed and following distance to eliminate stop and go.

I mean, if people were trained properly, you could already eliminate stop and go by just having everyone drive at a constant slow speed rather than repeated gas/break cycles.

Sometimes I'm not able to leave work at my usual time and get stuck in a bit of bad road design on my way home - I'm generally able to coast in first gear for 10-15 minutes, with the gap in front of my car growing and shrinking, while the car in front of me drives at the same average speed but with constant acceleration/braking. Me driving like this fixes flow for dozens of cars behind me, until the next jackass who can't abide maintaining a reasonable following distance.

Having robocars is essentially the same thing, except linking them together is far more efficient - since my car knows what the next hundred cars are doing, very little stopping distance is required and if an emergency stopping situation occurs the entire line of cars can brake simultaneously.

There is no ''near future'' where computer controlled driving is going to near adequate for any sort of road conditions that are not incredibly clean and predictable.

Human drivers aren't near adequate for any sort of road conditions that are not incredibly clean and predictable.

There are plenty of robodriving trials which are progressing fairly well.

Ah. So in your vision of the future some mega-computer actually controls everything which is occurring. That it wouldn't even allow you to drive if there wasn't a vacancy in its traffic control algorithms.

You don't necessarily need a mega-computer, it works fine with onboard computers that are more efficient than human drivers on their own, and that realize further efficiency gains when they've got a connection to stream data for other cars on the road.

It predominately comes from so many vehicles being in the same area. It is basic queuing theory.

In queuing situations, in the worst-case, you'd take the same amount of time, except there are loads of areas where human driving is sub-optimal.

At a traffic light turning green you can have the first 100 cars simultaneously accelerate, so you get through 80 more cars than human drivers would. And then cars 101-200 know the timing of the light, so they coast up to it slowly so that car 101 reaches the light just when it turns green again and then accelerates through.

 

[Silurian]

If you don't follow that a given diameter pipe can only move so much water per hour, or that a particular bottling plant can only fill so many bottles per day, I don't see how I'm going to convince you otherwise.

If there are many bubbles in the water, the pipe will be able to transport a lot less actual water and the bottles will take a lot longer to fill up.

A road with less air between the cars will be able to transport more cars.

 

[azzaman333]

I think the figures on head injuries for motorcyclists before and after mandatory helmets speak for themselves.

(Though, cheekily, I've not researched the matter.)

It's definitely something where, at first thought, it's something that only affects the motorcyclist, but then you consider the increased time and effort that paramedics, surgeons, etc have to spend on those without helmets, as well as much more difficult to estimate increase in costs associated with the mental trauma for witnesses of a more gruesome accident.

 

[Borachio]

If you don't follow that a given diameter pipe can only move so much water per hour, or that a particular bottling plant can only fill so many bottles per day, I don't see how I'm going to convince you otherwise.

But wait, your basic queueing theory will have told you that the fastest flow of traffic is to be found at the boundaries. Completely the reverse of what holds with a liquid flowing through a pipe.

I've got to question whether you know much at all about traffic flows?

 

[Farm Boy]

Human drivers aren't near adequate for any sort of road conditions that are not incredibly clean and predictable.

Many indeed are! Or are we going to regulate with only the lowest denominators in mind? That has interesting things to ponder as it relates to people's ability to pay for things, rather than their ability to be responsible.

 

[Cutlass]

I don't see how that logically follows at all.

I mean, it's true that higher traffic densities are associated with more accidents. But why computer control couldn't eliminate them, especially under these conditions, is beyond me.

What happens when humans control cars in a high density situation is that the drivers are constantly reacting to what all the others are doing. And since some of them are making unnecessarily aggressive moves, and some are being timid, there gets a lot of hesitation built into the system. Most stop and go happens because people are reacting just a little too much, or just a little too little, to something someone else has done. Which computer controlled cars, that aggressiveness and timidity is removed from the system, and the reactions are closer to optimal. And so the stop and go aspect of congested driving is reduced or eliminated, and so the whole of traffic flows smoother, and faster.

That's the theory, anyways.

 

[Farm Boy]

Reaganomics works in some theory too.

 

[Zelig]

What happens when humans control cars in a high density situation is that the drivers are constantly reacting to what all the others are doing. And since some of them are making unnecessarily aggressive moves, and some are being timid, there gets a lot of hesitation built into the system. Most stop and go happens because people are reacting just a little too much, or just a little too little, to something someone else has done. Which computer controlled cars, that aggressiveness and timidity is removed from the system, and the reactions are closer to optimal. And so the stop and go aspect of congested driving is reduced or eliminated, and so the whole of traffic flows smoother, and faster.

That's the theory, anyways.

Perfect zipper merges at the very end of the lane every single time.

Hell, if everyone is driving efficiently, a three to two-lane merge, instead of being a mess of congestion, simply becomes a spot where the traffic uniformly reduces by a lane and increases speed by 50% to maintain throughput.

 

[Formaldehyde]

If there are many bubbles in the water, the pipe will be able to transport a lot less actual water and the bottles will take a lot longer to fill up.

A road with less air between the cars will be able to transport more cars.

So now this theoretical control system would take even greater chances with the safety of the people than the current system supposedly does by providing the minimal amount of time necessary to react to situations? Because it would somehow automatically know when to slow at exactly the right moment?

The simple truth of the matter is that traffic congestion is predominately caused by too many vehicles. Automating the vehicles is going to largely take the bad drivers out of the equation. But it obviously isn't a panacea to traffic congestion caused by insufficient infrastructure.

Perfect zipper merges at the very end of the lane every single time.

Any reasonably competent driver already knows how to do this.

 

[Zelig]

Any reasonably competent driver already knows how to do this.

So a minority of drivers.

Probably the weakest of my points anyway, how many drivers properly crawl along at an average 5 km/h instead of constantly braking/accelerating in heavy traffic? 10% of drivers maybe.