Automated cars, your thoughts?

I am all for this myself. I consider myself a good driver, but I don’t find it fun or relaxing (except on a dead country road), and get stressed easy. I would much rather read a book or surf WiMax while going from point A to point B. (especially long drives cross country - drive at night and sleep, wake up and you are there)

Some people have concerns about safety, but I think the switch off would be a decrease in accidents. Of course there are those that would want to ban it if one computer malfunctioned and somebody died, forget the fact that several people die in the US every day in car accidents due to human error.

What are some thoughts on if this will happen some day? Do you think it is inevitable? Do you look forward to it, or do you think it is fantasy?

I know there are many people that love driving and are very good at it, and those people might resent having a computer control their car, but what if you can turn it on and off at certain times/places?

What do you feel about this, would you ever use it? Would you resent being forced to use something like this in rush hour traffic?

I would love to see positive feelings, but I want to hear the negative comments too and why you have them. Everybody I have asked (not that many really) has liked it, but didn’t think it would ever happen. I think it is inevitable. Technology is moving in that direction and since so many people are lazy (like me) and or don’t like driving, I think there will be quite a few people taking up the idea.

Thoughts?

The problem with automated cars is that you would pretty much need roadways made specifically for them. At minimum, you would need to embed some sort of RFID chips into the roadway so the car could locate itself. This would cost billions and take decades to perform, not to mention developing the vehicles themselves. It would also kill the car industry, since there would effectively be no point in getting a better or differentiated car.

Outside of practicalities though, I hope very much that I will one day see them become the primary means of transportation in the US. People dying due to car crashes is entirely needless even with today’s technology.

Given that even the best OS’s we have these days still can crash and thus require a reboot, I can’t see it happening any time soon. Even with a triple redunant system there could be problems, and presumably, the car would be able to connect to the internet for software updates, which means that it could be vulnerable to viruses and other forms of malware.

Seeing as how there’s so many inattentive drivers now, imagine how much worse things will be when they’re not required to pay attention at all. I can well imagine them staring at the controls with a blank expression on their face as alarm claxons go off.

I think automated cars are a great idea. Maybe we could expand the idea to a kind of automated carpool – since there are many popular destinations that multiple people want to go to, they could all get into the same automated car (built larger than the individual model) and read or surf the web on their way.

For safety, it may be necessary to put some of these automated carpools on dedicated tracks. Some may even go underground.

:wink:

Interestingly, DARPA is sponsoring an Urban challenge this year for cars that drive themselves. Link.

From that link:

So, there are people working on it. ETA: Oh, bolding mine, of course.

I don’t think it’ll happen within any practical period of time. It’s not technical - we can have the technology worked out withing 10 years, easy. It’s the damn litigitiosness of the American people. There’s no way an accident resulting from a failure, or even perceived failure of the auto-drive system would not result in a lawsuit for millions.

There’s just no financial incentive for private-sector auto-drive systems.

There are many businesses that are routinely sued for millions - airlines, hospitals, pharmaceutical companies, etc. Most of them seem to do OK. Even tobacco manufacturers still survive.

Anyway, automated cars would be nice, and should save a lot of energy by allowing closer spacing between cars (thus reducing air drag), and by allowing a more constant speed (fewer unnecessary accelerations). But even then, it seems wasteful to have one vehicle per person. I’d much rather see the money spent on a nationwide high-speed rail network and suburban/inter-urban mass transit systems.

I saw a thing on Future weapons the other day that had an all terrain robot vehicle that drove itself. It had sensors and an AI that navigated around things it couldn’t go over. The builder said it was doing this due to it’s algorithms alone, and there was no special programming or paths set. They said this technology could be available in less than 10 years.

So no special tracks or paths, just smart cars. Though the tracks could add extra safety.

I saw a comedian talking about this the other week. He said:

“You get in the car, you tell it where you want it to take you. You don’t have to do any of the driving. It drives you to your destination. You get off. Amazing! It’s like… The bus!”

Automated vehicles themselves are not a huge problem. Modern jet airliners have autopilots that can fly departure runway center-line to arrival runway center-line with no problems. Automated landings have to be tested frequently and it was possible that you were on a flight that landed automatically and never knew it. Trains and subways are also not difficult to automate.

You might think that automating the big vehicles like 747’s and trains would be hard and transferring that knowledge over to cars would be easy. You would be dead wrong.

Given unlimited money, dedicated roads, limited usefulness, and greatly added hassle it would probably be possible to design a system in a limited area that would drive cars around on something like tracks avoiding other cars. In reality, this type of system requires highly developed AI which both computers and their human designers have shown to be piss poor at. The test system would require insurmountable costs.

The difference between keeping a few trains or planes on track in a large but highly defined space is completely different than keeping 10’s of thousands of cars rolling along with their own computers, human inhabitants, destination, and inevitable breakdowns. Those problems would apply to well-maintained Interstate highways. Once you need to integrate secondary roads, the problems become infinitely worse.

The idea is so enormously complex that I don’t expect to see it in widespread use in the next 50 years or so of my lifetime.

Yeah, I’m not seeing the need to change the roadways. GPS and intelligent nav systems is all you would need. Implanting RF chips or whatever in the road isn’t even going to be necessary. So would tracks. I’m pretty certain with today’s technology we could do this, what is likely going to take so long is incorporating multiple levels of safety and more than just a few redundant systems.

If you can keep the number of breakdowns and wrecks below the current number, then it is a benefit, no?

Oh, and for what its worth, there are more than just a few people in Academia (as well as the army as cited in Future Weapons) working on this problem, I’m sure they are trying to find alternatives that don’t include massive changes in infrastructure. Links:

http://icbs.berkeley.edu/automated_vehicles.php
http://www.ece.osu.edu/citr/Demo97/osu-av.html
http://faculty.washington.edu/jbs/itrans/dualparn.htm

http://faculty.washington.edu/jbs/itrans/bishopahs.htm

interestingly, the one above has this:

Also, it appears GM is going to launch a self-driving car in 2008. Cite

Now, I’m not trying to change your opinions, just cementing other readers with the idea that this isn’t science fiction, and it may NOT be 50 years away, it may be soon. Sooner than you think.

At least, thats what I think. :wink:

That’s the problem, it is always much less than 50 years for all AI predictions. None of of them work as predicted because computers and the human brain have about .000000001% in common. One is completely natural at what the other does and vice versa but they simply cannot accomplish what the other is good at on a practical scale.

My job is as an Business Systems Analysis which is lightly related to this field and when I hear this stuff, my brain starts spinning through all possible scenarios, the costs, and the benefits of automation. This one fails miserably. It actually fails worse than flying cars which have almost been with us for more than 50 years and yet none of us have ever ridden in one. There are flying cars just as there are automated automobiles. It is just that they are both unworkable from several levels including AI (almost 100% failure), personal costs (no one can afford such a thing), and the chicken and egg problem of infrastructure.

GPS is probably the most promising technology bringing up the rear of the WWW. Unfortunately, it can’t deliver this stuff except in the lab because of AI and general control problems.

However, technology usually surprises all of us by ignoring everything we thing will come to pass and bringing in something even more amazing on its own.

Your car already has a computer in it, which if it fails the car doesn’t run. It’s not that difficult to make a OS/Software that doesn’t crash. The reason that your home computer crashes (which most of the new OSs do very rarely) is because you have an essentially infinite number of hardware, software, and periphial combinations. It’s impossible to make a fail-safe system with that number of different components. If you limited it to, say, one hardware combination running one piece of software, then you essentially get perfect reliability. At least much better than any human driver could obtain.

The other important thing is that this could dramatically increase fuel economy. Cars would be able to drive very close to each other, and draft off each other. They would be better at accelerating slowly, and conserving momentum by timing lights.

I’m not trying to make light of the engineering challenges, but with a concerted effort we could build one right now. It would take a massive amount of money to retrofit cars, but it’s certainly doable. The problem is that I don’t think anyone is willing to pony up the investment,

I’m not certain you need AI on the level of humans. What makes you think human intelligence is the best possible AI for navigating a vehicle at break-neck speeds? A relatively “dumb” system may be better suited.

Again, what infrastructure is needed? None according to my cites. Flying cars don’t work well because people are dangerous enough on the road, flying around presents more dangers than rewards. Computers specifically designed to do nothing but drive safely and contain redundant features for increased safety are nowhere near on the same level.

Truer words have never been spoken.

I don’t think it is going to be one huge concerted effort implementing this, more than likely it will be a gradual process over the next 10-15 years. Just like nobody went around handing out cell phones to everybody. It was a slow and gradual process that suddenly exploded. The internet grew like that too.

I personally think it will be cars that assist with driving tasks, like an improved Cruise Control for highway driving, GPS systems improving and implementations of road databases downloadable by vehicles. More and more computers, electronics and gadgetry are being implemented in a car, in 10-15 years, it will just be something that had happened. Thats how I see it. No sudden change at all.

I suppose it could be gradually introduced in some areas, but it’s still going to take an investment. It’s unrealistic to expect a system to be able to optically recognize things like stop signs, traffic lights, other vehicles, and the dozens of other things drivers need to recognize. It would be much easier, and much more reliable in the long term to switch these optical signs to radio broadcasts. Each car could be carrying a chip that transmitts things like it’s velocity, size, GPS position, etc. Stop lights would be replaced (or modified) with radio broadcasted signals.

I think the very best implementation of the self driving car would be for long freeway trips. Out here in the Western end of the US we have nothing like decent railway service and air travel is a fucking nightmare. If you do take a plane, you just have to rent a car at your destination in order to get around. So a dedicated pilot lane on the freeway which is only accessible to cars under remote control would be absolutely freaking great. Drive from your home to the freeway. Input your destination. Let the pilot lane take care of it as you sleep, relax, surf the internet, watch scenery, screw in the backseat, whatever. If you need to take a potty break, let the guidance system know and it puts you into the non pilot lane a mile from the next exit with services, you drive off the freeway, get food, take a pee, refuel, whatever, then get back on the freeway and resume the trip. When you get to your destination you have your own car to run around in. Less stress on the driver, fairly cheap method of travel, get to see the sights as you go–I’d sign up for it in a heartbeat. I roadtrip all the time anyway, and since I’m the only driver in the house it can get a little tiring, so leaving all the boring freeway driving to the car would be absolutely wonderful.

I think SmartAleq has it.

I can’t see fully-auto cars happening, well ever, because of the problems raised by Tuckerfan et al. However, a car that can slip into auto mode on dedicated freeways and back to human control on other roads night be workable.

Then, this raises other problems. What if the driver has since fallen asleep, and is coming to the end of the automated roadway at 60mph? Beep beep beep resuming manual control beep beep ZZZZZZZZZZZZZZZZZZZZ WOOP! WOOP! WOOP! ZZZZZZZZZZZZZZZZZZZZZZZZZ… There are arguments that even the level of automation we have in cars today (auto trans, cruise control etc) has serious secondary “knock-on” safety effects by increasing drivers’ “remoteness” from their surroundings.

There’s a part of me that thinks greater automation is a good idea. There’s another part that suspects road casualties would go down if everybody were forced to drive a manual transmission car with no cruise control, cup holders, telephones, etc.

I remember back in the early to mid 1980s on one of those “Towards 2000” technology shows, they had a system in Japan with vehicles on a test track operating under a type of remote control. It was meant to simulate a freeway. The lead car was manually driven, but others could slot in behind it about five feet off the rear bumper, and when the first driver braked, the other cars all did at the same time. This was supposed to increase the amount of traffic a given stretch of roadway could handle in a given time period. Even at the time, though, I could see a fundamental flaw in this: the cars were all of identical make and model. Bring in a Subaru WRX, a semi truck, an SUV, a Mini Minor, etc, all with differentr braking characteristics, and see if it works then.

I don’t think we’ll ever see this technology. The 50 year timeframe seems reasonable, but by then, we’ll probably be thinking in completely different terms, and the private motor vehicle will doubtless be performing a vastly different role to today (a smaller one, I’d wager).

No, it doesn’t. They weren’t sticking computers in 1981 Honda Preludes. :smiley:

While you’re right about why home PCs crash (which is, as you state, getting better), even idealized hardware situations like you describe are vulnerable to OS lock ups. NASA had to reboot the Mars rovers shortly after they landed and has had to do the same with different spacecraft. Presumably, automakers will do their best to work out all the kinks before selling the cars, but how long do you think that it’ll be before people start modifying their cars to get better performance out of them? Then, as the cars get older, people will start hacking the systems in them so that they can keep the cars running because they can’t afford the cost of the correct repairs.

You’d also be less likely to have people slowing down to rubber neck.

It’d also take a massive amount of money to educate service personnel, promote the idea to consumers, and modify the production lines. I agree that few people would be willing to pony up the necessary cash to get it done.