The Future of the Car

Gandul writes "Radar, lasers, wireless radio networks and other embedded tech will enable our cars to sense faraway traffic and stop accidents before they happen. But who will be in the driver's seat?"

Comments

Hopefully not people

[bigtrike]

The vast majority of accidents are caused because human beings are either incapable or unwilling to drive a vehicle safely. Because of this, we have lost many civil liberties. Due to safety concerns, the police can pull you over and search your vehicle at almost any time without real justification.

I'd rather have robots drive.

The changes that should be made

[Mishra100]

I really do think they need to start focusing on a rail type system that does all the driving for us. If we finally would convert roads to a electromagnetic railway system (like the bullet train) and just program cars to drive and stop when they need, then we would have a much much better system than we have now. This completly gets rid of Car insurance, gas, 95% of death related accidents(I would have the 5% is left for cars that malfunction), drunk drivers, pollution, and many other negative aspects.
I definitely think it would takes a lot of time to complete and would cost a ton of money. But we as citizens and as a country would save a whole lot more money having this implemented as a final solution to all of the stable and rising issues that circles around transportation.

Re:The changes that should be made

[egarland]

Rail can't replace roads. It's much less flexible than pavement, more expensive to maintain and not compatible with the existing transportation.

People aren't going to be running rails into their garages, to their front doors.. across the lawn to where you need to back up to hook up the trailer.. etc. When there's an accident a rail vehicle can't just drive on the dirt to go around which may not seem important until you think of the fire truck that's coming to pry someone out of the wreckage in that accident. Rail isn't flexible enough for a general purpose transportation system. That lack of flexibility is one of the two advantages you have with rail. It lets you predict exactly where things will travel and run things like power lines to them. That advantage is it's downfall when it comes to general purpose transportation though.

The other advantage is lower rolling resistance. As speeds go up air friction accounts for a larger percent of the energy used to keep the vehicle moving so as speeds increase this is actually less important.

Also, car insurance wouldn't go away it would just get cheaper. Gas may go away but you have to power the vehicles somehow and since we aren't building any more clean environmentally friendly nuclear power plants we'll probably be burning oil or more likely coal which dumps tons and tons of mercury into our food chain every year (anyone know what the half life of mercury is?)

The benefits you describe could be here soon, but the only realistic way to get them is if computers drive our cars. That's the right answer.

Re:The changes that should be made

[NaruVonWilkins]

Bullet trains are NOT electromagnetic railways. They are almost always standard gauge rail with catenary and messenger wire. There are a few isolated bits of maglev out there, and a test track in Japan for possible construction of a new Tokyo-Osaka alignment, but afaik, there's nothing in public operation that's longer than the Shanghai airport maglev - somewhere on the order of 50km.

Flamebait? wtf?

[coshx]

The vast majority of accidents are caused because human beings are either incapable or unwilling to drive a vehicle safely

This is plain truth. Most accidents are caused either because the drivers chose to drive wrecklessly and/or under the influence, or were caused simply because human reaction time is not as good as computers' reaction time.


Because of this, we have lost many civil liberties.

This is also true, and quite an insight. Think about random road blocks where you're tested for being under the influence even if you're NOT driving wrecklessly or even swerving. The equation is simple: am I willing to give up a little bit of my privacy to prevent myself from being killed? Generally, yes! Of course! But, if drunk driving didn't cause accidents because people weren't driving, there would be no need to pull this person over.


Mods, please please please stop modding based on your own beliefs, and rather based on the intelligence of people's responses -- I'm going to get modded down for that, eh?

Re:Flamebait? wtf?

[lazlo]

On a trip to Ireland I completely figured out why there's a drunk driving problem in the US: There isn't a pub within walking distance of your house. In Ireland, there is. It doesn't matter where you live, there's a pub just around the corner. Why drive there and back, when it's easier to walk?

So the solution to the US drunk driving problem is simple: Build more pubs.

"Virus kills hundreds on I-95"

[sexyrexy]

It has generally been the trend that the more complex a system becomes, similarities it will have to the foundations of the modern operating system. ATMs are a prime example of machines that started as moderately sophisticated PCBs and now routinely run Windows Embedded.

If a vehicle is "smart" enough to handle driving, it will have the computational power and flexibility to run reasonably sophisticated software. Consider that increasing wireless bandwidth (WiMax, anyone?) will lead to offloading the heavy-duty positional and map processing to a remote service over the Internet, with the software to display becoming a thin client for a remote database. A clever programmer will find a stack overflow in MapQuestClientForYourCar and BAM! Suddenly cars are automatically veering for each other instead of away.

The level of scrutiny and security applied to such systems will have to be on par, or higher than, such applications as air traffic controlling before it can be considered safe.

Re:"Virus kills hundreds on I-95"

[vandoravp]

I don't think virii will become a fatal problem simply because of the way the cars are likely to be controlled. I can't imagine any sane engineer linking the OS to the intercar communication system in a way that would allow direct control of the car from an outside source. The external communication/gps positioning is purely for additional road sensing and navigating-not for the actual driving and avoiding of obstacles-and will probably happen in a separate system. The core driving operation will likely be kept isolated from outside control. Also, remember that, in the early years of these systems, humans will still be driving. The extra sensing will only notify the driver of conditions and attempt to prevent (or at least alleviate somewhat) an accident should one begin to occur.

It certainly would be possible to suddenly make all cars lost but that wouldn't be so much of a safety issue. The real danger in this sense will be possible backdoors or dangerous easter eggs that could be inserted during the OS development process.

Who cares who is in the driver's seat...

[temi]

I mean when it comes down to it people may not be in total control, aka "the driver's seat" when this far-off car revolution happens, but they will be in the car, and that is all that matters. In general people use cars to get places so whether they are driving or not they really just want to get to their destination.

Of course non-destination based travel idiosyncracies(sp?) arise because it is taken for granted that someone has to drive the car. Hence you have "cruising the strip", "joy-riding", "drive-in movies", and other random and BASICALLY useless things.

The main point im trying to make here is that there is no BIG, SCARY revelation in... "OMG WHO IS DRIVING THE CAR?!?!!! W3 4R3 S0 PWN3D!!!"...who cares, we will always get where we need to be and if worse comes to worse i bet they will even let us sit in the front seats so we can feel like we are still driving.

Automatic speed control

[Migraineman]

Automatic speed control will never happen. Not because the tinfoil-hat brigade will be successful in lobbying against it, but rather it'll be the law enforcement personnnel who will kill it. Speeding fines are too large a part of the police budget, and that opportunity must be maintained. The police vehemently oppose any measure that justifies a reduction in the number of officers required. The insurance companies will probably oppose anything that eliminates fender-benders, too. Fatal accidents cost them money, but the fender-benders are income generators.

I'm not paranoid, just following the money.

Re:Automatic speed control

[Qzukk]

My last fender bender cost less than $1000 in all to repair both cars. Guess how much the insurance company raised my rates for the next 4 years or so?

Re:Automatic speed control

[mobets]

Because if they have to pay anything, they raise your rates. Then they add it into the statistics and raise the rates of everyone like you because yall are statisticaly more likely to cost them a bit of money.

Just a guess, but it sounds good.

Dual-mode vehicles

[Russ Nelson]

I think the future of the car is dual-mode vehicles. That is, a car which operates as today's cars do, but which can also drive up onto a monorail. One design is the RUF. On ordinary roads, it runs off batteries. Not a trunkfull of lead-acid batteries, but a modest battery, sufficient to get from home to the nearest monorail. Maybe a 50 mile trip max. Once on the monorail, electrical pickups power the vehicle. On the monorail, the vehicle is mechnically inherently safe. Braking works by gripping the monorail, not relying on the weight of the vehicle and a constant coefficient of friction with the road. So with reliable braking, vehicles can form a phalanx, to increase traffic density and reduce wind resistance.

Vehicles on a monorail will drive a 90 MPH, and do so with great safety. Even grandma, because the cars are computer-controlled on the monorail. You designate your exit, and the computer takes care of routing you. Each car does its own routing based on global traffic announcements. Just like BGP4 on the Internet.

Damn but I'd like to say "Take me to Boston and exit onto Boylston St." and then read a book, or fall asleep, or use the Internet access provided by the monorail connection.
-russ

Many problems yet to be solved.

[Pemdas]

I've actually worked for a number of years on autonomous vehicle technologies, and am more than passingly familiar with most of this stuff.

Wireless ad-hoc traffic information networks run into some major security issues. How do you establish trust? If the trust model is basically wide open, then antisocial people are going to put together systems which look at your route and start telling other cars "Avoid these roads at all costs! It's backed up for miles!", so that their own personal drive is relatively free of cars.

How do you prevent this? Do you require warnings from multiple sources before you believe them? Then you've just increased the required critical mass before usability by an order of magnitude. Do you trust that automobile makers can put together some sort of embedded crypto system that's "secure enough" and "tamper proof"? Well, that's worked so well for the DRM people, hasn't it?

Of course, if you're relying on the wireless system for safety, you're essentially giving the ability to swerve/brake hard to systems you don't own, so the matter of trust becomes even more significant, and liability becomes killer. Any way you tie the systems together to try to keep people safer, there's someone who's going to argue (with a non-negligable probability of success) that you should have done it a different way, and now you owe someone $millions.

In addition, liability is going to keep this stuff down for a while yet. No autonomous system is ever going to be perfect, and when dealing with loss of human life, liability more or less demands perfection. If I could put together a fully autonomous system tomorrow which provably had 99% fewer accidents than human drivers, I'd still get sued by the 1%.

This is the primary reason all remote sensing tech on the market today is in the form of "driver assist". If your system screws up, it's still primarily the driver's responsibility to avoid accidents.

I'm not a complete pessimist. I don't think the issues I'm raising are insoluable, and I believe we'll have good autonomous systems eventually. I just think the problems are fundamentally hard, and the legal environment doesn't help; it may be a few decades before the more exotic stuff gets into production cars.

Re:More to the point...

[L.Bob.Rife]

Are you aware that the #1 cause of accidents in my state is people that drive too slow?

Statistics

[SuperKendall]

Remind me where you live so I can stay the hell out of whichever state recommends driving faster to avoid accidents as safe driving.

Well I was actually trying to be funny, but in reality when they raised the speed limit in Colorado from 65 to 75, accident rates went down.

It turns out if you make smaller the DIFFERENCE in speed between other cars, you have fewer accidents - that's what people who dislike speeding cannot understand, overly slow drivers are actually just as dangerous as people who speed excessivley. Both cause accidents, as statistics from Colorado show when the speed limit was simply set to the speed people generally drove anyway.

So by avoiding the states that increased speed, you are putting yourself in harms way. Good luck wtih that!