Slashdot Mirror
GM Says Driverless Cars Will Be Ready By 2018
Gregor Stipicic writes "Cars that drive themselves — even parking at their destination — could be ready for sale within a decade, General Motors Corp. executives say. 'This is not science fiction,' Larry Burns, GM's vice president for research and development, said in a recent interview. GM plans to use an inexpensive computer chip and an antenna to link vehicles equipped with driverless technologies. The first use likely would be on highways; people would have the option to choose a driverless mode while they still would control the vehicle on local streets, Burns said. He said the company plans to test driverless car technology by 2015 and have cars on the road around 2018."
I'm sure the AI will drive much better than some of the people on these roads in Boston LOL
... but will they fly?
Just imagine the crossover deals this will open up.
"Drive to Pathmark"
"Pathmark is overrated. Destination modified to Walmart." *doors lock*
Too bad others have predicted the world will come to an end in 2012.
...that someone will have to come up with maps that are accurate? I don't mean ones that have pinpoint accuracy on the locations of roads, but thoroughfares with special conditions. I'd hate to riding in a car in autopilot that decided it could turn the wrong way down a one way street because the map data didn't show it.
I haven't driven a GM car in years. So yeah I can see GM cars becoming driverless within my lifetime.
I will remain pseudonymous, but I will say that my current area of research (I am a graduate student) is tangentially related to this field, related enough that I've looked into trying to convince GM to give me funding (so far nothing has materialized). Specifically my research looks deals with programming language design (e.g., making less-than-Turing-complete-but-still-useful programming languages structured in useful ways) to aid in static analysis. The aim is at safety-critical code (nuclear power plant code, industrial controller code, automotive software) such that you can say "barring hardware failure, this code is 100% guaranteed to meet hard realtime constraints", etc.
Anyway, at least publicly, GM is probably the most impressive car company in terms of researching these sorts of things. I feel kind of bad for GM. I hear they're selling terribly and are even selling at a loss on many cars, but their research department really is something impressive. Maybe they're a little bit Microsoft-ish in that their research department is heavily insulated from the rest of the company, I don't know. But GM is doing a lot of cool stuff and funding a lot of cool stuff with regards to "correct" software.
If it were some other random company, I would probably roll my eyes and say "oh they'll probably just test it really really heavily and then tell us that it works", but more than most companies, I trust GM to develop cool technology (such as novel static analysis techniques) to get this to work. Their R&D is active in a lot of areas, 99% I'm sure will never amount to anything, but I wouldn't doubt it if they could get the technology together to get auto-driving cars in 10 years.
Disclaimer: as I mentioned before, my efforts to get GM funding are still unsuccessful, and consequently I'm not on GM payroll in any imaginable way. I don't even drive a GM car (or any car). In fact their cars look kind of lame in general, but their R&D department in Cool.
Granted, it was only driverless because it was buyerless, but there is prior art here.
Eloi are stupid, throw morlocks at them!
So, when a driverless car runs a red light, who gets the ticket? The owner? The manufacturer? The software company? Hell, they have automated machines that issue red light tickets now, so will one pile of metal and software issue the ticket to the other? Will the machines develop their own monetary system, will driverless cars figure out hacks to avoid the tickets, and will the robot machines have their own jails and prisons? Capital punishment = execution by power surge or by fatal software virus? This smacks too much of a bad Twilight Zone episode.
"Every great cause begins as a movement, becomes a business, and eventually degenerates into a racket." -- Eric Hoffer
Apparently there'll be a copy of Duke Nukem Forever in the glove box.
A thistle is a fat salad for an ass's mouth...
I'm not an attorney (I'm also not an acronym kinda guy) - But it seems by assuming control of the car GM would also be assuming responsibility for the occupants of the vehicle and any other involved in a collision.
Correct me if I'm wrong.
It seems to me the only way this technology ever winds up on the road is if the owner of the car signs a waver at the car dealership to hold GM harmless and assume all responsibility for driverless mode accidents.
With the speed with which processing power and sensors become cheaper and more widely available, I think 10 years is definitely attainable. The tech is here, most of the problems are solved, we just have to wait for the price point to come down.
/.ed. ;)
[1] I was going to put our URL here, but the IT dept will kill me if the servers get
Any plan which depends on a fundamental change in human behavior is doomed from the start.
Between texting, eating, putting on makeup, smoking, futzing with the radio, surfing the Internet for the nearest Burger Doodle, and so many other things to do in the car, driving is SUCH a distraction.
Even if the technical issues were all resolved (which is not guaranteed by any stretch of the imagination), what about the legal and insurance issues? Until the insurance companies jump on the bandwagon, this will go nowhere.
It's not like ALL the cars on the road will be driverless. Who is responsible for a crash that occurs while you aren't driving and are reading or asleep (why else would you want a driverless car)?
They might have better luck putting driverless "taxis" in crowded downtown areas where traffic moves slowly - that would reduce the damage and injuries associated with accidents at higher speeds.
Never let reality temper imagination
Never let reality temper imagination
I need a driverless car. I want to be able to sleep/study/play video games while on the road. I want the the perfect attention and instant response of a computer keeping me safe. And I don't want to pay someone to do it for me.
Seriously, how can you even ask this question? Have you ever heard of the broken window fallacy?
A slashdotter who didn't build his own computer is like a Jedi who didn't build his own lightsaber.
GM had cars during the 80s that got better fuel economy than most hybrids of today (50-60 mpg, with carburetors no less). Few wanted them then and few want them today.
That's a very important point, and I hope more people take note of what you said. The primary barriers to this kind of thing are political, not technological. If I injure or kill someone through my driving, what's the most you could hope to sue me for? Maybe a million dollars. But if the car was self-driving, well hey, that's a company with deep pockets. You could sue me for a lot more!
Now who can handle the insurance policy on that?
Then, of course, inane regulation.
Never mind that these will be safer and less obstructive than 95% of drivers. Never mind that they'll end the problem of drunk driving. Never mind that they will massively increase productivity. Everyone has to get their piece.
Apology to Ubuntu forum.
How many times have heard the story that technology X is only a decade away, then another 10 years later Technology X is ust another decade away?
In my book, if you an't roll something out within 18 months, it's vapor. Talking about something you think is a decade away is just lip service clearly trying to generate some PR and drve up stock a few cents for the day.
- It wouldn't race to the next red light, but infuriate you (like I do my passengers) by taking its virtual foot off the virtual gas pedal as soon as the light ahead turned red
- It wouldn't waste gas idling at the green light with its finger up its ass
- It wouldn't go east to get west (unless Microsoft made its nav system)
- It wouldn't pick the route with the most stop signs
You can already improve your mileage on the interstate (or autobahn) by using your cruise control.-mcgrew
mcgrew's razor: Never attribute to stupidity that which can be explained by greedy self-interest
I'm sure designers have taken this all into account, but I'd still be concerned with control systems for pedestrian avoidance, sensors determining whether the small object in front is a newspaper or a rock to be avoided, and predicting behaviors of bicyclists, etc. Sometimes its better to run over a squirrel than break suddenly and risk being rear-ended or swerve around it.
GM's problem stems from the fact that they've had since the oil embargos in the 70's to improve their auto's milage. No intelligent person can argue successfully that GM's engines aren't a lot more efficient than they used to be. It's that they've wasted all of these efficiency gains on increasing horsepower to drive heavier cars more quickly.
GM's had 30 years to bring fuel efficiency & milage to the forefront of their goals. I have no sympathy for its demise.
wow homers got one of those robot cars!
(crash!)
one of those AMERICAN robot cars...
Website Just Down For Me? Find out
There will be mistakes and deaths, but they will be far fewer than we have today.
I think you're right. The uncomfortable part, though, is that when you're driving the car, you feel that you have the control over avoiding problems and accidents (I say you *feel* because in reality, there are some accidents you can't avoid). On the other hand, if the software BSODs and drives you off the bridge, you had nothing to do with it at all. Every time you get into one of these cars, you're putting your life into the "hands" of a piece of software/hardware.
Even on planes, which are very dependant on correctly functioning computers, there's a feel that the pilot could do something to save the day in case there's a massive failure on the plane.
Never underestimate the bandwidth of a 747 filled with CD-ROMs.
* Getting a computer to recognize the difference between a toddler crawling on the road, versus a doll on the road.
This can be done, there are object classification algorithms that can make the determination, especially if the toddler is crawling.
* Seeing well in the rain.
This can be done using near-IR cameras, in fact they can see better than you probably could.
* Telling the difference between a dishwasher carton (which might not have to be braked for, versus a fallen cubic meter of rock.
No, you break for both. If ANYTHING substantial in size is in the road, you break, I know I wouldn't go blasting through a big cardboard box.
* Telling the difference between a solid object and just a splash of water.
This one is easy using two cameras to get a 3d picture, or if you want to get fancy you could use LIDAR.
Keep in mind, this isn't the DARPA challenge. For the most part, roads are known quantities, and a good path following and obstacle avoidance package combined with an accurate GPS system would make this entirely feasable.
My Other Computer Is A Data General Nova III.
Well, there is no need for traffic lights if all cars are robots. Of-course the humans have to cross, but this can be done with overpasses or tunnels under the streets.
Once cars can really drive themselves, they should be in contact with other cars, road signs and such to maintain the best traffic conditions possible.
There will be no real reason for stop signs, traffic lights, speed limits, yield signs and such, all of this can be avoided once cars are driving themselves.
Of-course this requires an overhaul of the infrastructure and assumes all cars are driverless and communicating with each other, the road signs and such and that there are no others (pedestrians/animals/other obstacles) on the roads.
You can't handle the truth.
I agree. The advent of automated transportation is an oppostunity for big cities to ban all private vehicles altogether. Cities should purchase a whole fleet of self-driving vehicles, depending on their needs, and park them on the streets. City dwellers and visitors would then be given a wireless, GPS-enabled beeper to summons a vehicle when needed, at which time the nearest free vehicle would drive itself to the customer's location and take them to their destination. Car-pooling could be enforced, if necessary. There are already way too many cars as it is. Most of them are sitting idle at any given time. So, in that sense, self-driving vehicles may not be a good thing for the car manufacturers because it may lead to a drastic reduction in demand if a lot of congested cities adopt this plan.
As a motorcycle owner, I'd trust a computer controlled cage much more than a human controlled one.
Humans make a lot of mistakes including the stupid excuses "I didn't see you". With computer controlled stuff, the software will *see* everything down to a given size all the time. It doesn't get distracted or starts the 'stare into oblivions', both of which result in the same scenario.
Cars and SUVs and trucks are the largest obstacle to safety for cyclists, pedestrians and motorcyclists. Remove the recklessness from the former and the latter can only screw themselves by their own mistakes not because the cage driver just doesn't give a fsck about you because in a collision, you lose, not the cage.
I'd assume the tech works under ideal driving conditions, but what about the unforseen stuff that comes up in day to day driving? Can the car detect an icy patch on the upcoming corner? Will it swerve violently to avoid an empty cardboard box blowing across the highway? Will it still work around radar jammers? Technically illegal, but I'm willing to bet there's still a fair number of them out there. How stable is the processing unit?
Personally, I'll believe it when I see it, and even then I think I'll be holding on to my steering wheel for the forseeable future.
Murphey's fighting Occam, and we're in the stands.
It ends the problem of drunk driving only if the automated system is 100% in control portal-to-portal and 100% reliable portal-to-portal. The drunk will be making the initial decision to take to the roads.
Yes, which is why the problem is greatly curtailed -- drunk merely has to be able to say, "God damn I'm wasted. Car, take me home." I'm willing to bet that most drunk people who get in a car will gladly take this option in preference to risking a DUI.
You are correct, of course, that part of the problem is simply not giving a damn how your driving affects others. To the extent that that is the problem, I agree that technology and laws are largely ineffective. For example, when someone tries to use a cell phone while making a difficult driving maneuver, the problem is not that he overestimates his skill, or that he deems the potential punishment small enough. It's that he's just not putting in the effort to be a safe driver, and no law or technology will change that.
Apology to Ubuntu forum.
In Soviet Russia, CAR drives YOU! Wait a sec...
Surely, the best way to do it is to get you to the highway, at which point you join a lane (similar to the US carpool lane) that has a weak AM transmitter down the middle. Your car has a couple of sensors on the bottom, to make sure it stays in the middle, and just accelerates to it's highest economy setting.
Better yet, it could slip behind another cars slipstream and take the energy savings for granted. Half-second gaps between cars, with sensors in front and narrowbeam transmitters on the back to alert for stationary vehicles up ahead. Modulate that AM transmitter, and you've got yourself traffic information to plot a better route, and could be encrypted to prevent mis-use.
Why hasn't this been done? And if it has (even if a different system than AMRF) why hasn't it been implemented for economy long-distance driving?
I don't think that better fuel economy is neccesarily a given - cars currently support a very minimal set of electrics and on many cars there is no surplus of power - a lot of small european cars for example noticeably dim lights when electric windows are engaged.
There also comes the issue of redundancy which is not currently an issue, as well as the increased weight if the control components and sensors.
A thistle is a fat salad for an ass's mouth...
Driverless cars by 2018. Like there aren't enough of those out there.
Slightly better fuel efficiency by 2020, only 30 years after it was first proposed.
Government actually requiring that cars not be totally dependant upon gasoline, which would be practical? Crystal ball can't see that far ahead.
Laughter is the Spackle of the Soul.
Come on, what's holding them back? There are already warehouse trucks that drive automatically, avoiding obsticles, making cars do it is only a public awareness campaign away. With Galileo completed in 2013, I'd say thats a perfect year for this too. Maybe not in city streets, but on highways, why not? The technology is already there, we've see articles for years showing how it would work, and we already have cars with parts of these systems.
Just make the system, prove in some spectacular way how safe it is, and then sell it. I can't wait to see the movies where cars are pushed to their absolute limit to avoid a whole series of accidents, all of it happening too fast for a human to register.
Obviously, most countries will initially make sure this is illegal, but there will always be some small country that goes the other way just to be different, and the first manufacturer of these cars for that country will get some major publicity.
There will not be a mass-produced flying car though. That simply requires too much energy and we have a large enough energy problem as it is. Unless you want to use a derigible there is drag induced just by the act of flying which causes an additional amount of energy to be consumed as opposed to staying on the ground.
You don't read enough. There already exists a perfectly good sky car . 20mpg on pure, clean-burning ethanol, and completely safe computerized navigation and flight control. And it's quiet. And it goes well over 200mph, and can take off and land vertically, right in your yard. I can't believe you don't know about this vehicle. It's even red!
Don't disappoint your bird dog. Go to the range.
All things considered, if the tech works as well as GM is hoping for, then accidents would be
far, FAR fewer than what we see today. Owning a self-drive capable car might even LOWER your
insurance payments as you're taking the human out of the equation. Think about what causes most
accidents. Hardware failure ? Um. . no.
Usually it's stupidity on the drivers part. Driving too fast, ( or too slow in the wrong lane )
didn't see the vehicle next to them, drunk, racing, rubbernecking, on the phone, whatever.
Remove the driver from the equation and 99% of the traffic fatalities will probably go away.
Once the tech arrives, it would probably take 5-10 years to get the changeover completed. Once
that happens, most of the accidents and the reasons behind them would vanish. Talk all you wish on
your phone. Eat your breakfast and rubberneck till you are blue in the face. The computer won't run
the light, blow the stopsign or try to race the idiot next to you. Freeway traffic will likely be
self-drive ONLY.
Hell, they may even RAISE the speed limits. The ones we have now have to factor in the idiot
equation. Remove the human problem and higher speeds navigated via computer will be just as safe
( if not safer ) than the lower ones driven by their flesh and blood counterparts.
the lower ones.
With my drive testing the limits of my sanity on a daily basis ( ~80 miles roundtrip to the office
through the worst traffic Houston has to offer ) I'll be first in line if / when this tech becomes
available.
I guarantee the number one application in cities will be telling your car to circle so you don't have to find parking. Traffic will get way worse, then this will be outlawed.
Honestly, I would expect to see self driving semi-trucks before anything else. Shipping companies have an incentive to cut out their drivers in favor of time maximizing, gas mileage aware, AI workhorses.
Actually, a driverless car might be just in time for a lot of the Baby Boomer generation as they get into their 60's and 70's, just as their mounting health problems take their toll on driving skills. Accident rates tend to be lowest for drivers in their 40's and 50's, when mature judgment backed by decades of experience more than compensate for slower reaction times and loss of motor skills they had in their youth. Accident rates start rising again as people get on into their 60's and the effects of decreased vision, the onset of heart disease, diabetes, and the host of other problems which affect the elderly start to take their toll. Some people are good drivers well into their 80's, but others may lose their ability to drive safely much earlier as a result of strokes, heart disease, sleep problems, and the general dulling of the senses and slowdown of mental functions.
Often the elderly gradually retreat from driving, but it remains a vital lifeline to their independence. They shy away from driving at night or in bad weather, or in situations where traffic is heavy and complex, but they treasure their ability to get to the store, the pharmacy, or the local bridge game without having to depend on others. For many it is the difference between being able to remain independent or having to move into "assisted living" or back in with their kids. Having to give up the car keys ranks right up there with losing a spouse or having a major setback in their health for many.
Even a partially driverless system would be a boon to those elderly who, while they might not have the ability or stamina to handle a 6 hour drive from say Philly to Pittsburgh to visit the grandkids, could handle the drive from their home in say Levittown to the PA Turnpike (only a few miles), put it on autopilot and relax for 6 hours, then go back on manual when they near their destination, perhaps aided by a GPS navigation system. Heck, I would even like to have the ability to put it on autopilot for a while even for my 75 mile commute each day!
As an aside, I've seen in my own family that the elderly have widely varying attitudes towards tech. My 78 year old Dad has always embraced tech, he's said that the ability to utilize modern technology gave him an edge in business and in life, he could fix just about anything, and often still can. He owned and knew how to use an IBM XT and had a cellphone way back in the '80s, was on the internet with Compuserve in the 1992, 4 years before I went online. He grew up a farm boy, served in Korea in Ordinance, where he learned about computers and electronics, then went on to a successful career in the diesel engine business. On the other hand, I have an 82 year old aunt that refused to get a microwave oven until one of her kids brought her one a couple of years ago, does not own a computer, and still had a rotary dial phone until the phone company made her replace it recently. She isn't destitute or uneducated, she is a retired teacher living on a pension, and owns her home outright.
TFA mentions things such as how to deal with road obstacles and other problems of real-world driving, with one of the solutions being to alert the driver. Great. The car will drive itself, but someone still has to sit behind the wheel and be alert enough to deal with situations. Heck, you probably couldn't even take your hands off the wheel or feet away from the pedals, because by the time you react and get them back where they need to be, it will be too late.
Self-driving car technology will be ready for prime time only when it's sufficiently advanced that I can get in the back seat, tell the car where to go, and take a nap while it gets me there. And yes, that takes into account needing to be so relaxed and confident in the entire process that I actually *can* fall asleep in the back of a car that's driving itself.
If that happens in my lifetime, I expect it to be about the time when I'm too old to safely drive myself, which will be a good ways past 2018.
...expect them from Toyota in five.
"When information is power, privacy is freedom" - Jah-Wren Ryel