Slashdot Mirror
Would You Need a License To Drive a Self-Driving Car?
agent elevator writes Not as strange a question as it seems, writes Mark Harris at IEEE Spectrum : "Self-driving cars promise a future where you can watch television, sip cocktails, or snooze all the way home. But what happens when something goes wrong? Today's drivers have not been taught how to cope with runaway acceleration, unexpected braking, or a car that wants to steer into a wall." The California DMV is considering something that would be similar to requirements for robocar test-driver training." Hallie Siegel points out this article arguing that we need to be careful about how many rules we make for self-driving cars before they become common. Governments and lawmakers across the world are debating how to best regulate autonomous cars, both for testing, and for operation. Robocar expert Brad Templeton argues that that there is a danger that regulations might be drafted long before the shape of the first commercial deployments of the technology take place.
If "yes," then it's not self-driving.
Car drive you!
A self driving car should be able to be aware of it's surrounding, of the other cars around it, and if anything goes wrong, either set itself to the side of the road automatically or just Stop, while the others cars around that are also self driving, slow down and avoid the broken down car. Never requiring any action from the dumb*ss behind the wheel.
There will be detractors, luddites, and evangelists, sociopaths and attention whores all vying for a moment in the sun.
Welcome to the human race. I'll go get my popcorn.
Silence is a state of mime.
For any and all events that happen while that car is being driven, legally and otherwise. Better hope you can take over when the computer says "ohh, can't handle this situation" and hands off control to you.
Do pilots still need licenses in the age of autopilot? Well yes because machines aren't infallible.
For a long time, an autonomous car will not be driverless. People need to get over this notion that next year a car will drive itself and you'll sit in the back with a Martini and the paper. That probably wont happen in our lifetimes.
Initially, fully autonomous modes will only be permitted on certain roads (think limited access roads like highways, freeways and autobahns). This will last years as engineers are even more conservative than law makers. The next step is likely to be special lanes on A roads. It will be a long time before autonomous cars are good enough to operate on a B road or suburban street.
Ultimately, because the law requires someone to be responsible for the operation of the machine it means a qualified operator will need to be at the controls whilst in operation. Same with a lot of other automated systems (such as long distance trains).
Calling someone a "hater" only means you can not rationally rebut their argument.
Should you need a license to operate a computer ?
well, considering facebook , the answer is YES
if there is any manual control whatsoever then yes absolutely license should be required if it has no user overide for manual control like an automated shuttle then no one shouldn't be needed but then if you can't manualy control it do you really own it?
The concept of a full autonomous driving car is still really far way, I estimate in the range of 20-30 years or even more. My understanding is that current autonomous vehicles don't cope with rainy or unknown conditions. The more interesting aspect would be how augmentative safety technologies might reduce insurance premiums, or even reduce certain kinds of legal requirements. For example if your vehicle has auto assisted braking could that allow some one to drive at a different alcohol limit range? It might be that autonomous vehicles at first must be licensed to travel on specific routes. So you would have to register your autonomous vehicle only on specific routes you have 'taught' it first. Only then you could allow it to be 'autonomous' and be able to pay no attention while being transported ( i.e. you don't need a license to be in the vehicle).
Should all car drivers be accomplished horse riders? Well yes obviously! You never know when your car will break down, run out of gas, etc. and you'll need to hitch up a horse to get you home.
I think that it's pretty clear that within a few 10s of years the car with a driver will be the anomaly. The economic advantage in large areas of transportation (trucking, taxis, deliver, etc. etc.) are so huge that the technology will be adopted, and the transition to home vehicles is inevitable because the cost is minimal and the advantages great.
These discussions will look really stupid, probably before mid-century.
And I stopped listening right there.
Only fucking MORONS want this sort of thing.
When you're in a piece of heavy machinery, like a car, even if you're NOT driving it, you DON'T want to be impaired in case of an emergency.
So, drinking in a self-driving car is pretty much out. And for many of the reasons this dipshit talked about. MALFUNCTIONS.
Before you bring up bus and rail transport. Keep in mind, there are people actually driving those. And, in the case of long distance trains, crews full of people. All better trained at running the transportation than you are.
Chas - The one, the only.
THANK GOD!!!
This also makes me think, will you need insurance for a self-driving car? If two self-driving cars are involved in a collision, who is responsible for the damages? You could say the manufacturer is responsible - but what if it's a collision between a self-driving car and a human-driven car? Or, will manufacturers be willing to take on the burden of providing insurance for each car they sell?
So let me get this right. You are in a 'driverless car'. Yet your job is to painstakenly hover over the controls trying to double guess the AI every single second you are in the vehicle?!???!! Good luck with that because if the AI fucks up you have a second or two tops to stop yourself from becoming road paste. It sounds like a massive copout from manufacturers wanting to sell autonomous features before the technology is mature enough to realistically insure and assure customer safety.
or interchanges. If the cars are well guided and coordinated you could have full speed ground level crossing where the cars just space out enough to weave past each other. Would be terrifying at first but people would get used to it.
I love Jesus, except for his foreign policy.
They do not exist, and they are unlikely to exist for the foreseeable future. All of the demos:
1. Rely on highly process intensive data capture and analysis,
2. Do not handle changing road conditions
3. Do not handle road hazards.
4. Do not handle emergency signals.
An automated car that I have to be attentive to 100% of the time is worthless. A automated car that I have to be attentive to 10% of the time is an accident waiting to happen. The expressway equivalent to a rail line may be possible (carefully controlled, monitored externally, special lanes etc...), cars driving thru city streets and neighborhoods is a pipe dream.
Say what you want about automated pilots on airplanes, but a medium size city has more cars on the road at one time than the entire world has planes in the air.
[A curmudgeon tired of marketing BS]
See subject.
Why don't we just incorporate them?
Then they'll legally be people, and they can get their own driver's licenses!
When the automobile debuted, the UK passed the infamous Locomotive Acts (otherwise known as the Red Flag Law), requiring someone to walk in front of a "horseless carriage" waving a red flag.
Requiring a license for a self-driving car is the modern red flag to avoid spooking the lawyers.
Frankly, I like driving when I'm not in the city, it's relaxing, it can be fun, controlling the vehicle is great.
The only time I'd ever want the car to drive for me is in high density situations where most people can't drive for shit, thus I find it too stressful - in those cases, autonomous driving is perfect.
Outside of that scenario, if I couldn't drive my car myself, I simply wouldn't ever buy one.
But it win;t be called a ;drivers licence , it will be called a government issued photo ID
You need one now to cash a check, vote etc.
(I would have used the term "state issued ID" but SCOTUS still has to decide whate the term 'state' means.
Note to grammar nazis - would have is pronounced would of
The state needs the ability to track the movement of the populace. If they had unlicensed cars on the road, people would be able to move about freely, without being surveiled. Imagine the safety implications there...
Also, they would lose an avenue for much needed recurring revenue, and something to hold over the head of criminals.
It's almost like you think you live in a free country?
Why does a free citizen of a free country need government's permission to drive on public roads to begin with?
It is a right, which the Judiciary might take away from the bad — upon successful prosecution by the Executive. Not a privilege to be granted (or not) by the Executive government themselves...
Brace yourself before attempting a rebuttal — make sure, whatever you state in support of the need for these licenses, would not also apply to a permission to walk the public sidewalk...
In Soviet Washington the swamp drains you.
Until self driving cars can be held individually responbile for mishaps humans must be ultimately responsible for whatever damage they cause. I can forsee a future where self driving cars don't yield to pedestriasns walking againjst a lightl. A human would respect life and stop simply because a life is a life and humas respect life above all else., a self driving car would insist it was their right to run the pedastrian down simply because that was the rules of the road.
Time is what keeps everything from happening all at once.
Considering driver licenses are de-facto identification schemes: yes.
You should know better than to make false assertions when we have plenty of evidence countering your assertion that technology will ever be this good. Since the 1960s we have been automating space travel and airlines, and still need pilots and astronauts because when the shit hits the proverbial fan humans are required to intervene. Sometimes to correct problems with the technology, and sometimes to bypass it and fly by hand.
Drones require people to pilot them too, so don't try to go down a bad path.
I don't see this as a problem of litigation, I see it as the only sensible approach to having technology. Nothing, and that is an absolutely nothing, has ever been made by man which has been perfect. We try for "the best we can" but stuff breaks and the unexpected does occur. With an estimated 250,000,000 cars on the road chances are high that something will go wrong pretty damn fast. With motor vehicles already being the number one killer in the US annually, we want human intervention early and often. That means trained drivers behind the wheel.
As stated above, a half a century has not perfected "self driving" anything else. It's much better today than 50plus years ago but not even close to the point where you can fly without a human.
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
When the automobile debuted, the UK passed the infamous Locomotive Acts (otherwise known as the Red Flag Law), requiring someone to walk in front of a "horseless carriage" waving a red flag.
The first Locomotive Acts were passed in the 1860s.
Forget the "horseless carriage." We are talking about road trains, huge and heavyweight steam powered agricultural tractors, bailers, threshers, bulldozers, steam shovels and the like.
To this day flagmen and escort vehicles serve the same purpose.
I didn't bother to read TFS.
Yes. You do. Ask if the author (or any sane person on this planet for that matter) would fly in a plane that didn't have a pilot.
The difference with automated trains and trolleys is that they are on tracks.
You would need a 'driver' for a driver-less car as much as you need a horse for a horse-less carriage.
I think the key to making cars that are really 'self driving' will be to have the on-board systems backstopped by a call center rather than anyone sitting in the vehicle itself. Autonomous aircraft are really designed with a computer to handle the routine flying and then pass things off to a remote pilot for the interesting bits. An autonomous car could handle the freeway and major streets by itself quite well but might need to call up a licensed operator to negotiate a parking garage or a work zone.
Something I can see happening to lead to this will be commercial trucks that are self driving, and unmanned, on the freeway but that pull into special truck stops where a pool of local drivers are available to get the truck the last few miles to its destination.
Lets say my self driving car runs someone over... who is liable?
I've decided to stop wasting my time responding to AC trolls/sockpuppets... so if you want a response from me... login.
I think a more interesting question is will gaurdians be allowed to put children in cars alone?
( Jimmy, your parents just called and said they would be late, I'm going to call yoiu a robotaxi. )
If there a manual control, Then yes. Is it a sealed box/no steering wheel, then no.
See planes for current example. Hell many are able to take off land on thier own now,
lag and bandwidth issues make remote control a poor choice. Also data roaming costs for live video can very fast hit the cost of a new car.
Do pilots still need licenses in the age of autopilot? Well yes because machines aren't infallible.
This is a terrible analogy. First autopilot for a plane cannot taxi the aircraft so it is not feature complete. Secondly the consequences of mechanical failure in a car are far less severe and you can probably solve most of the ones which do not themselves involve the engine dying by having a kill switch and a steering wheel: all you have to do is yank the switch and steer the now rapidly braking car out of trouble. A kill switch on an aircraft is a somewhat less viable option which is why you need a pilot. This is also why commercial pilots have far more training than bus drivers.
hardware sensors reliability is an issue as well look at air france 447
Why don't we just incorporate them?
Just what we need a bunch of robot cars going around telling us that "You will be incorporated". Will they come with a red laser pointer strapped to the roof too?
car makers need to have patches for at least 5 years for free and no you must go the dealer from them.
let's see how bad can things get then they say after 1 year no more updates go buy a new car.
Trained drivers may be what we want, but we'll have autonomous cars first.
Since the 1960s we have been automating space travel and airlines, and still need pilots and astronauts because when the shit hits the proverbial fan humans are required to intervene.
We have pilots to make passengers feel good. We have astronauts because we can't make a robot as dextrous as a human yet.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Nothing, and that is an absolutely nothing, has ever been made by man which has been perfect.
A self-driving car does not have to be perfect. It just has to be better than the alternative.
With motor vehicles already being the number one killer in the US annually, we want human intervention early and often.
Isn't the fact that motor vehicles are already the number one killer in the US annually actually an argument for automated cars?
As stated above, a half a century has not perfected "self driving" anything else.
Five centuries of work before that never perfected heavier-than-air flying machines either, until one year, presto, all the necessary preconditions were finally met and airplanes became a reality. There's nothing linear about progress.
I don't care if it's 90,000 hectares. That lake was not my doing.
Figuring out who will give them the most money now, who will give them the most money in the future, and figuring out how to maximize that income.
/cynical?
// not me
/// nosirree bob
//// Re: Jeb Bush and Terry Shiavo
Really, if someone has some scenario of the car malfunctioning, and a hero driver catching it before it hits a wall or goes over a cliff, they have another think coming. Nobody would be that quick. These self-driving cars are going to either be good enough that you can get in and snooze all the way to work, or they will be worthless.
The thing that will really stop SDC's from happening are the laws. The gov't isn't much going to like giving up its highway robbery known as speeding tickets so will not alter the speed limit. The SDC will have to be programmed for the speed limit, while "regular" cars will go flying by probably 15 - 20 mph faster. SDC occupants will not only be in grave danger from getting hit from behind, but will be unhappy at taking far longer to get anywhere than the lawbreaking "regular" drivers that are supplying the state will all the ticket revenues.
SDC's will work most everywhere else in the world except the USA.
I'll just use the same license I will have for my fucking starship. Because seriously, if the mathematics is discovered to make hyper advanced software necessary to make self driving cars even REMOTELY feasible, it will also allow breakthrough physics enabling faster than light travel. I mean shit, can slashsuck stop with the hipster startup paper quadrillionaire stock fluffery for just five seconds???
The point of having pilots in modern airliners is precisely to intervene when automation fails. This happens. There are procedures for it.
My biggest issue with them is I can't get people to tell me how they work. If you are coming up to a blind corner, and the "safe" speed (the speed at which you could stop if there was a hidden brick wall at the point of least visibility) is 20 mph, yet the average driver takes the corner at 55 mph, and the car can physically take the corner at 80 mph, would you, shoud you program the car to go 20 mph, or 80 mph, or some other speed?
The problems with the discussion are that we don't know what we want it to do, not that we are worried it won't do it. Can the "driver" tune the car to "most safe" or "best time" or "average traffic" modes?
Whether it works is not up for discussion until someone can answer what "works" is.
As stated above, a half a century has not perfected "self driving" anything else. It's much better today than 50plus years ago but not even close to the point where you can fly without a human.
That's a legal, not technical issue. We have self-driving drones. But we won't trust a self-driving drone with flying a human around. That's the difference. We have 100% automatic driving (cars and planes), but don't use them for legal, not technical, reasons.
Learn to love Alaska
What can't driver's training, especially for robo cars, be an app you run on your phone. Have it test you for a short period of time once or twice a year to renew your license. Easy. You don't need a full on simulation of real driving to have a valid test, you can get away with training that is interactive and similar enough to the reaction and skills necessary. Multiple choice questions don't count as interactive or similar in skill as those kinds of written tests are usually pretty easy to deduce without any study or practice.
For every anecdote of a human taking over and saving the day, you can find a similar one of the human taking over and crashing. It mostly boils down to the amount of training that the pilot has had - and even the ones that end up crashing in situations where the automatic systems would probably have managed have had vastly more training than almost any driver on the road...
I am TheRaven on Soylent News
The point of a self-driving car is that you don't drive it. It drives you (there's a soviet Russia joke in there, somewhere) Do you need a license to get into a taxi?
There's a whole raft of issues with getting autonomous cars to work how we want them, but I don't think any of the problems are insurmountable. With regards to a blind corner, it would be neat if the first vehicles taking that corner would be cautious (20mph), but as they uploaded information about that particular corner to some kind of driving knowledge base, subsequent cars would be able to take that same corner quicker. Also, if the cars could share sensor information, then it would be possible for a car to "see" round the blind corner if there was another car already round there.
You're a temporary arrangement of matter sliding towards oblivion in a cold, uncaring universe
There is no way on Earth that the state governments will let those annual license fees flip through their fingers... there are trillions of dollars in unfunded unionized state worker pensions to be funded and vote-buying pork-barrel projects to be funded.
The politicians are also always looking for more excuses to claim that a free people are not actually free to do various things - without the consent of their rulers and certification that they are competent (and by default, government presumes itself competent of everything and the population incompetent on everything. People who like a government that presumes they are stupid children, deserve every bit of disrespect and abuse they get from that government.
We have trained operators in cars. What we need is a higher percentage of competent operators.
Well, I might have a way, but it only works on a semi spherical planet in a vacuum.
Maybe in 20-30 years. Before then, "self-driving" will be no different to cruise control or auto parallel parking.
Then, the DMV. No traffic violations, car chases, or accidents. Why would anybody be required to carry a license or other ID except as a method of social control.
It is trivial to envisage situations that occur every single day during a commute that would baffle a self drive vehicle and would cause it to want to hand control back to to a human.
The closest that we are likely to come to driverless cars are those operating on closed loops, e.g. between airport terminals where the road layout and the number of parameters is manageable. Even then there is probably some guy sat in a booth somewhere who can take over the controls if the car does something dumb or gets confused.
Auto-driving trucks might be even easier to implement than cars, since they can be guided by wires embedded in the pavement and their movement limited to interstates, (automated) gas stations, and freight terminals.
The first auto-driving cars will have their paths blocked by mobs of out-of-work truck drivers.
The story of the AF447 crash is precisely that: the human took over, and crashed.
What I wanted to show by bringing up this example is that in current airplane design, there are circumstances in which automation is known to fail (in this case, unreliable/defective sensors). In these circumstances, the systems are designed to give control back to the pilot. The rationale for this is quite clear. It could be argued that fully working automated systems are safer and more reliable than humans. However, automated systems with detected failures are not.
So the pilot is not there to make passengers feel better: he is a part of the automation backup system. Of course, sometimes this backup does not work: no system is perfect.
For automated cars, the situation is a bit different. As you pointed out, drivers are not trained for such contingencies. And if a problem happens, the car can just stop on the side of the road, while the plane does not have this option.
Sure. It will work similarly to the way it does today. In order to have a valid vehicle registration you will need to provide proof of insurance on the car. In the event of a collision or other incident which causes damage, law enforecement will record information about the incident, lawyers will argue fault, and a judge will determine the outcome. In the case of an autonomous incident, the carmaker will almost certainly be listed as a defendant, and will have a staff lawyer as part of the defense team. Or, as more likely happens, the lawyers for the insurance companies will get together and decide the outcome out of court.
Car makers already have insurance for legal problems (and/or are self-insured). Every manufacturer does. It's part of the modern landscape.
Is it just my observation, or are there way too many stupid people in the world?
Actually, this would be a problem. The USAF is currently struggling with some of this - they automated their drones too much, operators don't have enough to do to keep proper attention on the drone in case something does happen. They're actually considering removing some of the automation...
I don't disagree that this is the most likely current situation, but it's going to be virtually impossible to keep the driver from doing other things as you remove more responsibility and control from them.
I don't read AC A human right
I like driving too, but I only get to drive for pleasure a handful of times during the year. I often have to sit for extended periods - sometime stretching to hours - unable to do anything else productive (or entertaining) while travelling between destinations.
To be honest, I don't own a car which is a pleasure to drive. I own a minivan for transportation of equipment and people, and a truck for hauling things and for when my van breaks down. If could drop in a self-driving option in the van, I'd do it yesterday - there is no pleasure in that vehicle. Maybe I could even get more work done while on the road and make enough to buy a car I could drive for pleasure.
Is it just my observation, or are there way too many stupid people in the world?
Have you never heard of faulty brakes, faulty accelerators, faulty [insert car part here] which have resulted in human deaths, and both human and property injuries?
Those things don't change with a driverless car. Liability still exists. And just as an "accidental" fatal impact with a human by a human driver becomes primarily a financial burden (usu. involuntary manslaughter, suspended or limited sentence for otherwise "good" person), the financial implications to a automotive manufacturer are significant.
When a driver hits someone, the car is reviewed in excruciating detail to determine if any failure of the automotive systems caused or contributed to an accident. The driver is also scrutinized. The victim is also scrutinized and if they have a car their car is evaluated for faults. A driverless car removes the ambiguity of the driver from the fault path.
Plus, as the sibling poster points out, the systems in a driverless car are going to be engineered with additional fail-safe mechanisms which reduce the overall performance/utility in favor of safety in the case you posit.
Is it just my observation, or are there way too many stupid people in the world?
I think there will still be requirements for a license but it will be about the state apparatus' interest in controlling movement of citizens and not about driving the car.
At first, it will still be about "driving" as I think that the transition to self-driving cars will be somewhat gradual. I don't think it will be the case that this year's model is manual and next year's is fully automated. Automation will be phased in where the car can handle more and more routine driving situations until eventually no driver control will be required, and during this transition it will still be possible (and necessary) to actually drive the car.
Even the first fully automated cars will probably allow some kind of user overrides as to where the car goes, how fast, etc, so you will still need to have a driver responsible.
But after cars become fully automated, it won't be about "driving" anymore, it will be about the state's interest in controlling who can go where and when.
Some more:
1.1 Explorer program (1958â")
1.2 Pioneer program (1958â"1978)
1.3 Echo Project (1960â"1964)
1.4 Ranger program (1961â"1965)
1.5 Telstar (1962â"1963, commercial project with NASA contribution)
1.6 Mariner program (1963â"1973)
1.7 Lunar Orbiter program (1966-1967)
1.8 Surveyor program (1966â"1968)
1.9 Helios probes (1974â"1976)
1.10 Viking program (1975)
1.11 Voyager program (1977)
1.12 High Energy Astronomy Observatory 1 (1977)
1.13 Solar Maximum Mission (1980)
1.14 Infrared Astronomical Satellite, IRAS (1983)
1.15 Magellan probe (1989)
1.16 Galileo probe (1989)
1.17 Hubble Space Telescope (1990)
1.18 Ulysses (1990)
1.19 Upper Atmosphere Research Satellite, UARS (1991)
1.20 Discovery Program (1992â"2011)
1.21 Clementine (1994)
1.22 Mars Global Surveyor (1996)
1.23 Cassiniâ"Huygens (1997)
1.24 New Millennium Program (1998â"2006)
1.25 Gravity Recovery and Climate Experiment (2002)
1.26 Earth Observing System (1997â"2011)
1.27 Mars Exploration Rovers (2003)
1.28 MESSENGER (2004)
1.29 New Frontiers program (2006â"2011)
1.30 Mars Scout Program (2007â"2008)
1.31 Dawn (2007)
1.32 Lunar Reconnaissance Orbiter (2009)
1.33 Mars Science Laboratory (2011)
Waterfox - a Firefox fork with legacy extension support, security updates and better privacy by default.
A self-driving car does not have a human driver. The question that should be raised (to show how flawed the article is):
Today's passengers have not been taught how to cope with runaway acceleration, unexpected braking, or a car that wants to steer into a wall.
Help! I'm a slashdot refugee.
People seem to be conflating licenses and training, which is cute considering y'all just got done beating up on tech certifications in another thread. Should you need government permission to travel in your automated car? Hell no. Should you know what you are doing when climbing into one? Ideally, yes.
quite a few those had human initiated burns to make their their trips, so how driverless is that?
world was created 5 seconds before this post as it is.
Solvable by having reduntant sensors, of different types. A plane can measure air speed and also get gps speed, for example. If the difference is bigger than realistic wind speed, sound the alarm.
So which is it ... do we legislate GMO foods to make sure they are 100% safe before they are ever sold, or not legislate self-driving cars until crashes happen and we know what to do to make them safe.
Ah ,,, the hypocrisy and FUD in it all. The GMO group wants to outlaw things they don't want and deny them from everyone because they know what is best, while the other group doesn't know enough yet and wants to make sure no one stops them from getting what they want.
I sure as hell don't want the first batch to be driven without someone behind the wheel for a few years. And that person needs a regular driver's license and the ability to take over if anything either fails or the car can't cope with a situation.
As someone who just drove over 2,000 miles, some of it in snow, I would love to have some of the tech that self-driving cars are making available. But 100% door-to-door service with 100% accuracy?? I just don't see it happening anytime soon. The GPS I used sent me the wrong way twice, and I just updated the maps before I left. Once was not a big deal, it just picked a route that wasn't as efficient. The second failure took me to a non-existent gas station which appeared on the map to be in the middle of a corn field.
We do need regulations. Or do you really want to see a bunch of people killed by an automated car, and the heavy-and of the law come down then.
There is no overriding reason to rush these to market, the percentage of the population that really needs them is very small. For the rest of us, they are just a convenience factor. It's possible they could lower traffic accident rates. Or it's possible that they could increase them since in the beginning, only a very small percentage are going to be automated, so they will have to deal with the millions of bad drivers out there.
I do have one prediction though .. self driving cars are not going to be the market share everyone thinks. Why?? I can almost bet that very few drivers/riders will tolerate a car that follows all the traffic laws, such as not speeding and coming to a complete stop at every stop sign.
I rarely read replies, it's my opinion and if you thought about your opinion a little more, I'm OK with that.
Until an auto is good enough to go curb to curb on its own, you still need to have an operators skills.
Five centuries of work before that never perfected heavier-than-air flying machines either, until one year, presto, all the necessary preconditions were finally met and airplanes became a reality. There's nothing linear about progress.
The most intelligent comment I have read on Slashdot in years (possibly ever), if only I had any mod points.
"He is so stupid. And now back to the wall!" Moe Szyslak
What I wanted to show by bringing up this example is that in current airplane design, there are circumstances in which automation is known to fail (in this case, unreliable/defective sensors). In these circumstances, the systems are designed to give control back to the pilot. The rationale for this is quite clear.
Yes, like I said, it's to make the passengers feel good. Because as we have seen, the pilots depend on the same sensors that the autopilot does. Airliners aren't fighters, you don't fly by the seat of your pants. By the time your inner-ear-gyro tells you that there's a problem, you're already screwed. Which was precisely what happened.
How in the shit are pitot tubes still icing anyway? Why is heating the tube not a thing which works? Heating elements are not new technology. We should really be able to manage this by now.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Why would anyone jump into the path of a self driving car?
License to own, license to operate, license to park...
Because money.
RIP 1992-2000.
.
My current car is 12 years old and has an accelerator pedal that is basically a potentiometer fed into the engine control system.
But the steering and brakes DON'T work like that.
And to my knowledge don't on any car today.
If I was to design a car that was only EVER to be driven by computer there'd be need to physically extend the steering hardware back into the cab. I might corner mostly by varying relative speeds of wheels except in low speed manoeuvres.
If I thought occasionally it might voluntarily ask for driver assistance in driving conditions it couldn't negotiate, then I'd add simple drive by wire input devices. (Steering wheel and pedals as used in gamin industry, perhaps).
But if the driver was to intervene when the car's computer had totally lost the plot I'd want an always connected hardware override. A real steering wheel that could be seen turning when the car steered itself.
Which of these would you legislate for ?
Drones require people to pilot them too, so don't try to go down a bad path.
How about people driving remotely, then? Someone who wants to play Car Driving Simulator can drive while I get to sleep/knit/watch tv for the whole trip.
Do you really have to ask this question? Do you really think that any government is going to pass up the chance to tax its citizens?
No, we don't. We have pilots with training who can take over in an emergency. Landing in the Hudson is probably the most famous case, but plenty of similar less dramatic events have occurred.
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
Motor vehicles are FAR from the number one killer. 10 times as many people are killed by cigarettes - the butcher's bill is split between that and fat people having heart attacks and poorly managed diabetes.
Motor vehicles aren't even the number one killer for a particular age group - they are edged out by such things as overdoses and suicide.
They do beat out homicide in a number of age groups though.
Figuring out the laws of physics required for lift is not the same thing as automating a complex task that limited numbers of humans can perform. I gave the example in automation, and you simply plucked something out of the air to say "nuh uh".
Show me where automating flight has been perfected to the point where we no longer require humans. We have not done so, and that is the measure we need to make.
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
That means trained drivers behind the wheel.
Sadly we have very few trained drivers behind the wheel.
Who the hell buys a car every 5 years? I'm driving a 1994 Dodge Dakota.
Wrong on just about every account. It is not a matter of the alternative, it's a matter of what happens when an incident occurs and how a human is still the fail safe. Under many circumstances I would agree that auto-pilot is better. Long boring drives where the weather is good and the car operates normally being one of those. Where the automatic goodies fail is always when the unexpected occurs. A deer jumps out of the trees in front of the car, road debris too small for sensors or human eyes destroy a tire, weak pavement gives way, a patch of ice on a stretch road, etc.. etc.. We see the exact same thing in flight, and interestingly people here attribute a crashing plane to the human even when the human had to intervene because the plane was crashing despite automatic controls.
One day all of these exceptions can be built into software making the computers reaction better, but we are not there yet. TFA is not talking about having untrained people in cars in a decade, it's equating with current technology. So are you by the way. Arguing that we have all of these things covered in autopilot is provably false. Google cars can't drive today in poor weather, and a blowing paper bag is see as the same thing as a concrete block to sensors.
As I said below you can't compare automating complex issues like this to figuring how the physics for how lift works. It is not the same thing and not the same level of complexity.
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
We need an insurance-based licensing scheme for self driving vehicles, as outlined here.
It's a 2013 model and has a really incredible and awesome laser system for adjusting the cruise speed. I use it every single day in slow or fast traffic. But, it is not perfect. If I get behind trucks that tow a grid trailer (think, lawnmower trailer), then the system can't get a solid lock and speeds up/slows down continually. On occassion I've tried to set the lock when I was a little too close to a vehicle, and it acts like that vehicle is invisible and wouldn't slow down for it automatically. But 97% of the time it's good.
Just leave the insurance requirements the same as they are now: it should remain the responsibility of the owner of the car to insure it.
Here's why this will work. No manufacturer is going to go to market with a product without knowing (and having a lot of data) that proves that it is a better, safer driver than any demographic average that currently exists. Insurance companies will see the wisdom of this, and will give owners of "auto-cars" a break on thier insurance premiums. If your auto-car still manages to cause an accident, you are at fault just as much as if you were driving yourself. Hey, shit will still happen, just not as often.
The vast majority of accidents on the road are caused by stupidity. Decisions made by impatient, exhausted, drunk, inattentive, or angry human beings. These types of accidents simply would not be caused by auto-cars.
Just as today automakers are sued class-action style for egregious faults (regardless of insurance), there will also be lawsuits over auto-car implementations that caused foreseeable and repeatable gross errors, resulting in accidents that normal drivers would never have caused. But this is irrelevant to the discussion about insurance coverage, unless it becomes so contentious that auto-cars are deemed completely uninsurable (very unlikely given the money and research being put into this: like I said above, the statistics will still show that they will be much safer overall than human drivers).
But my opinion is that for accidents that would have happened even with a skilled driver (white-out snow, fog bank, black ice, pedestrian running from behind a vehicle, etc.), the manufacturer of the auto-car should be immune from responsibility. In such scenarios, no court would consider finding today's car manufacturers responsible for these accidents.
So leave things as they are. Your car, your insurance. No need to make things more complicated.
Every single mission you mentioned required massive amounts of manpower to get into flight. Every single mission still requires humans to review data and make adjustments if necessary. I cautioned about using drones as an example, and should have included space missions in that warning. Sadly people can't make distinctions on their own when it may harm their fragile belief system.
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
No, we don't. We have pilots with training who can take over in an emergency. Landing in the Hudson is probably the most famous case, but plenty of similar less dramatic events have occurred.
When the NTSB tested the autopilots in the sims after the crash, they all safely landed the plane back at the airport. All the human pilots put in the same sim either crashed or ditched.
Bird strikes are a known problem. A double-engine loss on takeoff is unluckly and unlikely, but not unfathomable. Autopilots are programmed for that scenario, and they react faster and more accurately than human pilots. It's only in the emergencies we didn't imagine in advance where humans shine. The "Miracle on the Hudson" was not such a case.