Slashdot Mirror
US Regulators Issue Comprehensive Policy On Self-Driving Cars (vox.com)
An anonymous reader quotes a report from Vox: On Monday, [The U.S. Department of Transportation] released a surprisingly far-reaching "Federal Automated Vehicles Policy." The policy attempts to do all sorts of things -- we'll get into the details below -- but the overarching motivation is that DOT wants to accelerate the development and adoption of AVs. DOT views AVs as a safety technology that could reduce some of the 38,000 traffic fatalities a year in the U.S., 95 percent of which are caused by human error. It also sees AVs as an accessibility technology that could provide personal transportation to whole populations (disabled, elderly, etc.) who have lacked it. The policy comes in four buckets: What the vehicles need to do to be safe; What federal and state governments need to do; How DOT will use its existing regulatory tools; DOT may need brand new regulatory tools to deal with AVs. The "vehicle performance" section lays out a 15-point safety assessment, so that AV developers and manufacturers know the sorts of things that federal regulators will expect. It covers everything from cybersecurity to data collection to crash response. And then there are "ethical considerations." AVs will have to make life-or-death decisions. The second section addresses the division of responsibilities and authorities between the federal government and state governments, and suggests a model policy that states can adapt for their own use. The feds will retain their authority to set and enforce safety standards, communicate with the public about safety, and occasionally issue guidances about how to meet national standards. States will retain their authority to license human drivers and register cars, set and enforce traffic laws, and regulate vehicle insurance and liability. There are three broad ways that DOT communicates about standards with automakers: letters of interpretation, exemptions and rule-makings. It is promising to speed up all of them in regard to HAVs. DOT is considering a range of new authorities that may be necessary to properly regulate HAVs. The report adds that "DOT has officially abandoned the NHTSA's own levels-of-automation classification in favor of SAE's, which is preferred by the industry. Vox has neat graphic you can view here. President Obama also wrote a piece about self-driving cars in the Pittsburgh Post-Gazette: "In the seven-and-a-half years of my presidency, self-driving cars have gone from sci-fi fantasy to an emerging reality with the potential to transform the way we live..."
/cdn0.vox-cdn.com/uploads/chorus_asset/file/7129959/Driveless-car-scale.jpg){kind=link}
...to save me and my family from all the "excellent drivers" who are busy on their phones while speeding down the freeway. Some of them, no doubt, posting diatribes about Big Government taking away their right to maintain complete and perfect control over their vehicle's performance.
http://vignette3.wikia.nocooki...
Maybe in the distant future, but definitely not within 5-10 years.
This is one of the things needed to get this technology legal and on the road. And better, one of the government agencies in charge of seeing this happen has outlined a plan to get it to happen. Not just in terms of what the manufactures need to get done on their end to be legal but also an outline of the regulatory apparatus required to ensure the safety of everyone!
I call this a win on just about every level.
It's insane even considering implementing beta technology when people's actual lives are at stake. I weep for my country and its rampant greed.
We had such a socialist revolution. In 1917, in Russia. Ah, Cuba and North Korea had those, too. Do you like the results ?
Religous speak to God. Insane are spoken to by God. When all shut up, one can finally hear Shostakovich in peace
but what is evident from the ThinkStock-like illustration that goes with Barack Obama's article: this is going to greatly favour Uber, a company to be abhorred for its greed, stupidity and lack of ethics.
Religous speak to God. Insane are spoken to by God. When all shut up, one can finally hear Shostakovich in peace
This is the beginning of the end of private vehicle ownership. That is why the corporate controlled totalitarian communist government wants federal control. Central planning at its best.
I can't get an update on my 1 year old phone and now we want to have self-driving cars that have network connectivity... any chance we can get an openness requirement or a supported software lifespan requirement?
Btw, this is might be the right actually doc for this -http://orfe.princeton.edu/~alaink/SmartDrivingCars/Automated_Vehicles_Policy.pdf
How many jobs will automated cars eliminate? Cabs and couriers are my first thought, large trucks provide millions of entry level jobs. I know automation will come one day, but are we just making more poor people.
Being a spelling & grammar Nazi is a sign you do not poses the intelligence to contribute to the conversation
No. It is through proper private ownership of our property and our individual rights that we retain control over our technology. The state is the entity imposing draconian patent/copyright law as well as backing monopolies that prevents this (eg right to hack). In fact, the state is the biggest monopoly of all. They can and do legislate entire companies out of business whenever they encroach on its turf (eg what they're trying to do to uber and lyft). The state is also responsible for the emerging draconian surveillance society, with the relevant legislation often used as justification to further lock down technology.
Who do you think will demand remote access to all of these vehicles? Do you think it will allow owners to remove whatever spy/malware vendors embed? Hell no. The liability nightmare makes private ownership of these cars nearly unaffordable. This gives both the vendors and the state unfettered access and control over freedom of movement. It's a mutually beneficial relationship at our expense.
The wealthy cannot rule without a big and powerful state, as well as legions of twits like you who will happily vote to force your fellow citizens to fund their own oppression.
I will not buy one unless there is NO data broadcast off-vehicle. My 1992 car does not do that. There is no need for a new car to either.
But in this age of "big data" and hugely profitable data brokerages and behavior mining, I do not believe that is how it will go down.
So, sorry, self-driving car companies. No thanks. I don't want my car reporting things about my life back to you any more than I want my TV or refrigerate doing that.
Don't forget that the working people have there own weapon of when living in the local jail is better then the street the big stores will go down.
Seriously - the article link is to a Vox article?
https://www.transportation.gov/AV/federal-automated-vehicles-policy-september-2016
One area not mentioned, at least in Vox's summary of the new regulations, is smart road technologies. Here, sensors in and around roads would share information with automated vehicles, and possibly mediate information exchanges between vehicles. Once many of the vehicles on a road are talking to the road and each other, it becomes straightforward to: i) Route traffic around obstructions, efficiently merging cars into open lanes well before the point of obstruction. ii) Clear a path for emergency vehicles. iii) Spread traffic efficiently over available routes to clear or avoid congestion. iv) Organize "convoys" of vehicles going in similar directions to take advantage of drafting and traffic light priority. iv) Operate a marketplace in which vehicles can bid for access to express lanes or right-of-way. (Late for a crucial meeting? Pay other cars $5 each to pull out of the way.)
// DevsVult: The Machines Will It
Industry: We want one set of nationwide (or, better yet, global) rules covering features and liability.
Feds: What rules would you like?
Industry: It hardly matters, since we make money in any predictable environment. But since you ask, we do have some suggestions . . .
"We reject as false the choice between our safety and our ideals." --The American President (20.1.2009)
Technology seems unflappable when it works. But given a mass system of control like self driving automation and a significant loss of the control systems could effect many of these vehicles not just one. Sure we have plenty of bad drivers out there and we focus way more on providing safety equipment to protect us from those bad drivers. Then fixing the bad drivers. How many times must you get stopped for DUI before we take away driving privileges? What happens if technology doesn't stop the accidents, what if true and complete self driving automation is decades away from being able to handle all situations? We have already seen what can happen when human's become over confident in self driving technology. We don't trust to have planes go without pilots even though we could also argue humans are and have been responsible for plane crashes. I think it has yet to be proven that technology is better and we have many hurdles to get over before many accept this technology as ready for replacing the human driver.
It's too easy to think of "edge case" scenarios that happen alllllll the time. I can see self-driving cars being an ideal solution for places like Los Angeles where it's sunny and nice out 95% of the time, but not in places where weather is a bit more unpleasant.
Now if the technology is implemented like something of a "super cruise control" that can be turned on/off by a human at will, then I think that is far more palatable.
Self driving cars will spell an end to any shred of privacy.
Today if you do not carry a cellular phone most cars do not "phone home" with data about your travels. Self driving cars will exist to collect as much data about you as possible and report back to Google or whomever.
I wish I could mod the previous AC up, but I'll have to use my karma instead. Here's a link to the actual subject of the story:
https://www.transportation.gov/AV
Amazing that neither the /. editor nor the linked/plagiarized Vox article bothered to provide it.
The fear of guillotine kept French elites in check. The same principle applies here, fear of communism kept Western elites in check up until collapse of USSR. Did you not notice that collapse of USSR coincided with decades of wage stagnation, loss of pensions, and general collapse of the middle class?
Best post in the entire thread. You win the internet today.
FUCK YOU, ASSHOLES.
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
We had such a socialist revolution. In 1917, in Russia. Ah, Cuba and North Korea had those, too. Do you like the results ?
I think he'd welcome those results as long as he's not in the cross-hairs.
I read this "comprehensive" guideline, because I have to for my job, and I have to say this is the most generic "comprehensive" guideline I've ever read. Just another set of govt. mumbo-jumbo.
We absolutely do not need smart roads. No. Smart roads with sensors all over the place are the opposite of what we need and will delay the adoption of autonomous vehicles if they become the focus of adoption or some sort of prerequisite. Not now and maybe not ever.
We need autonomous cars that are good at driving on dumb roads because dumb and sometimes poorly maintained roads are what the majority of roads will always be. Only after we have a good portion of the cars autonomous should we even begin to explore the need for further efficiencies that could be gained by integrating other data sources or subscribing to centralized route planning and coordination. Let the companies offer their own route planning and construction/incident avoidance solutions rather than make it part of any "smart road" centralized initiative. It isn't a safety issue.
The government should be focused on traditional roles of fixing potholes, maintaining the roads and maintaining line markings and appropriate signage for the benefit of both human and computer drivers.
Congestion tolls and taxes are another thing, but the solution is to just offer that data to whichever vendors want to subscribe to updates so that their systems can have the information and make appropriate driving plans based on passenger cost/time preferences.
This is one of the rare technology advances where the government's interests align with getting the technology to be pervasive (typically, you'd have to fight and/or circumvent the government to push disruptive technology... see SpaceX, for example). That will virtually guarantee government approval for mainstream use, and probably slightly before the technology is actually safe.
It's hard to imagine a better technology for the government, though. Track everyone driving, set speeds to whatever you want, stop any car at any time for any reason (goodbye high speed chases, hello stop-and-frisk on the highway), manipulate traffic arbitrarily (you want people to not drive through your neighborhood at all? just a $100M political donation, and it's done!)... the list of benefits for the government goes on and on. If I'm the government, the quicker and more pervasively I can push automation technology which I implicitly control into every aspect of people's lives, the better.
What is HAV? I understand the AV part is autonomous vehicle, but what is the H?
It certainly had an effect, although more from the opposition factor than any inherent value in the communist system.
Communism was a serious threat, and by threat, I don't just mean to the pocketbooks of capitalists. No one wanted what happened to Russia to happen to their country, and workers movements did get a bit of a lift from those trying to use them to block communism taking hold.
But make no mistake, its the political equivalent of using the threat of zombies outside the door to keep you focused on working together within. You're always better off with eradicating that sort of threat than trying to "harness" it to your advantage.
I won't even say we need a "better capitalism", but the ultimate solution needs to be better than a nuclear stalemate where the other side is oppression in the name of the "workers". I clearly remember the Cold War, and I have no nostalgia for the period in the slightest. At least today, I don't actually have to wonder if someone in Russia or NORAD is going to screw up and end the world on ten minutes notice. The threat is still there, but much more comfortably remote now.
Came here for baseless claims about the future, comments wholly devoid of a grasp of reality, and people fundamentally misunderstanding analogies.
Was not disappointed.
We have traffic lights that tell pedestrians to cross at the same time that they tell people to turn left into them.
We have increasing raddi turns.
The city planners who don't plan, and the city engineers who don't know anything about engineering usually ever get called out.
It's always cheaper to blame the pilot than the plane. It's cheaper to blame the driver than the make out cities coherent to drive in.
https://www.youtube.com/c/BrendaEM
from my cold, dead hands!!!!
The new tech could pass on productivity and improved living conditions to the general population, but nobody has figured out how to spread the wealth to consumers: it's bottlenecked at the top for some frustrating reason.
It's winner take most, #2 and #3 do pretty well, but 4 on down get hosed.
Capacity could greatly increase due to automation and a ready supply of 3rd-world labor, but the world economy is not "cycling right" to leverage that potential.
We'll probably have to experiment to figure out how to unjam the economy. Taxing the wealthy, and "helicopter money" are some possibilities.
Inflation is stuck at about 1.8%, but generally a strong economy needs about 2.2% or higher. It's like water pressure in a hydraulic system: too little, and it's sluggish; too much and stuff leaks and breaks. the needle has been hovering left despite low interest rates. Helicopter-money-theory is something that may increase the "money pressure" to a normal level, in part by giving regular consumers money to spend.
Table-ized A.I.
Practically speaking, the local jail (and all the rest of the imprisonment system combined) isn't large enough to handle the number of people that kind of outlook would bring if it becomes economically advantageous to any significant number of citizens. So I don't think what you're suggesting here is actually possible in any meaningful sense.
Something else would happen. Probably wouldn't be good, either.
I've fallen off your lawn, and I can't get up.
FYI:
The "breaks" are what you're trying to avoid - getting broken. The brakes are what you engage to decelerate the vehicle.
I've fallen off your lawn, and I can't get up.
No, you do see them -- they just aren't competent in as wide a range of situations as (some) human drivers are. They can take control of the vehicle and drive without any input from the driver in quite a range of circumstances. This is nto to say they are perfect, but neither are humans. In re humans, there are driving circumstances, various, that some humans are competent in, and some aren't; it's not like people are uniformly homogeneous in driving skills. So variation in skill is the norm, not the exception.
While humans aren't really improving much, self-driving vehicles are; and there's no reason at all to think they'll stop that development where humans typically do (and in some ways, they're already better -- for instance, the whole "I'm on a cellphone and not presently paying attention" thing.)
Sure, further to go. But they're going there, definitely on the way. No question about it. Quickly, too.
I've fallen off your lawn, and I can't get up.
Here you go.
I've fallen off your lawn, and I can't get up.
Clearly you either forgot to include the check, or it was too small. I suggest you try again, but fix whichever of those errors you made.
I've fallen off your lawn, and I can't get up.
What you're describing is why significant economic and social change are waiting in the wings as we discuss this. These changes are coming; and they will be qualitatively and quantitatively different than any changes prior, such as trains, fore and aft rigged ships, and the cotton gin.
Grab the popcorn.
I've fallen off your lawn, and I can't get up.
Hardly rare circumstance. Autopilot is useless on twisty roads (see below). It might be safer to have it turned on the motorways, but it would drive you off the road very quickly in a lot of country roads here in Europe.
They should have defined a driving test and force manufactures to limit the times the feature is turned on to the times it passes a thorough driving test similar to human driving tests. They haven't done that here, they took a more 'top level' conceptual approach and an implicit assumption that any claimed self driving car CAN SELF DRIVE. Which is dangerous, there's a stock scam in progress where the self driving car is a very crude neural network and they're trying to blindly teach it driving. THAT IS FUCKING DANGEROUS. If Autopilot works on the motorways, then it should only be possible to turn it on when its known to be safe on those motorways (proved to the regulator). When the person engages the autopilot, they should be sure it will work, the regulators should be sure it works, and Tesla should have satisfaction in knowing that an proper independent check has been done.
But actually writing down a first draft and getting their shit in gear is a positive step. Everyone needs to know the hurdles they need to jump, the DOTs slowness in doing their job left everyone in guess work limbo. I don't think vague fluffy top level document is it yet.
Video of Teslas autopilot being complete crap on a twisty road:
http://jalopnik.com/how-does-teslas-autopilot-handle-canyon-carving-1785428974?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+gizmodo%2Ffull+%28Gizmodo%29
Autopilot would fail a driving test on those roads if it was a human, so it should not be possible to attempt to turn it on on those roads till it can prove it can handle them.
Nope, it says here you're a caterpillar. That's the way it's going to be.
Christ, all 'is and meta morphing comments, too.
I've fallen off your lawn, and I can't get up.
it's bottlenecked at the top for some frustrating reason.
"helicopter money" are some possibilities. Inflation is stuck at about 1.8%, but generally a strong economy needs about 2.2% or higher.
I'm honestly not sure if this is meant to be a joke. Do you think that helicopter money will go to poor people or rich people/banks?
2004 Nobody won the DARPA grand challenge... no car completed the course.
2005 5 vehicles completed the course.
2007 they switched to an urban course having to obey the rules of the road and six teams finished the course.
That is rapid progress.
From 2007 to 2016 we have seen pretty steady progress with commercially available features for things like automatic parking, automated braking and collision avoidance, widespread use of GPS navigation (via smartphones and built-in) and more recently the fully autonomous highway driving from Tesla (yes I've seen the people reading books while "at the wheel").
And Google has been pretty open in their fully autonomous car project with two different cars one based on an off the shelf lexus and another custom built electric vehicle: 1.5 million miles driven and "currently out on the streets of Mountain View, CA, Austin, TX, Kirkland, WA and Metro Phoenix, AZ"
And we are seeing Uber's autonomous efforts play out in Pittsburgh. Multiple companies, multiple projects, multiple on-street implementations that are getting better and better.
Google wants to ditch the steering wheel altogether and it was California regulations that held them back a few years.
Anything can be fouled up if you execute it wrong.
Print more direct money, and spend it on repairing infrastructure. That would create a good many construction and repair jobs. Much of that stuff has to be fixed anyhow; postponing it just makes it grow riskier and even more expensive to fix later.
Table-ized A.I.
presidential motorcades = full manual or auto with rules like
Run over ok
Rundown ok
Drive anywhere ok
Driver anywhere with non blocking damage risk ok
No speed limits ok
Run red lights ok
Run rail gates at drivers choice
Well, if by "out" you mean you can buy one, no. But if by out, you mean, humanity has this technology, then yes, we do.
I've fallen off your lawn, and I can't get up.
lol. Free doctor, eh?
Someone's been feeding you a line of propaganda.
You get incarcerated, if you're *lucky*, you'll see a doctor after a few months, and even then, probably only if someone beats you within an inch of your life.
The free medical care? Mostly myth, and absolute bottom feeder quality if you do get any. Too little, too late if it's serious.
The free food? Trust me, you don't want to eat it. Any of it.
The free room / roof? Sure. Complete with free beatings, rape, and deaf/blind guards. And an hour a day in the yard. Pick your power group.
The free education? Awful. You can't use it anyway, you're a felon. Start figuring out who and how to rob, because that's the only decent-paying job you can get hired for now.
Jails and prison in the US are hellholes, the inevitable fruit of the popular retribution-centric mindset combined with scare-the-populace political maneuvering. You approve, okay, that's your call but don't try to sell those places as poverty relief. There's no relief to be found. Of any kind.
I've fallen off your lawn, and I can't get up.
The NHTSA FAV policy is clearly influenced more by politics/news and someone trying to score points in a press conference than safety. Of 14 points of 'safety', many have nothing to do with the safety of self driving cars.
* Privacy(omg googles is spying on me )
* System Safety(isn't this about safety )
* Vehicle Cybersecurity(omg their hacking our cars)
* Human Machine Interface(this shouldn't be that new , we have cruise control rules,AV should extend those )
* Crashworthiness(Already have crash standards)
* Consumer Education and Training (there are no standards for items already)
* Registration and Certification(should really be in here, you must follow their guidelines to be approved. )
* Post-Crash Behavior(relevant)
* Federal, State and Local Laws(Federal already supersedes local laws, ie federal speed limits example)
* Ethical Considerations( don't be Volkswagen or Mitsubishi)
* Operational Design Domain( needs "system engineering approach" and we need pretty pictures to understand)
* Object and Event Detection and Response(relevant , wouldn't be very autonomous if it is didn't do this though)
* Fall Back(relevant)
* Validation Methods(really we have nothing related to this already?)