An Anonymous Coward writes:
As predicted by many (including posts here on SN), extensive testing now shows that if the driver's workload is reduced to near zero they are in no position to intervene should the autonomous system get in trouble.
The Detroit-based company has tried many ways to keep its engineers alert during autonomous car test runs, employing everything from alarm bells and lights to even putting a second engineer in the vehicle to monitor their counterpart. "No matter — the smooth ride was just too lulling and engineers struggled to maintain 'situational awareness,'" said Ford product development chief, Raj Nair.
Ford's strategy of eventually removing the steering wheel and pedals from self-driving cars has ignited a debate between automakers on how to approach the development of Level 3 self-driving vehicles, or if Level 3 should even exist at all.
BMW, Mercedes-Benz, and Audi will introduce semi-autonomous Level 3 vehicles next year that require human intervention within 10 seconds or the vehicle will slow to a stop in its lane. However, other automakers like Nissan and Honda have upcoming systems that give the driver 30 seconds to prepare and re-engage the vehicle or it will pull to the side of the road.
The article continues with quotes from other manufacturers and US DOT. As a reminder, levels from 0 (no automation) through 5 have been defined by SAE. Level 3 is "conditional automation" and it's starting to look like this level is not such a good idea.
(Score: 0) by Anonymous Coward on Tuesday February 21 2017, @06:58PM
How about driving software that would sense traffic and buildings (and optionally, pedestrians, bicyclists, motocyclists, stray animals, etc.), then render them as colorful Tetris-like blocks? The human would stay engaged and entertained. Or instead of Tetris, it could be modeled on Exidy's Death Race.
(Score: 2) by bob_super on Tuesday February 21 2017, @07:16PM
> Exidy's Death Race
That's the problem here: automatic cars don't break the rules of the road... which is extremely boring, outside of the feeling they have to be getting about imminent rear-ending by semis.
Driving with the flow, while showing the user the 360 annotated view of other cars, might keep passengers interested for a little while (engineers, at least).
But to keep people awake, you really need to start scaring them occasionally, at least by having trigger-happy warnings about cars drifting erratically (there's always one), or by speeding and swerving a little more than necessary.
(Score: 2) by krishnoid on Tuesday February 21 2017, @07:45PM
But to keep people awake,
Maybe stop overworking your engineers and let them get a full night's sleep?
(Score: 2) by tibman on Tuesday February 21 2017, @07:59PM
Fine. But we're taking it out of their pay.
SN won't survive on lurkers alone. Write comments.
(Score: 2) by krishnoid on Tuesday February 21 2017, @10:09PM
"To save costs, we're going to decrease your salary by $20,000 per year. In return, you will be *guaranteed* a 40-hour workweek."
"Flex time?"
"Yes -- that 40 hours will be flexible. Now who's volunteer-- no no no !!!"
Detroit Free Press: Ford Engineering Director maimed in riot.
(Score: 0) by Anonymous Coward on Tuesday February 21 2017, @07:53PM
I meant to say, this would not be fully automated driving software. The computer would sense the surroundings and draw them as an animation with bright colors and realistic sound effects, making driving fun for the human player. A traffic jam turns into a puzzle. A pedestrian turns into a hopping frog.
(Score: 1, Funny) by Anonymous Coward on Tuesday February 21 2017, @07:42PM
We could give the 'player' a controller that looks like a steering wheel and two or three buttons located at their feet. I also have a solution for delivery of 'life-like' video-quality which involves a translucent crystaline structure which we place in front of the player.
(Score: 2) by bob_super on Tuesday February 21 2017, @08:00PM
Don't be silly, we don't have yet the tech for Glasses-free 3D at these kind of resolutions.
I hope your solution included at least real-time @POTUS twitter feed on the crystalline structure, so that players would keep being reminded that they are in an alternate reality.
(Score: 2) by davester666 on Wednesday February 22 2017, @10:23AM
I vote for requiring the engineer to hold a "steering wheel" (that doesn't actually control the vehicle) for it to move, along with a gaze-tracking camera, and if it determines you have nodded off, you get an electric shock. And removing your hands from the wheel results in a panic stop along with an airbag deployment.
(Score: 4, Insightful) by requerdanos on Tuesday February 21 2017, @07:04PM
I am no engineer, but that does not strike me [smugmug.com] as a great failsafe plan.
I am no engineer, but that does not strike me [flaglerlive.com] as a great failsafe plan.
A "Failsafe" [oxforddictionaries.com] is usually a good idea, because if it literally fails, it is still literally safe.
This level, or its equivalent, works well and safely every day in aviation, but then you almost always have to demonstrate competence to perform aviation. Not so in operating a terrestrial vehicle.
(Score: 2) by ikanreed on Tuesday February 21 2017, @07:35PM
Failsafe in a metal box hurdling at 30 m/s are never going to be entirely failsafe.
Failsafe is impossible. Failsafe shouldn't be a goal. Least net harm should be a goal.
(Score: 4, Funny) by art guerrilla on Tuesday February 21 2017, @07:58PM
heh, buddy, thisy here is 'murika, we do our metal box hurdling in mph...
and ain't it just like a nerd, people are crashing and burning, and his first thought is he wants the least harm to the internet...
nice...
(Score: 2) by requerdanos on Tuesday February 21 2017, @08:06PM
Failsafe shouldn't be a goal. Least net harm should be
Okay; fair enough. What response by the autopilot causes the least net harm? Pull over, slow to a stop in your lane, or something else?
You can't always pull over; on some roads, like the one pictured in the link above, it's impossible on one side (there is a cliff face there) and a bad idea from a safety and stability perspective on the other (there is a steep dropoff there). Something (almost anything) else would have to be done here.
Slowing to a stop in whatever lane you happen to be in would be fine if another automated vehicle is behind you, but if a human is behind you and he does not see an obvious reason you should be stopping, then he is much less likely to stop in time to avoid smashing into you. I could see someone not paying attention to their level 3 car autopilot slowing down in their lane because they weren't paying attention, then getting rear ended and suing the manufacturer if he survives.
My post was kind of tongue-in-cheek above, but the questions are real ones.
(Score: 1, Funny) by Anonymous Coward on Wednesday February 22 2017, @04:02AM
The famous trolley problem. I recommend multi-track drifting.
(Score: 3, Touché) by tangomargarine on Tuesday February 21 2017, @09:27PM
but then you almost always have to demonstrate competence to perform aviation. Not so in operating a terrestrial vehicle.
They let you drive without a license?
"Is that really true?" "I just spent the last hour telling you to think for yourself! Didn't you hear anything I said?"
(Score: 4, Insightful) by theluggage on Tuesday February 21 2017, @10:35PM
This level, or its equivalent, works well and safely every day in aviation, but then you almost always have to demonstrate competence to perform aviation.
That's true. Plus - I Am Not A Pilot, but in my ignorance I assume that following things don't happen much in planes:
Seriously - as I understand it, Air Traffic Control start filing incident reports if two planes pass within several hundred metres of each other. CF two lanes of opposing ground traffic a metre or so apart with closing velocities of more than 100mph...
What does happen in a plane (...and we're talking airliners and general aviation here, not Top Gun) that needs a reaction within a second or two (esp. when the autopilot is likely to be engaged)? Turbulence?
(Score: 1) by khallow on Tuesday February 21 2017, @11:29PM
This level, or its equivalent, works well and safely every day in aviation, but then you almost always have to demonstrate competence to perform aviation. Not so in operating a terrestrial vehicle.
Autopilot is not level 3.
(Score: 2) by requerdanos on Wednesday February 22 2017, @02:30AM
Autopilot is not level 3.
According to http://www.techrepublic.com/article/autonomous-driving-levels-0-to-5-understanding-the-differences/ [techrepublic.com], level 3 is characterized as:
Drivers are... able to completely shift "safety-critical functions" to the vehicle, under certain traffic or environmental conditions. It means that the driver is still present and will intervene if necessary, but is not required to monitor the situation in the same way it does for the previous levels.
This is basically the case with autopilot avionics, which can sometimes but not always land a plane, usually but not always fly an entire route, and never but maybe someday take off.
The following article, http://www.cnbc.com/2015/03/26/autopilot-what-the-system-can-and-cant-do.html [cnbc.com], specifically says this:
Before takeoff, the pilot will enter the route into the computer, giving [the autopilot] a start and end position and exactly how to get there... Generally, the pilot will handle takeoff and then initiate the autopilot to take over for most of the flight. In some newer aircraft models, autopilot systems will even land the plane.
Occasionally, ... the autopilot will disengage itself in the event of extreme turbulence, for example, at which the pilot will be alerted to take over control of the plane.
That's a pretty high correlation with level 3's "able to completely shift 'safety-critical functions' to the vehicle, under certain traffic or environmental conditions. It means that the driver is still present and will intervene if necessary". No, autopilot is not level 3, but it matches level 3 much more closely than levels 0, 1, 4, or 5.
And it exactly parallels TFS's "if the driver's workload is reduced to near zero they are in no position to intervene should the autonomous system get in trouble." Somehow pilots in the same circumstances are usually able to keep paying attention, maintain situational awareness, and take over if necessary. The aviation standards are higher.
(Score: 1) by khallow on Wednesday February 22 2017, @08:18AM
And it exactly parallels TFS's "if the driver's workload is reduced to near zero they are in no position to intervene should the autonomous system get in trouble." Somehow pilots in the same circumstances are usually able to keep paying attention, maintain situational awareness, and take over if necessary. The aviation standards are higher.
Pilots aren't in the same circumstances for the most part. For example, there are two to three pilots on such planes, often working in tandem, and the demands on the autopilot and the pilot are over a longer time frame in which situational awareness is easier to retain and regain.
And standards can only do so much to compensate for bad system design.
(Score: 2) by choose another one on Wednesday February 22 2017, @11:31AM
I agree that level 3 is much the same as aviation autopilots, the auto industry is now going where aviation went 40 or so years ago.
Somehow pilots in the same circumstances are usually able to keep paying attention, maintain situational awareness, and take over if necessary. The aviation standards are higher.
Drivers are trained (a bit) to drive, pilots are trained to fly and trained to monitor, particularly in commercial (usually 2 pilot) aviation where one pilot monitors the other or monitors the autopilot. Monitoring is emphatically not the same skillset as doing, and drivers are not currently trained to monitor at all.
Even so, there is still much concern in aviation about over-dependence on automation (e.g. google "children of the magenta line") and lack of situational awareness, and there are several notable accidents that are arguably a result of it (AF447 for one). Notably, however, aviation is still safer overall than it was in the "old days", there are plenty of old pilots now flying armchairs and bemoaning the new generation's lack of stick and rudder skills, but they do tend to forget that they themselves are a self-selecting sample of what was - the "old days" pilots with poor skills are not here anymore...
The real question is whether the car industry, and all road users, will have to learn again the painful lessons that aviation already learned (in some cases still is learning) about automation and automation dependency.
(Score: 5, Touché) by DannyB on Tuesday February 21 2017, @07:23PM
Try changing the control labels.
[_] Notify me of impending death or injury.
[x] Do not disturb my sleep unless we reach destination.
The lower I set my standards the more accomplishments I have.
(Score: 4, Informative) by meustrus on Tuesday February 21 2017, @07:39PM
See the actual policy which defines levels 0-5 [sae.org].
More commentary on the differences between each level. [techrepublic.com]
Hyperlinks were not invented to drive traffic back to your own web site, MotorTrend. They were invented to streamline citations so that we can all see the same primary sources that you saw. Of course that would make you a glorified aggregation engine instead of a news site, wouldn't it? Not that aggregation engines are unsuccessful - says the commenter on the SoylentNews aggregation engine.
If there isn't at least one reference or primary source, it's not +1 Informative. Maybe the underused +1 Interesting?
(Score: 1, Funny) by Anonymous Coward on Tuesday February 21 2017, @08:23PM
drive traffic
I see what you did there.
(Score: 0) by Anonymous Coward on Tuesday February 21 2017, @09:20PM
For the record, I submitted the article as AC, and I don't work for Motor Trend. There is a source link in the M/T article but it was to a paywalled site, http://www.autonews.com/article/20170217/MOBILITY/170219851/fords-dozing-engineers-side-with-google-in-full-autonomy-push [autonews.com]
(Score: 3, Interesting) by meustrus on Tuesday February 21 2017, @11:49PM
I have no issue with the summary. I just found it really annoying that all 10 links in the article (6 of which are for specific makes rather than actual information) point straight back to MotorTrend, even when they are clearly referencing actions that third parties made. Third parties which clearly make press releases. The articles themselves don't link to any concrete information either, even going so far as to describe their source as "NHTSA" with no link or even any identifying citation.
The paywalled article is even worse, having no links whatsoever. And after closing and reopening the tab a couple of times, it also has no content. Not sure that's much of a change from before since I don't give 2 shits what some e-zine I've never heard of has to say about a a source that it is deliberately withholding.
If there isn't at least one reference or primary source, it's not +1 Informative. Maybe the underused +1 Interesting?
(Score: 2) by BenJeremy on Tuesday February 21 2017, @08:27PM
...based on my daily commute into Detroit, Ford engineers don't exercise a lot of "situational awareness" when driving their regular "Level 0" cars.
I witness legions of people on auto-pilot, not the good kind - changing lanes without signalling, tail-gating (which leads to stop and go traffic) and late merging (not zipper, which requires both lanes to be the same speed). In general, some people are decent drivers, but even decent drivers get lulled into leaning on their subconscious to do the driving.
Autonomous driving will probably be much safer, at any rate. Even Level 3.
(Score: 1) by khallow on Tuesday February 21 2017, @11:31PM
...based on my daily commute into Detroit, Ford engineers don't exercise a lot of "situational awareness" when driving their regular "Level 0" cars.
I too find this hard to believe, but not everyone in Detroit is a Ford engineer. And having situational awareness is not a consequence of being an engineer.
(Score: 2) by BenJeremy on Wednesday February 22 2017, @01:14PM
Well, there are a lot of engineers on the roads during my commute... GM, Chrysler, vendors.... not many people are situationally aware, no matter what their employer or job.
My main point was that it isn't a circumstance of the automated driving, but rather just a reality of a disengaged participant, and that many are already driving on their own "autopilot" that is far more dangerous. Removing some of the control from people who would already be somewhat oblivious isn't going to promote more alertness, it will only make it easier for the driver to disconnect.
(Score: 3, Funny) by Gaaark on Tuesday February 21 2017, @08:34PM
Have the car temporarily swerve into oncoming traffic! The adrenaline rush and the smell from their pants will keep them awake for a while.
--- Please remind me if I haven't been civil to you: I'm channeling MDC. ---Gaaark 2.0 ---
(Score: 2) by JoeMerchant on Tuesday February 21 2017, @09:27PM
Nobody has mentioned chemical stimulants - give the engineers a good onboard coffee dispenser.
🌻🌻 [google.com]
(Score: 2) by number6 on Wednesday February 22 2017, @03:00AM
Discard the idea that cars need to be fast futuristic-looking slick desirable luxurious things; take your design cues from the simplicity, efficiency and utility of bicycles, trains, and the lifestyles of the Amish culture.
i.e. design cars with utility and efficiency as the only priority; super-low rolling resistance thin large diameter spoked exterior wheels like penny-farthing bicycles and old vintage cars; also super-springy leaf-spring suspensions; this allows you a super-comfortable efficient ride over rough terrain without needing roads which look like bowling alleys .....................stop manufacturing roads from concrete/asphalt/tar; build road systems by hydraulic-tamping the dirt and consolidate its surface with some environmentally friendly polymers.
Restrict the horsepower and speed to sane levels and discourage the idea of sports/muscle cars; .........allow the front windscreen to fully swing up exposing the occupants to be exposed to the oncoming wind (in nice conditions or when feeling drowsy).
Design transport infrastructure around longer time-frames. Re-define distances and densities and spaces and storage elements and supply chains. Slow down and smell the roses in all areas of human endeavor.
Make a thing which is more repairable and recyclable by a single human than the last created thing.
My concluding opinion on this whole driverless/autonomous car phenomenon:
It's a big joke being played out by technology maniacs without the will to retract their mouth from the dick of wealthy old oligarchs who've lost their moral compass;
this project does not include higher-order thinking in its parameters.
(Score: 1, Informative) by Anonymous Coward on Wednesday February 22 2017, @03:45AM
build road systems by hydraulic-tamping the dirt and consolidate its surface with some environmentally friendly polymers
Someone likes mud and ruts. That 'polymer' will do nothing. Lets be nice and say a car weighs 1 ton (older models can be as high as 4). That means each wheel has 250 pounds in a 1/4 ft area. The thinner the tire the more weight on the area. The bigger the area the more distributed it is.
Now do the same thing with truck. The legal weight is 80,000 pounds. Or about 4444 pounds per tire. They can also get permits to go overweight so long as they stick to particular roads.
Now add in water, ice, grit, dirt, and other things. Those other things become abrasives. Many states have tried making roads out of plastic. Usually recycled 2liter bottles. They found the grit ate the roads up in no time at all. So they went with mixed and it was better but they still did not last as long. Also add in weeds. So you better have some way to remove them. As weeds do not stop. So now you are talking chemicals and frequent spraying.
The cement and asphalt is actually cost effective. Repaving roads is expensive in time, money, and traffic jams. The material itself is actually the CHEAP part (2-3 cents per pound). Properly grading and water run off design is the part where all the cost goes. Many times they regrind it in place and re-mix with new just to save cost. Cement has an amazing amount of compression strength but poor lateral strength. Dirt has neither. Asphalt is usually slathered over cement (not always). Because it has an interesting property. It is slightly gooey and easy to lay down a new layer. It can sorta self heal in hot weather. Which is nice as you do not have to come out and tear up the road every 3-4 years but maybe every 10. However, it is very brittle in cold weather. It is why you see it more in the south and not as much in northern areas. You see it in cities more as they like the speed at which it can go down. Which means minimizing the disruption tearing up the roads creates.
For some areas your idea may be quite feasible. Others not as much. Sorry. There will be a long cross over of driving/self drive cars. For cars the cycle is around 10-15 years. Someone will pay the price for all of that new infrastructure you want. Who will do that? I can 100% guarantee it will not be those rich oligarchs. It will be you and me with a good portion ending up in their pocket. They will make up a tax for it. On the other end what do I get out of it? A lesser system that has more limitations and costs more to do anything with.
(Score: 1, Insightful) by Anonymous Coward on Wednesday February 22 2017, @07:52AM
there are vehicles that need to move as fast as possible: ambulances, fire trucks. and police cars, although some may argue against that.
society will not allow you to slow down these vehicles.
strong roads are also strategically relevant.
if you need to move many soldiers very fast, you need good roads for big heavy vehicles, including tanks.
(Score: 0) by Anonymous Coward on Wednesday February 22 2017, @08:13AM
" thin large diameter spoked exterior wheels "
That's what Americans are putting under their cars already in every car building tv show and IRL. Those chromed, low profile, spoked wheels, that look just like the cart wheels, except a little wider. Holy shit they are ugly.
(Score: 2) by carguy on Wednesday February 22 2017, @03:06PM
> ... design cars with utility and efficiency as the only priority ...
I think you are reinventing the Holsman -- https://en.wikipedia.org/wiki/Holsman_Automobile_Company [wikipedia.org]
A good friend drives his 1906 model for short trips (Honda minivan for everything else). I've driven it, it didn't take long to get used to steering with the tiller and using the twist grip throttle (reversed from motorcycle, wrist-moves-up for faster). It's kind of like driving a lawn tractor, but with seating position higher than most pickups and SUVs. 35 mph is possible, but not a good idea to turn at that speed. He's nearly 80 now and hand cranking the opposed two-cylinder engine was difficult, so he added an electric starter.
The original customers were often country doctors in the midwest who replaced their horse and buggy with something that could negotiate the rural dirt roads of the day and keep them above the mud.
(Score: 1) by anubi on Wednesday February 22 2017, @06:52AM
What time interval does one usually react in to avoid a surprise encounter with another car, pedestrian, animal, or other unforeseen obstacle?
My empirical experience indicates usually less than a second.
If I am paying undivided attention to my driving, I usually *barely* react fast enough. I do not like even playing the radio or talking to anyone when behind the wheel. ANY inattentiveness on my part sharply increases the probability that I fail to react fast enough, and will now require insurance adjusters and lawyers.
Its been my experience that only things like flat tires, engine problems, being lost, or out of gas, give me the luxury of more time to react to the problem.
"Prove all things; hold fast that which is good." [KJV: I Thessalonians 5:21]
(Score: 2) by ledow on Wednesday February 22 2017, @08:05AM
Prima facie evidence that humans shouldn't be trusted, shouldn't be responsible, thus shouldn't be *able* to take control of these vehicles except on an emergency override basis.
And if we're heading that route, you have a much tougher sell ahead of you, car manufacturers, because it's no longer "our" car.
But one day, this will come up in court to say "Look, the manufacturer's own engineers couldn't keep focused when they had no direct control, how do you expect a man on the street to, therefore I shouldn't be held responsible for the car swerving into that playground and taking out that little girl" and you're going to have a hard time fighting that lawsuit because your own guys couldn't keep awake doing the same under test conditions.
The only logical outcome of automating a human's actions is removal of the human from those actions entirely. Anything else is a nonsense. Sadly, the technology isn't up to driving on its own either, so by pressing ahead with it, all you've done is open yourselves up to liability for your "AI"'s actions and now it's going to get messy.
- Want to take control from the human.
- Don't want to take responsibility for that control yourselves.
- Provably can't trust the human any more when their control is taken away.
One of those is going to have to give.
As I tell my boss, in every workplace I can have the responsibility AND the control, or neither.
(Score: 0) by Anonymous Coward on Wednesday February 22 2017, @03:16PM
One possible scenario -- you program a destination when you get in the car. The car checks out the route and decides if this is a route that it can drive autonomously. Due to incomplete maps, bad weather (snow over the lane markings...), cell-network-gaps, recent road works, etc., the autonomous system may refuse to engage and the human has to drive manually.
In this case, there is no hand off mid-trip. At least this solves the inattention problem -- if the car "knows" the chosen route the driver can safely snooze (well, as safe as possible when moving in a box at high ground speed).
If the driver wanted to go autonomous when available, they could program several way-points and the car might agree to drive for some of them, but not others. Again the hand off would be done at a known safe place to stop.