Slashdot Mirror
US Regulators Investigating Tesla Over Use of 'Autopilot' Mode Linked To Fatal Crash (cnbc.com)
An anonymous reader quotes a report from CNBC: The U.S. National Highway Traffic Safety Administration said on Thursday it is opening a preliminary investigation into 25,000 Tesla Motors Model S cars after a fatal crash involving a vehicle using the "Autopilot" mode. The agency said the crash came in a 2015 Model S operating with automated driving systems engaged, and "calls for an examination of the design and performance of any driving aids in use at the time of the crash." It is the first step before the agency could seek to order a recall if it believed the vehicles were unsafe. Tesla said Thursday the death was "the first known fatality in just over 130 million miles where Autopilot was activated," while a fatality happens once every 60 million miles worldwide. The electric automaker said it "informed NHTSA about the incident immediately after it occurred." The May crash occurred when a tractor trailer drove across a divided highway, where a Tesla in autopilot mode was driving. The Model S passed under the tractor trailer, and the bottom of the trailer hit the Tesla vehicle's windshield. Tesla quietly settled a lawsuit with a Model X owner who claims his car's doors would open and close unpredictably, smashing into his wife and other cars, and that the Model X's Auto-Pilot feature poses a danger in the rain.
It was bound to happen sooner or later.
Luckily for Tesla this sounds like it couldn't have been avoided in any way.
There will be more... but, like Tesla says, their Auto-pilot system has thus far proven VERY safe. What remains to be seen is how the world reconciles the fact that there will always be outliers...
That's still pretty impressive if it's twice as safe as letting a human drive.
Even more so after seeing all the videos on youtube with people in the back of the car letting tesla drive.
Minimum threshold fixed. Thanks!
"the first known fatality in just over 130 million miles where Autopilot was activated," while a fatality happens once every 60 million miles worldwide. Autopilot is only allowed on highways, whereas I am sure they are comparing 60 million miles against normal driving which is inherently more dangerous than all cars heading in the same direction with barriers between the traffic flow. Apples and Oranges.
If my car is in autopilot, and I take control of the vehicle just before dying in an accident, is it considered an autopilot fatality?
Tesla's autopilot mode is still in beta, so it isn't a big deal. Because it's apparently OK to sell cars that have only been beta tested at most.
Tesla noted that customers need to acknowledge that autopilot "is new technology and still in a public beta phase" before they can turn it on. Drivers also acknowledge that "you need to maintain control and responsibility for your vehicle."
>bottom of the trailer hit the Tesla vehicle's windshield
aka driver got decapitated smokey and the bandit style :o
Who logs in to gdm? Not I, said the duck.
And some of you want us all to get into a so-called 'self driving car' with no manual controls that would at least give you a chance to save yourself. I'll just keep driving myself the old-fashioned way, thanks anyway, because I want to live. Ask me again in 50 years. I'll still say 'Hell, NO!', but at least then you'll have decades of data instead of a measly few years' worth.
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
According to the article this was not something the driver could see and avoid, and while the autopilot could not see it either, based on the data from this crash it *could* see it the next time. Drivers learn from their own experience and fatal crashes terminate their learning experience, while autopilots learn from ALL autopilots on the road, and there are no "fatalities".
Of course it is always the fault of humans in the end, in this case the tractor trailer was not supposed to be there, so we'll only have perfect records when we get rid of all the drivers and have all the cars on autopilot.
Violence is the last refuge of the incompetent. Polar Scope Align for iOS
This would not have occurred I suspect.
The new model from Tesla will transform into a coffin
Yea most Tesla's are in Orange county Silicon Valley, London (assuming they export) etc. However if you take the world into considerations with traffic everywhere then this statistic may not be as straightforward.
If you recall Tesla cars for this, then you should recall all human driven cars as being unsafe for allowing human drivers.
No statistic is, especially when coming from the PR department...
What's that saying? Lies, damn lies and Statistics? No, the other one... Figures never lie, but liars figure..
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Until "autonomous" means exactly that, we will have people lulled into not paying attention, and a driving system that cannot handle everything that is thrown at it. The result will be crashes.
No manner of EULA, or cries of BETA will get around that predictable result.
Expecting human nature to change to match your product's limitation is a fool's journey.
Drivers learn from their own experience and fatal crashes terminate their learning experience, while autopilots learn from ALL autopilots on the road, and there are no "fatalities".
Nonsense. Humans are capable of learning from situations that they did not directly experience. For example, I've never owned a Tesla and yet I learned today that Tesla autopilot is dangerously under-designed and is not to be trusted at this time.
As far as not driving into large objects and driving blindly, I already learned those things without killing myself. In fact, I would say that knowledge of my own mortality makes me a better learner in some cases.
BS. Death for autopilot means nothing. Human drivers value their lives (except where I have to commute) and will do a lot to try to save them, including paying attention to what happens in front of them. I doubt that you can force autopilot to value human life or their own "life".
"Neither Autopilot nor the driver noticed the white side of the tractor trailer against a brightly lit sky, so the brake was not applied," Tesla wrote.
The driver likely didn't notice because he wasn't paying attention. That is something drivers can see when driving. You can see the front of the truck cross in front of you, you understand the sky doesn't suddenly shift colors, you understand previously seen road doesn't suddenly turn into sky, and if all that escapes you you can still see the truck wheels along the road. The car probably saw them too and figured it would fit between them. Why didn't the radar see the truck? Too angled towards the road?
None of the articles mention who died. Did the driver die? If so, how do we know he didn't notice the truck? Did a ghost tell us? If someone else, then we can't trust the driver isn't trying to cover his ass. No one gets into accidents on purpose unless you're making YouTube videos.
No, this was something the driver did not avoid. We don't know if/when the driver saw it, or if the driver could see it.
It is likely the driver was staring at their phone and not looking at the road because they assumed the car's autopilot mode worked.
"Tesla Autopilot has crashed. Do you want to submit a bug report?"
(Too soon?)
Why didn't the radar see the truck?
The radar saw the truck's trailer, but misidentified it as an overhead sign, because it was so high off the ground.
Did the driver die?
Yes.
If so, how do we know he didn't notice the truck?
If he had noticed the truck, he presumably would have applied the brake. (we'll have to assume the driver wasn't feeling suicidal)
I don't care if it's 90,000 hectares. That lake was not my doing.
What's the point of having Autopilot if you have to stay fully involved? Might as well drive then and avoid the autopilot disengaging at some unfortunate moment.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
It would be a high priority to look for vehicles crossing unexpectedly. That is a major cause of accidents. And not that hard to do with just stereo vision.
However, the article contains no useful information. Did the AutoPilot actually see it? Did it interpret it correctly? Did it try to slow down at all? At what point after the truck started moving (but before it entered the other road) did the AutoPilot react?
Those are the critical questions.
I think it would be safe to assume that the auto pilot did slow down once the truck was in front of the car, but that, of course, would be too late.
It was a frigging tractor-trailer crossing the street. How can anybody in their right mind claim that this was impossible to see??? If you, as the driver, cannot see a tractor-trailer crossing the street (no matter whether it's white or any other color), then you shouldn't be driving a car. If the car's AI cannot distinguish it either, it's not fit to control your vehicle. Very simple. You can be pretty sure that any human that pays attention while driving would notice a truck crossing the street, and hitting the brakes in time (actually, you would most likely already see it approaching the intersection, and, if it doesn't slow down, you can probably guess that he's trying to make it across, and you would react accordingly). If, of course, you *don't* pay attention (or you have a color-blind AI driving the car), then all bets are off. But don't tell me "it's impossible" to see a white tractor-trailer...
If my Tesla Model S is parked on the street, and some truck smashes into it, sending my car onto the sidewalk where it rolls over a pedestrian... is it Tesla's fault?
This article reads like a hit piece.
In addition to the Tesla having Ludicrous speed mode, it appears there's a secret Kevorkian speed mode.
Some drink at the fountain of knowledge. Others just gargle.
It sounds like the truck crossed the lane without enough time for the oncoming cars to make it but all we hear is how the autopilot is at fault. I can understand how the the sensors missed the trailer and that is going to be something all developers will have to add to their tests (when seeing a rig with a space after it then check for tires).
We are going to see cases like this come up now and again with self driving cars but there won't be a need for a recall. What should happen is an alert go out to the owners of cars while the manufacturers check their systems. If their cars pass tests then they can send out messages to their customers. If not then they create an update, test it, verify it, and send it out. Until owners of the cars hear that the system has been verified then they need to be extra vigilant when such an event happens.
Tesla is clear in their manual and click-through agreement that autopilot is not a self-driving car, but a sophisticated driver assist. Most people won't read anything, though, so I suspect the majority of Tesla owners don't understand this.
What's the point? Most things people like about cars are subjective. I like the (fairly limited) driver assists in my car a lot. The intelligent cruise control is great for limiting your top speed when a cop is around, the "beep if you're in a dangerous situation" feature has helped me a couple of times when my mind was wandering when the guy in front of me decided to stand on his brakes. I think it's good stuff.
Socialism: a lie told by totalitarians and believed by fools.
It's because the autopilot as implemented in a Tesla car just doesn't make sense. It's bizarre to have a system that purports to allow you to not pay attention to driving, yet still have you have to pay attention to driving. Things that don't make sense tend to not have a place in the way people understand and use things. It isn't ready, it doesn't work with human psychology, and it should have never been put on the market in its current state.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
This type of accident is easily avoidable by a human. Semis are really slow moving when taking turns like this. The fact that the truck was that far into road means that a person, smart enough to not buy a piece of shit electric death trap, would have more than enough time to slow down and stop.
This moron deserved the death he got for, one buying a piece of shit electric car, and two putting a lemon like a tesla in auto pilot mode. All tesla owners are colossal morons. I look forward to more of them getting killed by these lemon death traps they stupidly over paid for.
If this had been a Google car, there would have been no accident. Google has much better radar which maintains a model of all vehicles in the vicinity before they turn into your direction, and it's high enough up that it would not miss a trailer.
Bruce Perens.
"so we'll only have perfect records when we get rid of all the drivers and have all the cars on autopilot."
And all the cyclists and all the children and all the senior citizens and all the animals....
Basically the only way to get perfect records is to create roads that are dedicated to driverless cars. Then a very complicated problem suddenly becomes fairly easy to solve.
If I understand correctly he was an enthousiastic early adopter, putting youtube videos up that showed how well the computer in his Tesla did. He was probably reading a book or something just to show off his smart car.
Autopilot mode isn't "hands off" or true autonomous driving. In fact, Tesla's implementation doesn't even use the GPS. It's really a more sophisticated lane keeping and cruise control system. It can change lanes, but you have to command it to do so.
In fact, if you take your hands off the wheel, you have about 30 seconds to get them back on (with the car steadily warning you) before it automatically comes to a stop.
So why have it? It can reduce driver fatigue - if you're in heavy stop-and-go traffic, the autopilot will keep pace and slow down with traffic, so you're not spending lots of time and mental effort accelerating and braking. Given it's an electric car, doing so efficiently can be a challenge, so a computer makes sense.
It also helps because highway driving is some of the most boring around, and remaining awake at the wheel is quite difficult, so a momentary lapse of attention is not so bad.
It's why it's called "autopilot" and not self-driving. It's like a plane's autopilot - a lot of more basic units are single axis units and they're really meant to help offload some of the pilot's work. Of course, more modern units (until you get to airliners, single and dual axis units are the norm) have an emergency recovery vbutton that tries to get the autopilot to help re-stabilize the plane.
Airliners have the fully automated autopilots with 3 axis plus throttles. Even then the planes can't take off by themselves, though if suitably equipped they may be able to land.
If you read the Tesla blog post that the (biased?) article is talking about, instead of the (clearly biased) summary, you'll see that Tesla actually points out that the US fatality rate is 1 per 94M miles (though the blog post does also mention the 60M worldwide number). Blame CNBC for picking the worse comparison point (and the submitter for choosing that article, I guess), not Tesla. What "they" are you talking about here?
There's no place I could be, since I've found Serenity...
Nah, it's not self preservation, it's the car. If I die, I'm dead, shrug. If I wreck the car there's a ton of insurance paperwork, I've got to sort out a rental replacement, I need to search for a new car, there're delays, there's hassle, it's expensive.
Easier to just avoid having an accident.
The way you want autopilot to improve is the way a human improves. First he's scared by a near miss. Then he tries not to make the same mistake again. My point is that you want the autopilot to be improved by examining near misses, not crashes. Fatalities are not required unless you're very unlucky.
Maybe it does make some sense. What if, for example, the autopilot drives in the most energy-efficient way? Of course people can learn that as well, but if driving in a certain way squeezes some extra percents of range, this could be important for EVs.
Ezekiel 23:20
Autopilot: Autonomous Piloting.
Autonomous: having the freedom to act independently.
Piloting: be the pilot of (an aircraft or ship).
That suggests it can run without intervention or monitoring until it alerts the occupier that it requires attention. In fact, Autopilot on planes are attended because it's required. It's in the air so the pilot can often "sense" when somethings wrong. Can you sense when you're about to hit a truck trailer?
Calling it autopilot is the absolute wrong thing to call it. It should be called "driver assist". Anything else is disingenuous.
According to people I know, it's much more relaxing.
Why couldn't it see the tires, undercarriage, and side reflectors?
Probably for reasons similar to why people sometimes don't see such things. Sensors are imperfect just like our eyes. They work very well under some conditions and very poorly under other. It's generally not hard to find a corner case where a given sensor can't see something very well. As long as you can control the conditions that isn't a problem but out driving in the real world you can't control the conditions so it's really hard to get a sensor suite that will cover all the numerous corner cases.
Here's the uncomfortable truth about automating cars to drive themselves. People are going to die. We can't always predict where the corner cases are and the only way to find them will be through cars getting in accidents. There is a saying in the navy that "The rules are written in blood". People had to die to find out what worked and what didn't and to develop the rules they use. You deviate from those rules at your peril. We don't know what the rules for safe automated driverless cars are yet and the only way we are going to find out some of them is for some people to get hurt. Some will die so that many may live.
Having the car on autopilot, but requiring the driver to "pay attention" and be ready to take over within seconds is the worst combination possible.
Not in all cases. We've had cruise control for many years which is a crude form of autopilot. It's been my experience that (surprisingly) it doesn't result in paying less attention to the traffic around you. If anything my experience has been the opposite. I tend to actually pay as much or more attention when the cruise control is on. I've spoken with other drivers who have experience the same thing. What you are saying has merit but there is some nuance there too. I don't think it is quite as simple as autopilot = less attention though that will be the case sometimes.
What is the reason one would switch the car to autopilot? Most likely so that they can be less attentive to the road.
Physical comfort for one. I use cruise control because keeping my leg locked in a single position for long periods is physically uncomfortable. I could see someone using greater degrees of automation for similar reasons. Again, you are not wrong but there is nuance here.
Autopilot still has a far better safety record than human drivers.
There is insufficient data to take that claim seriously at this point. Autopilot features are promising but need a LOT more miles in real world conditions before it is safe to make generalizations like that. I know a lot of people have high hopes for autopilot (myself included) but let's not let our expectations get in the way of scientific evidence.
One thing I do feel comfortable stating is that more people are going to die before autopilot features become truly safe. There are lots of corner cases that we're going to have a hard time predicting and we'll only learn about after some accidents occur. This is the case with any new transportation technology. Airlines are very safe these days but early on they were significantly less so because there were problems we didn't know about yet. Airline windows are rounded because we learned the hard way about stress fractures from sharp cornered windows which wasn't obvious at the time. People had to die to learn that lesson. This won't be the last time someone dies making autopilot safe.
There seems to be big flaw in the design of the trailer that allowed this to happen.
In the UK HGV trailers are required to have side and rear run-under prevention to stop this very thing from happening.
http://www.transportsfriend.or...
If an animal crosses a highway usually the first car can avoid it or break in time, it's the ones coming after it though that crash, often in a huge pileup. Driverless cars though would all stop as soon as the first sees the animal (there is no point having all-driverless cars and not networking them). While you can't avoid everything, you always get best possible outcomes.
"The radar saw the truck's trailer, but misidentified it as an overhead sign, because it was so high off the ground."
The radar+driver AI misjudged the hight of that thing in spite of knowing angle and distance to it.
The big automaker lobbies are behind this - paying the government to punish Tesla motors at every opportunity, because god forbid you innovate existing industries into obsolescence.
Big Auto and Big Oil will never allow Tesla to be a long-term success.
According to the article this was not something the driver could see and avoid,
Yeah I read that and my bullshit detector went off the charts. "There was no way the driver could've possibly seen or avoided a tractor trailer"? No way at all? Really?
Sounds an awful lot like an attempted cover-up by the manufacturer IMO.
All Tesla has to do is simply classify each Autopilot as a corporation, which under our current judicial rulings, would afford it even more rights than any of the meatbags either inside or outside of it.
If he had noticed the truck, he presumably would have applied the brake.
As anyone who has taken their eyes off the road for half a second knows, it only takes milliseconds to get into a situation you can't get out of.
So he could've seen it, but by the time he realised the autopilot hadn't seen it, even if the reaction was only half a second, it's too late to avoid collision.
..when my mind was wandering when the guy in front of me decided to stand on his brakes. I think it's good stuff.
You should just try paying attention. It has worked well for me.
Something I'm not hearing is at what time of day it was and what direction the car was going. Something from my early driving days that impacted me hard was driving through my neighborhood (tight, narrow, curvy streets with both sides of parking) in the evening and the sun was right infront of me. Speeds at about 10~15mph (military housing) and still only had about 2 seconds to react to a pickup truck coming out of the glare. Literally, the glare was so bad it was like somebody walking through a heavy bead curtain infront of me.
I was still going too fast for the conditions. My speed should have been dead slow, but the other guys conditions wouldn't be the same as mine in this case and thus may have been driving.
The radar saw the truck's trailer, but misidentified it as an overhead sign, because it was so high off the ground.
Then it's time to fix trucks and trailers. In the EU, those trailers are required to have an underride protection device. That would also have been clearly visible.
It still has to work with human psychology.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
Cool! Do you think it's just as infallible as their 'Maps' navigation system? Cause that would've killed me if I'd followed it blindly.
And do you think it's more reliable than their Android phone OS? Cause that regularly spontaneously reboots on me. I wonder how that would look on the highway...
If they can make good enough technology to drive you around safely, why can't they get the simple things right?
Sure, just be perfect. Great plan for humans.
Socialism: a lie told by totalitarians and believed by fools.
However, the driver may have noticed the truck if he was actually driving. Since the car was on autopilot, the driver probably was browsing the net or looking in directions other than the road, after all, the car is driving itself, there is nothing else to do.
Cruise control only keeps your speed constant, you still have to pay attention to the road or will drive into a ditch when the road turns.
And autopilot will steer the vehicle but like cruise control it is not sufficiently robust to permit you to stop paying attention at this point. People will forget that fact at their peril.
My car has a spring loaded accelerator pedal, where (maybe by coincidence, maybe by design), the spring can support the wight of my foot at any position of the throttle, I can essentially rest my foot there.
You still have to leave your leg in a fixed position so it doesn't really solve the problem. My father has had knee surgery so keeping his leg in one position for more than 60 minutes or so is quite uncomfortable for him. It's not about pressing the pedal, it's about moving the joints even if only slightly. My knees are fine but after 8 hours of driving my joints will start to ache too. Cruise control relieves this problem. I could see other autopilot features doing the same in different ways. It would be nice to one day be able to change seat position or recline or even lay down.
However, once you turn cruise control on, where do you place your foot?
Somewhere on the floor where I can reach the accelerator or the brake quickly if needed. Haven't you ever used cruise control?
If you needed to suddenly brake (moose on the road etc), will it take more or less time to slam the brake compared to if you kept your foot on the accelerator?
Roughly the same. It's not like you put your foot on the dashboard. It's still close by. In fact you actually can position your foot to be closer to the brake than if you had your foot on the accelerator pedal if you want.
This morning's NPR coverage of this story started that the driver was watching one of the Harry Potter movies while the autopilot was engaged. No wonder they didn't see the truck.
The data is insufficient to show that autopilot is safer than humans, but is sufficient to show that it is less safe than humans? Seems like a double standard.
Who said it is less safe? Certainly not me. We don't really know either way at this point. We SUSPECT it will eventually save lives and the early data seems to be pointing that way but we are just at the early stages of fully robotic driving. We have some data but the technology simply hasn't gotten into the public's hands to see what will happen on a large scale. We just don't know how it will play out yet. Any statements about its safety or lack thereof at this point are mostly pure conjecture.
The truck wasn't crossing a "street", it was making a left turn across a divided highway at a grade crossing intersection in the middle of nowhere, with no signal light. That's highway as in full speed oncoming traffic, and also as in left turning traffic has to yield right of way. How was the Tesla impossible for the truck driver to see? The truck driver didn't have right of way and should have waited for a gap in traffic before turning.
Meanwhile, this big thing suddenly crosses the path of the Tesla driving at highway speeds. Even if it had seen it, what are the chances that there would still have been an accident anyhow? Sure, it might have been able to slow down enough to "only" hit the rear end of the truck and not get capped, but that's a rather complicated sequence for mere silicon to comprehend.
#naabhaprzrag, #sverubfr-000, #agi-fcbafberq, negvpyr[pynff*=' negvpyr-ary-'] { qvfcynl: abar !vzcbegnag; }
As for self drive in general, prepare to see a lot of incidents like this. Tesla's features are quite limited and basically amount to smart cruise control and lane tracking and even they haven't got that working. The more complex, autonomous solutions are going to face far more complex realtime scenarios where the car will do something unexpected resulting directly or indirectly in an accident.
Except he didn't "get into" the situation. The truck made a left turn across rural highway traffic.
#naabhaprzrag, #sverubfr-000, #agi-fcbafberq, negvpyr[pynff*=' negvpyr-ary-'] { qvfcynl: abar !vzcbegnag; }
Apparently you missed the second half of my post, you can't claim that there is insufficient data while making such confident predictions about the future.
Sure I can. Every single major transportation technology we have ever developed has had some number of fatalities during its development. Every single one. There is no reason whatsoever to believe that self driving cars will be any different in that respect and we already apparently have one fatality already. I couldn't begin to tell you how or when or how frequently they will happen, merely that they will. One hopes that the number will be a tiny one.
If autopilot is safer than human driving already then what is this nonsense about the number of lives that will be lost in making it 'truly' safe?
It's not nonsense because those are separate issues. There WILL be lives lost developing autopilot and I would confidently bet large sums of money on that. This is a separate issue from whether or not is is safer than the current alternatives. Dont' conflate the two. Self driving cars could prove to be immediately safer from day one than human driven cars. (we just don't know for sure if they are yet) Separately there will be some indeterminate number of fatalities during the development of self driving cars until the technology matures and it is likely the number will never go to zero. Hopefully this will be a small number but I'm (sadly) entirely confident that it will be greater than zero. What we need to see is that fatalities from self driving or auto-piloted cars is less than fatalities from human driven cars. That would be a success.
You know, a country with money's a little like a mule with a spinning wheel. No one knows how he got it and danged if he knows how to use it .
Heh-heh, mule.
The name's Musk, Elon Musk. And I come before you good people tonight with an idea. Probably the greatest—Aw, it's not for you. It's more a China idea.
Now, wait just a minute. We're twice as smart as the people of China. Just tell us your idea and we'll give you subsidies for it.
All right. I'll tell you what I'll do. I'll show you my idea. I give you the Tesla Autopilotl!
I've sold autopilots to Plymouth, Oldsmobile, and Studebaker, and by gum, it put them on the map!
Well, sir, there's nothin' on earth like a genuine bona-fide electrified one-car autopilot! What'd I say?
Autopilot!
What's it called?
Autopilot.
That's right! Autopilot!
Autopilot, autopilot, autopilot, autopilot, autopilot...
I hear those things are awfully new.
It's safety tested just by you.
Is there a chance the car could crash?
Not on your life, now splash the cash.
First adopters must be brave...
They won't be given early graves.
Will this venture fund new green jobs?
No, good sir, I'm the new Steve Jobs.
We've killed off our whole space program.
Fund my SpaceX, my good man.
I swear, it's the country's only choice! Log in to PayPal and raise your voice!
Autopilot
What's it called?
Autopilot
AUTOPILOT!
But the economy's still all fucked and broken.
Subsidies, this man has stolen!
Autopilot
Autopilot
Autopilot!
Autopilot!!!
Auto...*CRASH*
Is it really too much to ask for you to read all the words in my post before responding?
Is it really too much to return the favor? I understood what you said just fine. You failed to comprehend the argument.
What you haven't shown is anything to back up your original vacuous statement: "more people are going to die before autopilot features become truly safe."
Exhibit A on your failure to read what I wrote. The argument is that EVERY transportation technology we have ever developed has resulted in fatalities. All of them. There is no reason to believe self driving cars will be any different especially since they have already have a fatality on the ledger. If you think that is "vacuous" then you have failed to understand the argument.
Think more, write less.
Comprehend more. Post less. We're done here.
Since you are still having difficulties, I've fixed your original statement for you:
"more people are going to die before any transportation technology becomes truly safe."
The car simply forgot to yell: " DUCK!"
Only half joking as it seems that would have worked in this case as the car mostly survived the collision.
when will cars relearn that critical lesson from aircraft?: "make sure you increase the cognitive load on the pilot as much as possible to keep them from daydreaming".
Fatigue ... wait, nope, they don't have the range for that. I'm a commercial pilot. I use autopilot routinely. However, I deliberately maintain situational awareness the whole time because, unlike arrogant millenials, I study my profession enough to know that automation will and routinely does fail in situations that are poorly predictable. I think lane assist is a great idea, as the automation gives greater awareness to the driver. Auto-braking is something I use routinely on airplanes and is a great idea in cars. However, the confusion between "auto-pilot" and autonomous cars is a deliberate deception by Tesla's advertising arm, and I doubt that the design and software behind Tesla's autopilot has had the same level of design rigor and oversight that does the autopilot on the A320.
Look up. Thousands of pilots do this every day. However, being professionals instead of asshats, we maintain awareness of the situation around us instead of watch Harry Potter on the highway.
Perhaps Tesla or any Auto manufacture can integrate similar product such as Furious S8 - A 8 cam dash cam (shown below), they could have figured out what went wrong (the culprit). There needs to be a way to capture/record the surroundings to ensure any new technology is safe and if an incident does occur it can determine rather caused by human error or the technology. Condolences to the family, and RIP Joshua! https://youtu.be/dTTP5SKc1Fk https://youtu.be/b9K6HmCb3Bg https://youtu.be/JGkQzWfbW3s
Look autopilot or not the truck failed to yield to the Tesla that had the right of way.
It's entirely possible that neither the driver or autopilot had sufficient warning to apply the brakes in time.
This would be a clear case of truck causing a fatality and wouldn't be national news except it was a Tesla.
Even still, "We will drive for you until we don't, and in that case fuck you" isn't a very good marketing philosophy.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
And if the image was so washed out that it couldn't make out the outline, then because it couldn't see clearly, the car shouldn't have been moving so fast. {...} It would also help to upgrade the camera to one with a wider dynamic range and/or more resolution so the image is less likely to get washed out again.
That or add an entirely different class of sensors.
There's a reason why Volvo supplements their forward facing camera and long-range radar with a LIDAR, same with some Mercedes.
Or that for their city safety (only slow speed in city collision avoidance. No lane detect) the little VW Up counts on a LIDAR instead of a camera.
A camera has the same limitations that human eyes have (and actually even worse, due to lower dynamic range and different resolution*)
If you cannot see it, chances are that the camera won't be able either.
On the other hand, because they are self-illuminated, and because they return a different kind of information (depth, instead of shapes of colour) a LIDAR might cleary perceive the contour of a white object on a white background.
It has its own limitations (highly reflective / no reflective at all surfaces), but would add new useful information to supplement what is provided by the current Autopilot camera (which might miss a white-on-white object like this trailer) and the long range radar and forward-facing sonar (which are usually mounted too low to be useful in this specific situation**).
e.g.: the LIDAR of my parents' Volvo has correctly seens object obstructing high. like a low blaconny under which I'm paking the hood of the car.
Though it has sometime a little bit pannicked (highly reflective barrier of the tool both on some french highway that is slowly raising + LIDAR = over-reaction, car is already in "ready to do an emergency break" mode although there's enough clearance).
It violated the Basic Speed Law just as surely as if it had been driving the speed limit in heavy fog, and that's a programming error.
That evokes 2 comments for me:
- Yes but that require the car being able to realise that it doesn't see.
It's not behaviour to react to something it's seeing, it's being able to predict that it's not seeing something, which is a much harder problem.
The car might simply NOT realise that it's missing something.
Most of the car monoscopic systems that I've seen demoed in detail (there are tons of example on Youtube, even some code for OpenCV floating around) use some simplistic projection to infer the 3D structure of what they see.
Then they detect 2D shapes on the pictures and use them to infer presence absence of edge in the projected 3D environment.
(Such system can't make the difference between "there is no edge 10m in front of me" and "there an edgeless mass right in front of me obstructing the view")
I don't have a way to know if Autopilot's monoscopic camera works this way** (or if it does pick extra 3D cue due to motion tracking and parallax/perspective).
But it will be definitely at a loss here.
Compare that with humans which have a *pair* of eyes, a stereoscopic system able to provide depth perception (unless the image is completely feature less - and then you CAN infere that there's something wrong) (the wrong usually being expressed as some form of sea sickness).
Same with cars using stereoscopic cameras (I've seen it on some Mercedes, and several japanese brands like Mazda), they'll definitely either have a complete 3D representation of the environment (not infered by projection, but actually based on image-pair correlation) or will lack one which is an information that there's something deeply wrong (e.g.: very desne fog).
Slightly similar with cars using LIDAR (though it has a bit better visibility in case of fog by virtue
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Sure, just be perfect. Great plan for humans.
Simply paying attention while operating dangerous machinery is a reasonably normal skill I would have thought....
Except he didn't "get into" the situation. The truck made a left turn across rural highway traffic.
We've all been in similar situations, when someone cuts us off or pulls out in front of us. Awareness and reaction times are key skills required to avoid such incidents. And every millisecond counts in such cases.
We can't know the precise detail, but we do know that large trucks don't move that quickly. For the car to go under the trailer (ie the back half of a long vehicle) at a speed to shear the roof completely off would mean the truck got at least halfway across the road with no reaction from the driver.
And we know the driver was operating in an "Autopilot" mode which suggests full attention and awareness was not a priority for him.
So the moment you turn on autopilot, or play with the radio, or use your phone, or eat food while operating dangerous machinery, you are getting yourself into a potential situation.
The truck might be the root cause, but from what I can gather from the nature of the accident, this driver's actions didn't his improve his chances.
RE: If I exclude parking, it is cheaper for me to drive than to take a bus or train
Nonsense. No way you calculated all the costs. Insurance? Amortized deductible? Amortized risk of losses greater than your insurance? Maintenance? Depreciation?
If you calculate all the costs driving is far, far more expensive. It is much more reliable and convenient. If it was actually cheaper there would be far more cars on the road, which would be bad for those who can afford to drive, so the costs stay high, in part because of that exclusivity.
My camera knows when it's missing details such as the edge of the "white side of the tractor trailer against a brightly lit sky". When the histogram clips either to the left or the right, a self-driving car should determine whether each blank area of the image is big enough to conceal a hazard (unlike glare on a piece of chrome, for example),
Yup, that's a possible strategy, that apparently wasn't added to the software yet.
On the other hand, I will only react in badly lit condition (white vehicle in an over-lit scene / black vehicle in an underlit scene), it won't necessarily react correctly in case of fog (everything is washed out even the mid-range) or snow (the scene is just a huge cluster fuck of motion noise).
But yeah definitely extra strategies need to be added. And I stand by my previous comment, an extra "sense" must be added for this specific range (like a lidar)
and if so, slow down and be prepared to stop.
- a car (not necessarily Autopilot, even simpler tech like collision avoidance is *always* ready to break).
- you've successfully demonstrated that the car can determine that it can't see a region, what should it be "ready for" if it can't see there ?
- the correct course of action (like with humans and with simpler collision avoidance) would be to adapt the speed/braking to the visibility range.
i.e.: if Autopilot determins that there's a zone it can't distinguish 50m a head, it should automatically slow down to speed where the braking range is shorter like 40m. So if suddenly an object "pops out" of the visibility range and is suddenly seen to pose a collision risk, the car can break in time.
- the car should also have a clear protocole to hand control back to the human.
I.e.: if visibility goes bonkers, to start slowing down quickly while sounding an alarm asking the human driver to pay attention to the road instead to the movie.
Speaking of which...
WHAT THE HELL ?!?!?
You're driving an experimental technology, and you're absorbed by a movie ?
I you want to completely forget about the road and look a movie, there are vehicle better adapted to that:
it's called "public transportation", we have plenty of it here around on the European continent...
Or you know, simply being a passenger on the ride, e.g.: car pooling...
Driving *alone* in a vehicle isn't terribly efficient (although in this case it's a Tesla which has been proved to be a little bit better to the environment, even if you factor in the construction of the vehicle and the production of the energy).
It's always better to share the ride, and if the passanger also knows to drive, that's an extra eye (a third one in addition to the human driver and to Autopilot).
I usually travel in this configuration when on vacation road trip: parents' Volvo got a collision avoidance system, I am the driver, and my passenger knows enough to whatch the road too, while the rest of the team parties hard on the back seat...
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
It is my understanding that Google cars can't drive at highway speeds. So yes, there would have been no accident because Google cars can't drive on highways.
According to the news, he was apparently watching a Harry Potter DVD. Truly an idiot.
The other options are only more "efficient" if you place no value on your time or personal convenience. You might choose to spend an hour on public transportation instead of a 30 minute drive to save $10, but I would usually not.
In several European cities, the situations is actually reversed:
- They are over crowded. Unless you want to get around at unusual hours, there is going to be traffic jams everywhere.
- Public transportation is usually circulating separately and isn't affected by jams (city trams and bus might just have separate lanes. Or the metro trafic is completely underground in cities that have them).
- Long distance train have high speed. Car speeds on highway are usually limited to 120 or 130km/h depending on the countries (...but not in Germany. Only local segments are limited, there's no general limitation), whereas the train usually drive at 150~180km/h (or higher depending on the local type of high-speed trains).
So depending on your destination, public transportation *might* actually get you there faster. (and much safer)
But on the other hand it costs a bit.
Same also for short in-city trips by bike:
- Lots of European cities have separate bike-lanes. (and in lots of jurisdiction, when there's no bike lane, bikes are supposed to form separate lines)
So bicycles aren't affect by traffic jams
- Altought a car's *top* speed (=the speed limitation in the street) is higher than a bicycle's, a bicycle's *avarage* speed ( = still close to the top speed) is higher than the speed of a car stuck in a traffic jam.
So, in summary, in several european big cities, not taking the car it the thing that actually saves you time.
(Or, alternatively, managing to have a work where you can shop up and go home outside of rush hours).
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
If so, how do we know he didn't notice the truck?
If he had noticed the truck, he presumably would have applied the brake.
or at least ducked