Tesla in autopilot mode crashes into parked Laguna Beach police cruiser
latimes.com
latimes.com
I can see the benefit to Tesla and future Tesla customers of essentially crowdsourced fleet learning. But what is the benefit right now if you strictly follow the rule of remaining alert enough to intervene at a second's notice?
In previous threads the best explanation I got from Tesla owners was that it frees you from the "details of physically driving" so you "can now supervise instead."[1] That just seems suspect. Supervising autopilot to the degree required to correct sudden mistakes seems, to me, to be probably very close to the mental load of the details of steering yourself.
Indeed. When reading Tesla's statement,
"The Palo Alto-based automaker, led by Elon Musk, has said it repeatedly warns drivers to stay alert, keep their hands on the wheel and maintain control of their vehicle at all times while using the Autopilot system."
...one can't help but think, "then WTF is it good for, if I still have to drive the car myself while using it?"
Supervising autopilot to the degree required to correct sudden mistakes seems, to me, to be probably very close to the mental load of the details of steering yourself.
I'd say the mental load is higher than manually driving --- not unlike being a driving instructor to a student driver who is OK most of the time but occasionally makes a possibly fatal mistake that you need to watch out for and quickly correct. Normally, a car will go in a straight line with no steering input, but the autopilot almost "has a mind of its own" and can steer itself in the wrong direction.
This is subtly different from cruise control, which lets you concentrate on steering and rest your foot on the brake so you can use it quickly if needed. Cruise control doesn't encourage lowering attention, but enhances it.
And why are Tesla burning karma by promoting this technology?
For SpaceX, self-landing rockets are a direct method of reducing costs.
For Tesla, sort-of-self-driving cars are a distraction from the core product and are tainting the brand. Stick to selling top-tier electric vehicles with human-driving, people will still buy them on merit.
To me it's am example of what happens when software developers run the asylum. You can do nearly anything in software, but should you?
Except if the rumors about turnover in the Tesla software team are true, the exact opposite is the case. The hot goss is that they're under pressure to meet unrealistic goals, not vomiting their own hubris into the product.
What has lead you to this conclusion? Autopilot is really just advanced cruise control. You can treat it the exact same way by leaving your foot on the brake and leaving a hand on the wheel. If people aren't using it that way, that is certainly something Tesla should try to fix. But I am just not sure I buy the argument that cruise control is inherently a safer technology than cruise control plus auto steering.
There's some threshold of interventions/minute that I think humans just aren't very good at. Which is why there are systems like Threat Image Projection, which turns baggage scanning into what is effectively a video game: https://www.rapiscansystems.com/en/products/rapiscan-threat-...
In my experience it’s LESS boring to drive with autopilot and I pay MORE attention to the actually important, potentially dangerous, things like what’s coming up (you’re actively aware of longer stretch forward, which also means you know you’re coming up to a stretch that autopilot won’t be able to handle well in advance and have plenty of time to disengage) and the cars around it (not watching lanes closely frees you to give all your attention to cars around you).
I engage similar system on my car before fiddling with the attention seeking touch screen, or when I need to give a quick look at the kids behind. Another is traffic jam: a lot of nothing going followed by a lot of things going on - focus on the wrong thing and you can rear end somebody. Minor benefit are some fuel saving if you are a "dynamic" driver, and it can make you more relax in stressful situation like the traffic jam above.
That's just to answer your question. I completely agree with you that the feature is overhyped, not worth paying much for it and actively promoting dangerous driving behaviour. However, it nowadays comes nearly for free as a consequence of other very useful systems: the whole set of emergency braking scenario (emergency, cruise control, safety distance, pedestrian, ...), line keeping assist, blind spot monitoring. It has its uses, if you remain aware of its shortcomings.
...in other words, the feature partially works around some of the problems they introduced themselves by using an overly complex touchscreen UI instead of the traditional no-eyes-needed buttons and knobs.
Autopilot basically adds autosteering to that. Seems to me that could lull you into more distracted situations that collision avoidance is unable to cope with.
Note: autopilot and collision avoidance both came out around the same time, so it's far from clear that autopilot has resulted in a reduction in accident rates, or if it was mostly just the collision avoidance.
intuitively, the load seems higher. this is because your brain must now include predictions regarding an additional third party who is not there in cars without autopilot or with autopilot off. it’s similar to how i am more stressed and tired after riding in the passenger seat with a less experienced driver at the wheel than i am at the wheel myself.
Of course, there's zero data about how many accidents happen because drivers were staring at their phones, so you rarely see articles screaming "<non electric car> driven by an idiot staring at phone screen crashes <into what ever catches attention>".
I was skeptical about Autopilot from day one, but one test drive convinced me that this may not lead to zero accidents, but it certainly is way way better than vehicles that crash when driven by distracted drivers
Living in a high traffic area I would appreciate the time I don't have to pump my damn foot up and down every 10 seconds for hours straight while waiting in traffic.
You have an extra failsafe for those moments where you fail to remain alert. People keep treating these incidents in isolation and failing to account for the fact that human drivers make exactly this kind of fuckup every hour of every day, and lives are lost. Every day.
You get that right? Cars. Kill. People. Every day.
Tesla is shipping systems in the absolute infancy of this technology that are already statistically indistinguishable from human drivers as far as safety goes. And the luddites here want to throw all that in the garbage because of "what is the benefit right now".
That’s a hell of a claim, I’d love to see some serious citations on it.
For comparison the overall casualty rate for the average car in the US is somewhere near 12 deaths per 1 billion miles driven.
I wouldn't say that Tesla's have proven safer, but I think saying they are "statistically indistinguishable" from human drivers is relatively fair at this point.
> Unfortunately, our website is currently unavailable in most European countries. We are engaged on the issue and committed to looking at options that support our full range of digital offerings to the EU market. We continue to identify technical compliance solutions that will provide all readers with our award-winning journalism.
On the other hand, I'm wondering how much I want to read an article from a website where they must track me to when I just want to read something...
http://archive.is/http://www.latimes.com/local/lanow/la-me-l...
And then sharing, trading (to expand the profile), and selling it to other companies? Really nasty stuff.
You'll see some sites have quite an extensive menu of opting in to various tracking etc. For example Engadget or TechCrunch have a popup asking you to give consent, or not, before using the website. This is how it should be done.
Then there's other websites that shows you all it's tracking and ask you to just accept, there's no option to refuse, which is against GDPR, as implied consent is no longer valid.
And then you'll have websites like latimes that just goes "fuck it, we'll just block 500 million people in EU instead!"
https://www.google.com/search?q=Tesla+in+Autopilot+mode+cras...
It may mean they are doing obscene tracking too, but you are making ideological assumption if that’s your only hypothesis.
Blocking EU residents as a GDPR protection is dumb.
And again using or not using the internet as intended isn’t the issue here but rather if they intend to provide services to the EU or not a message saying EU users should leave in the same manner as age conscent is verified is technically sufficient to indicate that you do not provide service to EU residents, geoblocking the IP also sufficient even if it’s not perfect.
How else should they do it? I concede it is not foolproof nor ideal - but for websites which you've not provided your location to yourself, how is it possible for them to determine whether or not you are in Europe?
Every time I ask this question, the response is "radar as a technology has too much noise, not accurate enough to detect stationary objects".
But it DOES detect stationary objects. Otherwise no one would even turn on Tesla AP at all.
Why is it so hard to detect stationary objects in front of you?
Tesla's fans always defend the failure of their AEB by pointing out flaws in radar, while entirely ignoring that Tesla's vehicles have contained front facing optical cameras for years now.
The Tesla shouldn't have hit the divider in the previous accident, the optical (visual light) camera should have seen it, even if radar didn't. Why doesn't Tesla's system work as well as Sabaru's Eye Sight?
How else would they do it than incremental?
How about... not?
I'm not being disingenious here.
If there's autonomus tech, then I find it scary that it might change its behaviour. You see, I've learned how (say) Autopilot responds in certain situations. I know when to trust it, and when not to.
If this behaviour changes (perhaps even if it changes for the better!), that's a very dangerous thing. Some of my experience is now invalidated, but I don't know which part. But it mostly works as before, which gives me a false sense of security.
Changing the driving behaviour of a vehicle, especially in totally un-obvious ways, is super dangerous, and such changes must have very strong justifications.
Just think of that poor guy whose Tesla drove into the divider. Oh look, it behaves just like before -- except for its lane following behaviour.
Still, this answer doesn't quite satisfy, because it ignores the very strong signal that exists right before a crash, which crash avoidance technologies do (seem to) use.
Note: Tesla also has a collision avoidance system, which is actually on all the time and a distinct feature from autopilot (although it uses the same underlying sensors).
as distance increase, the angular size of an object decreases inversely.
a radar requires an increase in power both to see at a longer distance and to see a smaller object, because the signature reduces with 2*distance^4
( this is an oversimplification to help introduce the topic, you can see the details here http://www.radartutorial.eu/01.basics/The%20Radar%20Range%20... )
a static object is the worst case scenario: you can't use its doppler signature because it's not moving against the background and the radar becomes less efficient the moment you need it the most (at high speed)
as an aside, given the problem statement it should be possible to find a maximum speed that allows for timely detection and stopping.
chances are it's still slow.
I wonder why the radar/lidar isn't helped by an ir coincidence rangefinder.
This could only be said by someone who has never worked on the problem. I don't get it, why is programming so hard, you are just telling the computer exactly what to do?
Obviously this doesn't seem to be the case but maybe you can enlighten us why this is hard?
Maybe radars can't be pointed at a narrow space, if that's the case why not use lidar instead?
A static object has a negative Doppler shift rate linearly in proportion to vehicle speed and a y-axis position greater than zero but less than four metres. What makes that difficult to detect or distinguish?
(Lidar really does matter, but also extremely rigorous testing.)
I mean that plastic bag could contain a rock or something, driving through it is just a bad idea.
it’s dangerous to stop in the middle of the road unexpectedly too
Even among non-Tesla cars, whose manufacturers don't try to "mislead", 71 percent of people believe automatic emergency braking can avoid all crashes [0]. Is this percentage higher or lower for Tesla, and is that difference warranted?
[0] https://newsroom.aaa.com/2016/08/hit-brakes-not-self-braking...
[1] Quote from above "When traveling at 45 mph and approaching a static vehicle, a scenario designed to push systems beyond the stated limitations, the systems designed to prevent crashes reduced speeds by 74 percent overall and avoided crashes in 40 percent of scenarios. In contrast, systems designed to lessen crash severity were only able to reduce vehicle speed by 9 percent overall."
Oh, and Tesla does disable Autopilot if you ignore the hands-on-wheel warnings.
[1] I think it's pretty debatable, and as far as I know it's still something of an open question to determine the minimum amount of hardware required for a fully autonomous car
I mean there is a warning telling you that it's not self driving. People are dumb, that's it.
The system exists in a bad gray zone — one that works well enough most of the time you stop paying attention. But it’s not full proof, so you get the worst of cognitive bias here.
Personally I think this faux self driving (literally falsely named auto pilot) should be outlawed. We cannot depend on people to take over when the car generally is handling the situation most of the time. Either manual or fully automatic — this in between will continue to cause too many bad accidents.
I'm far from an expert on the subject, but I assume real data must be acquired for that to happen.
Does Autopilot actually make the problem worse, or is it just more newsworthy?
I cannot find any current (2018) car that crashes into the parked car.
(I only spot-checked prior years). I found one in 2014 (audi q5) that hits the car at 1mph when going 12mph.
http://www.iihs.org/iihs/ratings/vehicle/v/audi/q5-4-door-su...
By contrast, the 2014 subaru outback avoids collision even at 25mph.
http://www.iihs.org/iihs/ratings/vehicle/v/subaru/outback-4-...
I can also find no data that suggests any other TACC is crashing into vehicles (no NTSB studies, investigations, etc) at all.
So yeah, i'm going to conclude, based on what i've seen so far, that autopilot is part of the problem.
If you have data that says autopilot is not worse, i'd love to see it.
A lot of police officers were being struck because you tend to steer in the direction you stare.
I recently mentioned self-driving cars to somebody and her response was "Oh, one killed a woman recently, didn't it?" That was about the extent of what she knew about them because she watches the news.
What I don’t get is all the HN commenters acting like a few crashes proves that Autopilot is irredeemably dangerous.
Of course, as usual, their obstacle detection failed to detect a stationary obstacle that didn't look like the rear end of a car in the same lane. We need a minimum standard for a vehicle which takes automated control of steering and braking. It must reliably stop for obstacles.
Today I read Tesla is getting into the insurance business because premiums are getting out of hand. It makes some sense since internally they’ll have more data they can use to deny claims.
It’s interesting how reality plays out.
https://electrek.co/2018/05/29/tesla-insuremytesla-insurance...
Claiming that the driver is still responsible for driving the car, while using a system designed to encourage the "driver" to not pay attention is a dumb idea, and all I can see is this kind of thing leading to regulations the delay actual self driving cars.
I'm not a self-driving car fanboy or anything. I don't think it's just around the corner, but it's clearly going to happen /eventually/ and I can't imagine it being more dangerous than regular drivers. But this kind of "self driving but not" features feel like the sort of thing that calls to government agencies for regulation, and sufficiently "dumb" mistakes from these systems seems like the sort of thing that triggers a reactive over-regulation.
> Audi development engineer Thomas Müller and his team sent “Lu Ban” and “Kong Ming”, two piloted Audi A7 Sportback prototypes, on a very special kind of test drive.
This is a very far cry from what Tesla and Musk are doing. For starters you can't even buy the car yet. If you look for information on "Audi Piloted Driving" [2] you can see they're just as cautious as you'd imagine a German company to be:
> In 2017 Audi will introduce what’s expected to the world’s first to-market Level 3 automated driving system with “Traffic Jam Pilot” in the next generation Audi A8. The system will give drivers the option to travel hands-free up to 35 mph, when certain conditions are met prior to enabling this feature — for instance, the vehicle will ensure it is on a limited-access, divided highway.
Note that it will only work on highways, but also only at non-highway speeds. Stop and go traffic.
> In 2020-2021, Audi will introduce a Level 4 “Highway Pilot” feature—technology similar to what has been demonstrated in our concept vehicle “Jack”—that offers hands-free driving at posted limited access highway speeds in which the vehicle can execute lane changes and pass cars independently.
Seems pretty cautious to me.
[1] https://www.audi.com/en/innovation/piloteddriving/piloted_dr...
Tesla is outlier mostly in the disproportionate amount of flak it takes.
http://www.autonews.com/article/20160801/RETAIL03/308019956/...
This is just like Fox "some people say" - Fox people say! Or Javascript Bitcoin miners - Monero is not Bitcoin! It's snowing so there is no global warming - It is "climate change" but the entire globe is warming! This hurricane is just weather, not climate change - it is the 10th 1 in 500 year hurricane we have had this month!
I could go on. Sure, talking about the future self driving features (and selling them) is poor form to avoid this from Tesla, so they aren't helping - but they have always been clear that it is not self driving. It is the listeners/ readers that are incorrectly connecting the dots.
You have to go way down to "Enhanced Autopilot" section to read the smaller text that mentions that its subject to regulatory approval, and that the driver is responsible.
If you don't think that's misleading advertising, I don't know what is.
They should also retrain their sales staff because on my test drive the saleswoman specifically told me the car could drive itself and that I should take my hands off the wheel to test it out.
They should also retrain their CEO who keeps talking about "self driving" in reference to existing cars:
https://www.tesla.com/blog/all-tesla-cars-being-produced-now...
> "All Tesla Cars Being Produced Now Have Full Self-Driving Hardware"
Those people being Tesla's marketing team, presumably.
> they have always been clear that it is not self driving
That is very far away from the truth.
> what I say is always what I mean, and if you misunderstand, it's your fault.
Linguistic pragmatics 101: communication is a two way street.
A spokesperson pointed out that the owner's manual reads, “Autosteer is not designed to, and will not, steer Model S around objects partially or completely in the driving lane”.
I haven't seen the encouragement to not pay attention. In some ad on TV, maybe it was for a new Cadillac, the driver opens a soda with both hands off the wheel, and that seemed to be as far as they were willing to go.
Cadillac has an eye-tracking system that deactivates the self-driving features if you aren't watching the road, and their system is limited to specific highways they have mapped.
Tesla has a ridiculously inadequate attentiveness control, and they do little to nothing to remind people who regularly fail to keep their hands on the wheel. They don't nag people because they want them to believe the car is more capable than it actually is.
And, I say that as a proud owner of a Model S with Autopilot v1.
I don't even get the point of that. Why even take one's hands off the wheel(or if knee-steering, one's foot off the pedal) if they can't even take their eyes off the road? I haven't driven one of these vehicles so maybe I'm missing something. It seems intuitive that automation would free us to do other tasks, but I'm not seeing that in these early "self-driving" implementations. It's more like these car companies are actually selling these cars as experiments, at the expense of people's wallets and possibly their lives.
They should be legally required to call it something like "lane assist".
An autopilot maintains heading, and that's all. Many sailors are not techies --- an autopilot on a boat will run you right into an obstacle.
Self driving vehicles have been used at Rio Tinto's Pilbara mine for a couple of years. Sure, it's on private land and the complexity of the task is far smaller than for public roads, but it's happening now.
http://www.abc.net.au/news/2015-10-18/rio-tinto-opens-worlds...
Personally, I don't see mining operations as anything remotely close to public roads autonomous vehicles because you have a controlled environment in which paid workers that have been trained to interact with equipment and be aware of its limitations and dangers are performing the said interactions in a strictly constrained manner with the autonomous equipment clearly visible as such and while protected by all kinds of insurance and work-specific regulations.
There's none of that even in the testing stage of current "autonomous" driving vehicles.
All vehicles can safely communicate with each other and are controlled from central location.
It is automated driving but not even close to "self-driving".
Accidents still happen up there.
https://www.youtube.com/watch?v=5WZCweaakOg
(*video may not explicitly of an autonomous truck, I'm not entirely sure, but just using it as an illustration)
This is a lot like rocketry. A rocket that meets 95% of its design requirements explodes, crashes, or fails to reach the proper orbit. Its damn near 100% or failure. Too bad Elon can't see this.
I think Tesla's obsession with autopilot is misplaced. Give me a solid reliable affordable EV with good range. Autopilot would be nice but only if it works. I can wait for that. But a solid EV is something I'm actually more excited about believe it or not.
I'm sort of afraid Tesla will throw the whole game in pursuit of self driving tech.
Eventually, you're not marketing to consumers but to fleet owners.
Outside of that, he's in the grey area where he can build product excitement and fall back onto the "Keep your hands on the wheel" type of fineprint. I agree with those saying that it's far from ideal.
I'm not saying that's what happened, but it's certainly possible. The mere fact that it's possible to me casts doubt on this entire design philosophy.
> "Tesla has always been clear that Autopilot doesn't make the car impervious to all accidents, and before a driver can use Autopilot, they must accept a dialogue box which states that 'Autopilot is designed for use on highways that have a center divider and clear lane markings,'" a Tesla spokesperson said in an emailed statement.
From the picture at the top of the story, it is clear there is no divider on the stretch of road where the accident occurred.
Maybe it should detect when it is being misused, and start blaring loud noises, do its best to make the driver carsick, or simply immobilize the car.
https://www.youtube.com/watch?v=AO33rOofFpg&t=2m18s
Also takes his hands off the wheel while using it in a manner that if someone else did and died, Tesla would slam them in the press for doing so, using dodgy-from-a-privacy-perspective and potentially-inaccurate (frequent reports of hands on wheel false negatives) telemetry.
uh, isn't a self-driving car you have to drive an oxymoron?? does anybody understand words??
He's been promising "self-driving" for years, except the damn technology isn't ready/safe enough for what people are clearly using this from. The manufacturer puts a warning label, but the usage model/UI is clearly flawed.
His engineers know this, which is why Tesla's self-driving group has had such a high turnover rate over the last few years. Musk keeps overriding them.
Not just the model/UI but the actual sensor suite. Remember, the Tesla that resulted in a decapitation had over-saturated cameras that couldn't see a tractor trailer. Of course, that's where the UI comes into play, because if the driver hadn't been watching Harry Potter, he would have seen the tractor trailer as well.
This was proven to not be true: https://jalopnik.com/the-ntsb-to-partially-blame-teslas-auto...
Maybe with AR the car could highlight the objects detected so that the driver can brake if the car doesn't see an object it knows about, but a head mounted display seems pretty clunky.
Reports at the time said he was watching it on a portable DVD player, not the Tesla screen. Not sure about the NTSB report someone just cited though. If that report really said he wasn't watching anything, I wonder how that rumor started?
It seems pretty clear to me that Volvo only has Pilot Assist in the first place as a method to gather data via telemetry for the next generation of self-driving vehicles, because it otherwise doesn't seem to add anything useful to the driving experience. Consequently, I haven't heard of any notable incidents where Pilot Assist caused an accident.
I haven't test driven Enhanced Autopilot but I see people on YouTube taking their hands off the wheel completely for much longer than the Volvo let me do. Tesla seems like they're pushing something half-baked out and representing it as something it isn't.
It’s not going to drive for you. But it does an amazing job of basically canceling out crosswinds when you find yourself driving down the highway. It makes a huge difference, it makes it much less tiring/frustrating.
Going slow on a windless day? Straight street? Not that useful.
For helping with very gentle curves or very annoying crosswinds it’s quite nice.
It may be, but is not necessarily “consequently”. That Volvo is not cool to write about plays a role. You don’t read about boring daily crashes either.
Self-driving tech has been there for decades. Convoy driving of trucks (one human-driven truck in front, several trucks following its lead) on highways was developed in the 90s.
On well-maintained infrastructures, automated driving is a problem that was solved before deep learning. The fact that we don't have an automatic lane-driving mode for highways as a standard feature in new cars is a political failure.
In 1997 a huge and successful demonstrator for an automated highway system was made. And legislator did not follow up by allowing on the road the systems they spent more than 10 billions to develop. [1]
In my opinion, a technology should be considered ready/safe enough if it is x times safer than the average driver. Zero accident is a crazy requirement. 10 times safer as human is much more reasonable but no politician will accept that.
And obviously it is easier to develop a safe system on some well known and well maintained road segments (e.g. highways) than on 100% of the roads and dirt roads on the North American continent.
So yes, the only way for an economic actor to get into there is to bend the existing rules, as rules for automated vehicles were never made. You can't sell a car saying "no need to look at the road, I'll drive for you" so you put a disclaimer, while making sure people understand that you actually made a self-driving car.
You can't sell a self-driving car and argue that your rate of accident is acceptable because the laws for such a notions are not there. A single accident in your self-driving car (that may be 10 times safer than a manual car) and you will be considered responsible.
[1] https://www.smithsonianmag.com/history/the-national-automate...
If you want to fix the problem, just charge him with attempted murder of a cop.
It is not really surprising when our whole system of existence has the stupidity and weakness of its inhabitants at its foundation. Point is, this will only get worse. Because any "progress" along the current lines asks the common man to be more stupid...
Outrageously false. These incidents are so rare that they have whole news stories written about every single one. How many crashes were there in Laguna Beach yesterday?
Tesla accident rates in aggregate are quite good. Under autopilot the data is noisy but still underneath the level of an average car (not as low as luxury vehicles in the same price range, though).
Claiming that it "isn't ready" is innumerate nonsense. It could clearly be better, sure. So could you. And no one is demanding to revoke your license, because you aren't a scary AI.
So we get to the problem of "if it can't do handle the specific obvious case where it should be absolutely superior" (I would argue an automated system should not have ever been able to make either the police cruiser or the 280 crash in any circumstances. Those were both in perfect conditions - clear weather, etc, etc), how can we trust it to handle any of the complex situations that a self driving car necessarily must handle?
The Tesla system solves this by simply saying "the driver is still driving, we're merely assisting", but that assistance leads to inattentiveness - that's human nature, you have to design around it - which should be fine, if the systems were able to handle that, but that clearly are not.
I would argue that given the configuration of the Tesla system they are less safe, not necessarily due to net accidents, but rather because of the failure modes of the accidents. Bumper to bumper accidents on a freeway happen all the time, even driving perfectly you can end up in them because not everyone else is driving perfectly. The problem with things like the Tesla system is that it's generally fine, except in those cases where it drives at speed into objects because it gets confused.
If your vision or mental state leaves you unable to safely account for large obstructions you aren't allowed to drive. Because it isn't safe.
Seems to me that it's premature to me making bold claims like this. Of cars costing over $100k, what's the accident rate?
What data are you basing this statement on? Or is this just your vague gut feeling based on reading about a few accidents in the news? I thought statistical data showed otherwise, that the technology was safe enough (ie. better than humans).
Edit: removed "debunked stats".
US interstate deaths are about 1 per 183 million miles, so while Autopilot may be safer, it's not as clear depending on the uncertainty of the 320 million mile figure. Also, German drivers on the autobahn have roughly half the fatality rate, which is better than the claim for Autopilot. All in all, I think an honest assessment would be that it's probably as good as an average driver in good conditions who isn't drunk. But rhetoric about saving 900K lives raises my hackles - I do not want to trust the source of this sort of PR with my life.
To align with your statement, I can state; "I really wish all people would stop this bullshit "auto pilot means it can drive itself and I don't have to do anything" nonsense.
I feel this is quite a critical analysis of an ambiguous product title; [1]'Telsa Autopilot (& Enhanced Autopilot)'.
I read the title; "Tesla in Autopilot mode crashes into parked Laguna Beach police cruiser"... and wondered how this article is even news..
Where, besides future plans for their vehicle product (Which, as suggested by Wikipedia - may be an independent product itself, and not an extension 'Tesla Autopilot / Enhanced Autopilot'), does Tesla's product in question indicate it's self driving? Autonomous Car? I do admit, the 'summon' feature of their vehicles makes this slightly confusing...
Additionally; what is the criteria to be 'self driving'? I assume; you are talking about 'auto pilot'..?
If so; this over-simplifies the entire scenario.
The very definition, and scope, of the title 'autopilot'/ 'automatic pilot' is far from specific. In fact; most of them are applicable only to [2]aerospace.
> Claiming that the driver is still responsible for driving the car, while using a system designed to encourage the "driver" to not pay attention is a dumb idea, and all I can see is this kind of thing leading to regulations the delay actual self driving cars.
This part of your post, I actually agree with. Having said that; I believe this is more so associated to an applied opinion of what a 'self driving' car is, and what we are seeing with these vehicles - and not because of any lack of technology advancements.
But again; I urge caution in applying your, or anyone else', opinion of what a 'self driving' car is.
FINALLY: I really wish we could start taking some responsibility of what we agree to. If the driver, even once, must agree to a clear scope of requirements to use the product - then blatantly disregards this; how, or why, is this even news?
We all accept terms and conditions of use, and privacy policies - with almost every product we use - and don't expect to be shown them every time we use that same product. I'm one of the few that read these documents - and more often than not; refuse to use the product. I made the choice not to agree.
There's other comments to this article that clearly indicate they believe a 'once off' popup is questionable at best - but I really don't understand this. Without a doubt; having to push a button to enable this Telsa/Enhanced 'Autopilot' feature, then scroll through a dialog to the end (Yes, okay, exaggerated) - and then click 'ok' would be a huge hazard.
I don't believe the user should be prompted with the same notice every time they use a product.
However; I do not necessarily believe prompted once is the right answer.
What if someone else uses the vehicle? Or you sell it?
I do not pretend to have the answers...
--
Disclaimer; I know Wikipedia is not gospel.
[1] https://en.wikipedia.org/wiki/Tesla_Autopilot [1.1] https://www.tesla.com/blog/dual-motor-model-s-and-autopilot [1.2] https://www.tesla.com/autopilot [2] https://en.wikipedia.org/wiki/Autopilot
As an example, rationally speaking credit bureaus and financial institutions are 100% responsible if they give someone a loan or open an account for them under your name. But they were allowed by the legal system to put the onus on the victims of the fraud to prove they are victims not perpetrators of it, and they called it "identity theft" instead of "bank fraud," and now we have victims who never did anything wrong or could have done anything differently battling CRAs and FIs for years to get out of a situation they should have never been put into in the first place [0][1].
If the laws allow companies to sell Level 5 autonomous vehicle while keeping the passenger (because it's not really a driver anymore) responsible for all legal matters, they would do so, and they would lobby against any law which puts the burden correctly where it belongs. And we would be forced to put our financial, mental, and legal well-being on the line just to get from point A to point B because it will be impossible to avoid autonomous tech then like it is impossible to avoid credit bureaus now.
[0] Like this case: http://www.cbc.ca/news/canada/manitoba/credit-report-error-f...
[1] Obligatory Mitchell and Webb: https://www.youtube.com/watch?v=-c57WKxeELY
Well, law is allowing Tesla to sell a Level 1 vehicle as Level 5. They have to start applying false advertisement laws first before they can think about other laws.
Secondly, I have a Model 3. The idea that the name is misleading people is the dumbest thing i’ve ever heard. Even if people were confused, their confusion would be cleared up very quickly within a few days of use. When you pick up the car they talk to you extensively about what it can and can’t do, and there are many warnings in the car that explain it. People who make this argument probably haven’t tried using AutoPilot.
Put this is context. you can’t just release a perfect system on day 1. In order to get the software to that level earlier versions need to be seeded into the public. I am 100% convinced the software is a huge safety and convenience win as it is.
Auto pilot literally means vehicle operation without human participation. The name autopilot should be banned.
Finally, you could use this as input to a planner like Model Predictive Control where a higher cost is incurred for approaching anomalous objects.
This video (https://youtu.be/wsixsRI-Sz4?t=1h18m28s) shows Elon Musk, two years ago, saying the following:
"I basically consider autonomous driving to be a solved problem".
"A Model S and Model X can drive with greater safety than a person, already. Right now."
"We are less than two years away from complete autonomy".
I'm not trying to downplay the impacts of the driver or Autopilot here, but even if a police officer was in the cruiser, it doesn't look like it would have been as catastrophic as the article implies.
> "Tesla has always been clear that Autopilot doesn't make the car impervious to all accidents, and before a driver can use Autopilot, they must accept a dialogue box which states that 'Autopilot is designed for use on highways that have a center divider and clear lane markings,'" a Tesla spokesperson said in an emailed statement.
Does this dialog box appear every time the car is started or is it a one time thing, just like terms of services?
Basically this is a well-known issue in aviation. As automation gets better and better, humans rely on it more and more and get less and less able to handle even minute failures in it. Additionally humans are absolutely terrible at being able to be dropped into a complex situation and have to make immediate rational decisions about it - thus the failure mode of "in case of error, tell human to handle it" is a bad idea if time between "tell human about it" and "crash into things" is under ten seconds or so. This has been the subject of many NASA studies and NTSB reports and the above article does a good job presenting this info in a form a layman can understand.
There are currently no known easy solutions, sadly.
(French Guyana, Martinique, Réunion etc.)
Tesla really is sticking it to the man!
Nobody is going to set a cruise control to a certain speed and look away from the road!!!
But somehow they (or just Elon?) seem unreasonably averse to doing that. The man is losing hero-status by the minute.
Why can’t we solve the most basic use case? We don’t have the sensors and it’s all done with cameras?
We clearly don't have an epidemic of cars with forward collision warning systems slamming on their brakes for no reason, so those systems must be able to deal with the false positives somehow.
The first milestone for vehicular automation is not perfection, but to simply be better, and in the worst case - comparable, to humans. Given the media's tendency to try to make big news out of every single self driving vehicle crash, it seems like we're already approaching (if not passing) this first milestone, as these articles are relatively rare while these vehicles have driven hundreds of millions of miles in autopilot. I can't find the latest number, but back in October 2016 Tesla already had 222 million miles under autopilot. It's safe to assume it's now some substantial multiple of that.
Some time it would be interesting to see media releases around the time we transitioned from horses/carriages to automobiles, to see how the state of the media has changed. Or perhaps it has always been this way. "I will add, that the man who never looks into a newspaper is better informed than he who reads them; inasmuch as he who knows nothing is nearer to truth than he whose mind is filled with falsehoods & errors. - Thomas Jefferson, 1807"
https://penguindreams.org/blog/self-driving-cars-will-not-so...
I realize this particular case is different with us talking about assistance/safety features, but I still feel we're going the wrong direction here. These features reduce your awareness. They're like a teenage driver you constantly have to monitor, but worse.
They teach people to be comfortable and less alert with systems that have potentially fatal bugs that we can't yet identify.
We can build mass transit now, at a fraction of the cost of developing the tech for automated vehicles. There are so many corner cases, with highways, city driving conditions that simply cannot pan out for autonomous vehicles without a lot of testing; and a lot of people getting injured and dying.
I wish people would shelve this pipe dream for now. Maybe once we have really solid rail transport again in the US, like we use to decades ago, and rich and poor people can have a means to work even in smaller cities; then we can work on frivolous junk that will benefit the few that can afford it. Right now we just run the risk of killing more people who can't.
Knowing you can get on a train at 2am and make it home does play a significant role in reducing drunk driving, and can benefit people of all income levels.
That's right. It's your fault either way, driver, just like it is in airplanes with autopilot.
How many crashes do we have from people using regular cruise control features? Is the incident rate per mile when using Autopilot from Tesla higher or lower than regular cruise control?
Those are the right questions in my opinion.
- accident rate of Tesla drivers that don’t use autopilot
- accident rate of other cars in the same market/category as Tesla’s (are wealthy people better drivers?)
- accident rate of cars with Tesla Autopilot engaged
- accident rate of other cars with similar ‘assistive’ technologies.
So why would the Tesla have been heading toward the roadside? Was the cruiser perhaps parked in the middle of the roadway?
Edit: so they don't use lidar... Is this a machine vision failure, then?
This is because almost everything in the environment is stationary to the car, so it is hard to distinguish between stationary things in your way and stationary things just off to the side.
So if the car in front of you slows down, it's obvious that you'll hit it if you don't slow down.
A car parked partially in your lane might just look like regular noise to the sensors. That is, using the current technology, if they (correctly) classified the parked car as an obstacle then they would (incorrectly) classify so many non-dangerous situations as obstacles so as to make the system infeasible; the car would emergency brake constantly as it drives down the road.
Some would say that until those obstacles can be classified correctly such technology should not be used at all, but I think it's probably doing more good than harm at the moment and hopefully a good solution can be found quickly. Even if they do fix this problem soon, I'm sure we'll hear plenty of stories where something goes wrong anyway, but hopefully they won't seem so obviously avoidable (who hits a parked car!)
Either having more than one sensor or using mirrors or something.
In fact, I'm sure that the camera systems in place are already doing this to some degree. I think it's more that it's not yet good enough at picking out these stationary objects that are in the path of travel.
Issue is, with every complex system, there are few ins and outs that users need to know - like lane merge buffer lines, sharp lane merges by another vehicle in front of view, etc. Within your first few rides(below 10), users could pretty much guess when it works and when it might have issues.
On labeling the feature, it is marked clearly BETA and wiggle prompt is shown pretty much in all the confusing cases - so alert is never missed for me.
My only gripe is with all these incidents, Tesla is forced to update the feature to make it less user friendly or take out completely. I genuinely feel safe and come home every single day with less stress. Elon's tantrums are not helping the situation either.
there’s not a single piece of software i trust in my daily life. i mean, something as simple as an iPad, phone, computer, tv, etc. hardly goes a day without a glitch, but yet people are ready to trust “self-driving” auto features as if they’ve never experienced a software failure before. and normally, it isn’t a software failure in the strictest sense. it’s a human error embedded in software. so i ask myself, do i trust the however many overworked, sleep deprived, wide-eyed engineers’ abstract and concrete thought now embedded in cars’ active driving systems? the answer is no way.
It seems Teslas autopilot cannot reliably do much more than conventional lane assist etc.
Tesla could disable autonomous lane changing and autonomous changing of streets, requiring driver's input for all changes of direction (that are not smoothing out, but staying in lane), and it would be really good.
But Tesla needs the impression of game-changing, not just best-in-class for financial markets reasons.
I believe authorities should order call-backs and force Tesla to disable all traces of autonomy.
"Unfortunately, our website is currently unavailable in most European countries. We are engaged on the issue and committed to looking at options that support our full range of digital offerings to the EU market. We continue to identify technical compliance solutions that will provide all readers with our award-winning journalism."
Maybe we should develop a habbit here on HN not to link to geo-fenced sites?
I can't read the article. Weird. is this because of GDPR?
That Tesla is taking so long to learn what "human factors" are is painful, especially since HF has been part of other industries for decades.
How convenient for non-autopilot auto manufacturers... to have a competitor out there boasting semi-autopilot capabilities...
I wonder how many of these "autopilot crashes" could be machinations... if not already, then in the future perhaps?
Since they're required to, the us dept of transportation puts it at just over 6 million.
Priceless is of course the right answer, but engineering wise you're in the 5-10 million range.
Its AutoPilot, why the hell are all those people expecting it to Pilot the car Auto matically? ....
Three times in a row, has this article been written by a robot? It doesn't feel like something a human would write.
Perhaps Tesla should consider a new CEO and Elon take a move to a more figurehead position?
After multiple deaths now the head of marketing at Tesla should be criminally prosecuted (or whoever is the big shot who signed off on this name).
I have a lot of problems with Tesla, including the way they market this, but the name is accurate.
Some one should study you. Because I don't know how someone with a brain can be so stupid...Not considering the possiblity that you are paid for it, because everyone knows that there are no shills on hackenews...
But think about it like Elon Musk would: If AutoPilot is statistically saver that a car without AutoPilot then it's good enough to be sold on production cars. Tesla cars are saver that normal cars statistically.
Off course, picking one crash and saying AutoPilot sucks is statistically not the right approach. I do agree that Tesla should improve on this. But crashing into a parked car probably means that the driver was playing with his phone, or the cars touch screen too much. Tesla should have systems in place that detects the driver not paying attention. (something better than touching the steer)