"Petit began by recording pulses from a commercial Ibeo Lux Lidar unit. Discovering the pulses weren’t encoded or encrypted, he could simply use them at a later time to fool the unit into believing objects were there when they weren’t."
That appears to answer my longstanding question as to what might happen when multiple vehicles using lidar are on the road at the same time. (Pulse timing notwithstanding of course, but given enough cars and enough time it could be periodically problematic.)
However, regarding stopping a car with fake laser generated obstacles - that's nothing that can't be done with real "obstacles". A roll of paper or aluminium foil stretched across the road can't be discriminated from a solid object, and a a few empty cardboard boxes are as good as concrete blocks to a self driving car.
Or even just standing in the road yourself if you have sufficient faith in the technology.
I suspect that the lidar signals could be modulated in the same way that cell phone signals are, so that a population of cars can share a single laser wavelength without confusing one another.
Laser gun sights are very common in the US. So for some people, it's more than just "feeling threatened". They're arguably too paranoid. And irresponsible. But being right about that won't help if they shoot you.
The design is very much intentional. A huge barrier to adoption for self driving cars is perception of safety; Google wanted a car that looks safe not dangerous.
Perception of safety... to other drivers on the road. This is a research prototype, there is no target consumer yet. The main goal is having a car that people will feel comfortable seeing driving itself on the street, not buying, yet. So, "toy car" is what they should be going for, at this point ;)
I guess I'll go with the default assumption that people at Google are smart. What worries me is that a company as smart and as experienced as DEKA could produce something as geeky as the Segway. Don't want to see Google afflicted by the same nerd-blindness.
Ok, that makes more sense. If I were at Google, I'd still be very upset about those images being disseminated, as that design will likely provoke ridicule among the mass market.
Kind of the same way that a Segway is neat tech, but riding one is a defacto declaration that you don't care how goofy you look. Not good marketing.
OTOH, some of the smartest and most successful people (at creating products) that I know express themselves like this. So ... tolerance works both ways. Don't flip the bozo bit based on a word -- there are many cultures on the internet.
EDIT> Your suggestion is prudent, but I don't want prudence and inoffensiveness to suck the life out of my communications, leaving only a dry passive-voiced husk.
Nobody likes a self-righteous censor, I get that, but I'm just being honest with you: it comes off as lazy and juvenile without really helping your case.
And in general, I'm pretty sure I could drive in a sufficiently confusing way during my commute tomorrow to confuse human drivers enough to cause an accident that would have been theoretically avoidable if they had responded better.
"Attacks" on self-driving cars that could already have been conducted on humans, and clearly aren't, because it just isn't worth anybody's time, aren't very technically interesting. One could make a case that there's a socially interesting aspect in that someone could use such an attack, in conjunction with social fears of self-driving cars, to do something like trash somebody's stock price or something, but the "social fear of self-driving cars" would be an irreducible element of such an attack.
It's the attacks like the Jeep attack, where I can hit fleets at a time, because the cars are all intimately networked together or something, that are interesting.
You give a billion people these cars and I guarantee there will be attacks against the cars. Just like there are uncommon incidents of people trying to murder people by sabotaging their vehicles.
I think an important distinction is that a human driver under attack will likely know he/she is under attack very quickly, whereas the passengers in a self-driving car might not know that the self-driving system is under attack without significant delay.
Not to mention that the attack is undetectable. As there will be no witnesses and not recollection of who did the attack. So no legal deterrent to the would be attacker. This is a major problem.
My guess is that either the driver's vision will be quickly impaired or his/her eyes will quickly start hurting. In that case, the driver will quickly know something is wrong even if he/she may not know it's an attack.
Human driven cars can also be disabled with funny sounds or women wearing short skirts or a bit of spray paint on the road.
We really need to remember that humans are atrocious at driving on average, and doing better than them is a pretty low bar. Not that we shouldn't strive to improve automation, but don't let perfect be the enemy of the good.
We really need to remember that humans are atrocious at driving on average, and doing better than them is a pretty low bar.
Let's revisit that assertion when autonomous vehicles manage to even minimally cope with heavy rain or snow, neither of which are barriers for those "atrocious" human drivers.
The reality is that, for all their flaws, humans vastly outperform autonomous vehicles in imperfect driving conditions. That's just reality, despite the HN tendency to pitch this technology as though it could replace all human drivers tomorrow.
Maybe your bad example made it difficult to understand what point you were trying to make. :P
My point is simply that there are tasks where human drivers outperform computers, and there are tasks where computers outperform human drivers, so simply looking at some scenario where humans do better doesn't tell us anything. You have to look at the whole picture.
And yes, currently humans are way better than computers in the general case. I'm not disputing that. I'm just saying that "computers can't handle situation X, which humans can" is a pointless thing to say by itself.
And yes, currently humans are way better than computers in the general case. I'm not disputing that.
ROFL, no, with absolutely no qualification, you flat out said:
We really need to remember that humans are atrocious at driving on average, and doing better than them is a pretty low bar.
I guess what you're saying is that computers are so bad, they still can't get over that "pretty low bar"?
If so, I suppose I agree, though I think you're dramatically underestimating just how capable human beings are.
so simply looking at some scenario where humans do better doesn't tell us anything.
Which is really strange, because that's exactly what everyone does when talking about autonomous vehicles. The discussion always surrounds current trials, where vehicles are allowed to operate in cherrypicked, near-ideal conditions: perfect, southern California climate, with minimal rain, zero snow, and perfect visibility.
So I agree, let's stop looking at just some isolated scenario. Let's focus on the general case... where computers, unfortunately, currently fail, and will likely continue to do so for a very long time (dealing with highly variable conditions likely requires major infrastructure changes to actively support autonomous vehicles, and that ain't happening any time soon).
> I guess what you're saying is that computers are so bad, they still can't get over that "pretty low bar"?
Yes, that is exactly what I'm saying. So you can skip over the parts where you talk about how other people, not me, praise the abilities of current autonomous vehicles, because I'm not doing that.
Human drivers, on average, are not "disabled" by any of that. Distracted momentarily, sometimes, perhaps, but not disabled - city driving is replete with "bits of spray paint" and "funny sounds" and sometimes "women wearing short skirts." On average, human drivers manage a great deal of complexity without incident. We use the same visual acuity, adaptability and instincts to negotiate traffic as we did millennia ago to avoid predators and survive our climb up the food chain, and we're actually quite good at it, albeit obviously not perfect. Human drivers are capable of not mistaking the reflections of other cars for other cars, and know the difference between a cardboard box, a human being and another vehicle.
The bar is actually quite high, if you want to do better than humans on average.
Humans usually aren't disabled by that, but sometimes they are. How often is a human distracted by one of these things, and before they can recover they drive into an obstacle? It happens with great frequency.
And in this particular case, humans not only can be disabled with a laser pointer, but they can be disabled permanently, not just for that driving session, but for the rest of their lives. Can't swap out cameras in a human driver. Yet somehow this is not a crisis for manually-piloted cars.
I doubt it happens that often, but I can't back that up with numbers, and I can't be arsed to try.
But, humans can vary in their driving ability. In the case of autonomous cars, you would have to assume any flaw in one is shared by all. The problem isn't the laser pointers, or that humans can also be disabled by laser pointers, but that autonomous cars aren't yet smart enough to deal with errant signals. If all the autonomous cars can be stopped by an ever increasing list of simple conditions that (with a couple of exceptions, like a laser pointer in the eyes) wouldn't stop an average human driver, then it implies that the current state of the art does not nearly live up to the hype.
We can come up with a long list of simple conditions that wouldn't stop an average computer but would stop an average human, too. For example, a computer won't have any trouble driving for 48 hours without powering down, or with a stinging insect in the cabin.
That's not to say that computers are superior to humans. It's quite clear that the current state of the art in autonomous driving is way behind what humans can do. My point is simply that comparing scenarios which one can handle and the other cannot is not very informative.
"Google, Uber and even Apple’s potential self-driving car can all be foiled..."
So two cars that haven't even been seen in public yet - one of which might not even exist - can be foiled by this technique?
Might be true, might not be - no way for anyone involved in this story to know though. Is this really what passes for journalism now?
In fact, according to the article, this was never even tested on an actual autonomous car's software - he just spoofed a commercial LIDAR system and extrapolated from there.
"Ultimately, this latest car hacking is yet another signal that industries usually disassociated from data security now need to start taking it seriously."
Ever since I first played Need For Speed 2, I've fantasised about a real-life version of the ROADRAGE cheat code I could use to cause those doing the rest of us on the road a disservice to flip over with a mere honk of my horn. We're not there yet, but this is a clearly a step in the right direction.
I remember how when laser pointers started proliferating there was a media panic over their potential use in downing aircraft and blinding people. Schools swiftly moved to ban them(I recall being in a few assemblies where someone brought in a pointer and caused some mild, easily ignored disruption).
Regardless, the world hasn't blown up. You can still legally own a laser pointer in most countries, subject to various power limits. The risks and responsibilities involved in their use are real, but have to be weighed against every other hazard.
The police will crack down on you as hard or harder than if you just pointed a laser pointer at human drivers. You can get very long prison sentences already for doing this.
58 comments
[ 2.7 ms ] story [ 104 ms ] threadThat appears to answer my longstanding question as to what might happen when multiple vehicles using lidar are on the road at the same time. (Pulse timing notwithstanding of course, but given enough cars and enough time it could be periodically problematic.)
However, regarding stopping a car with fake laser generated obstacles - that's nothing that can't be done with real "obstacles". A roll of paper or aluminium foil stretched across the road can't be discriminated from a solid object, and a a few empty cardboard boxes are as good as concrete blocks to a self driving car.
Or even just standing in the road yourself if you have sufficient faith in the technology.
I love the tech, and I love the goal, but I might refuse to ride in one of these just because they look too retarded.
EDIT> This may seem frivolous, but I see it as a huge barrier to adoption. At least make them look cool, not like preschool toys.
Kind of the same way that a Segway is neat tech, but riding one is a defacto declaration that you don't care how goofy you look. Not good marketing.
No one did the industrial design. It isn't a product, it is a prototype to test a technology.
Not to derail, but I'd avoid using language like this if you want to be taken seriously.
OTOH, some of the smartest and most successful people (at creating products) that I know express themselves like this. So ... tolerance works both ways. Don't flip the bozo bit based on a word -- there are many cultures on the internet.
EDIT> Your suggestion is prudent, but I don't want prudence and inoffensiveness to suck the life out of my communications, leaving only a dry passive-voiced husk.
I'll try to be more tolerant...
http://www.livescience.com/21707-lasers-eye-damage.html
"Attacks" on self-driving cars that could already have been conducted on humans, and clearly aren't, because it just isn't worth anybody's time, aren't very technically interesting. One could make a case that there's a socially interesting aspect in that someone could use such an attack, in conjunction with social fears of self-driving cars, to do something like trash somebody's stock price or something, but the "social fear of self-driving cars" would be an irreducible element of such an attack.
It's the attacks like the Jeep attack, where I can hit fleets at a time, because the cars are all intimately networked together or something, that are interesting.
So will the self driven cars, if you read the article.
My guess is that either the driver's vision will be quickly impaired or his/her eyes will quickly start hurting. In that case, the driver will quickly know something is wrong even if he/she may not know it's an attack.
I'm no expert on laser safety or human biology, so I left a question on biology SE: http://biology.stackexchange.com/questions/38603/how-quickly...
Hopefully somebody can give an answer.
And if the driver experiences no symptom, then the attack is just ineffective.
We really need to remember that humans are atrocious at driving on average, and doing better than them is a pretty low bar. Not that we shouldn't strive to improve automation, but don't let perfect be the enemy of the good.
Let's revisit that assertion when autonomous vehicles manage to even minimally cope with heavy rain or snow, neither of which are barriers for those "atrocious" human drivers.
The reality is that, for all their flaws, humans vastly outperform autonomous vehicles in imperfect driving conditions. That's just reality, despite the HN tendency to pitch this technology as though it could replace all human drivers tomorrow.
My point is simply that there are tasks where human drivers outperform computers, and there are tasks where computers outperform human drivers, so simply looking at some scenario where humans do better doesn't tell us anything. You have to look at the whole picture.
And yes, currently humans are way better than computers in the general case. I'm not disputing that. I'm just saying that "computers can't handle situation X, which humans can" is a pointless thing to say by itself.
ROFL, no, with absolutely no qualification, you flat out said:
We really need to remember that humans are atrocious at driving on average, and doing better than them is a pretty low bar.
I guess what you're saying is that computers are so bad, they still can't get over that "pretty low bar"?
If so, I suppose I agree, though I think you're dramatically underestimating just how capable human beings are.
so simply looking at some scenario where humans do better doesn't tell us anything.
Which is really strange, because that's exactly what everyone does when talking about autonomous vehicles. The discussion always surrounds current trials, where vehicles are allowed to operate in cherrypicked, near-ideal conditions: perfect, southern California climate, with minimal rain, zero snow, and perfect visibility.
So I agree, let's stop looking at just some isolated scenario. Let's focus on the general case... where computers, unfortunately, currently fail, and will likely continue to do so for a very long time (dealing with highly variable conditions likely requires major infrastructure changes to actively support autonomous vehicles, and that ain't happening any time soon).
Yes, that is exactly what I'm saying. So you can skip over the parts where you talk about how other people, not me, praise the abilities of current autonomous vehicles, because I'm not doing that.
The bar is actually quite high, if you want to do better than humans on average.
And in this particular case, humans not only can be disabled with a laser pointer, but they can be disabled permanently, not just for that driving session, but for the rest of their lives. Can't swap out cameras in a human driver. Yet somehow this is not a crisis for manually-piloted cars.
I doubt it happens that often, but I can't back that up with numbers, and I can't be arsed to try.
But, humans can vary in their driving ability. In the case of autonomous cars, you would have to assume any flaw in one is shared by all. The problem isn't the laser pointers, or that humans can also be disabled by laser pointers, but that autonomous cars aren't yet smart enough to deal with errant signals. If all the autonomous cars can be stopped by an ever increasing list of simple conditions that (with a couple of exceptions, like a laser pointer in the eyes) wouldn't stop an average human driver, then it implies that the current state of the art does not nearly live up to the hype.
That's not to say that computers are superior to humans. It's quite clear that the current state of the art in autonomous driving is way behind what humans can do. My point is simply that comparing scenarios which one can handle and the other cannot is not very informative.
So two cars that haven't even been seen in public yet - one of which might not even exist - can be foiled by this technique?
Might be true, might not be - no way for anyone involved in this story to know though. Is this really what passes for journalism now?
In fact, according to the article, this was never even tested on an actual autonomous car's software - he just spoofed a commercial LIDAR system and extrapolated from there.
Garbage.
Regardless, the world hasn't blown up. You can still legally own a laser pointer in most countries, subject to various power limits. The risks and responsibilities involved in their use are real, but have to be weighed against every other hazard.