Self-driving cars can be fooled by fake signals
You'd think that self-driving cars would be most vulnerable to remote hacks, but the biggest danger may come from someone nearby with a handful of cheap electronics. Security researcher Jonathan Petit has determined that you can fool LIDAR (the laser ranging common on autonomous vehicles) by sending "echoes" of fake cars and other objects through laser pulses. All you need is a low-power laser, a basic computing device (an Arduino kit or Raspberry Pi is enough) and the right timing -- you don't even need good aim. Petit managed to spoof objects from as far as 330 feet away in his proof-of-concept attack, and he notes that it's possible to present multiple copies of these imaginary objects or make them move. In other words, it'd only take one prankster to make a self-driving car swerve or stop to avoid a non-existent threat.
There's no guarantee that this will be a major issue if and when self-driving cars become commonplace. Petit's technique only works so long as LIDAR units' pulses aren't encrypted or otherwise obscured. While that's true of many commercial systems at the moment, it's possible that production-ready vehicles will lock things down. Still, this is a not-so-friendly reminder that car makers have a lot of work ahead of them if they're going to secure their robotic rides.
http://www.engadget.com/2015/09/05/self-driving-car-lidar-exploit
Just one reason those 'magic' self-driving Google cars are nothing more than PR.
They work, but it'll be generations before they are widely trusted and used.
= new reply since forum marked as read
A whole bunch of jokers seem to think it's hilarious to shine laser pointers into airliner cockpits overhead. 
Not you, of course, but it needs to be kept in mind.
There's a LOT of pushback against High Speed Rail and against light rail and other public transit.
And a lot of opponents say that Uber and Google cars and V2V technology will save the day, and that V2V technology in particular will triple the capacity of highways because we'll all be driving in "car trains".
Well, it's just not so, V2V only increases density where traffic is light and moving, by decreasing inter-vehicular distance.
A lot of cars have collision avoidance, I don't know how long before a significant number are autonomous for braking and steering, but there will be glitches and a growing number of people cannot afford a new vehicle.
I'm all for just killing the auto era altogether where possible, and millenials are looking for a world without cars, too.
that's all, rant off!
And, in case you remain unconvinced consider its stated accuracy, 3-4 meters. I don't think that's good enough, do you? Maybe close enough for targeting a bunker buster, but not for navigating I-80 through the Rockies in winter, let alone summer.
So the GPS argument just doesn't play. Plus, have you ever tried to access a satellite in a snow storm? Sorry, your autonomous vehicle just drove you over a cliff.
There is no way an autonomous vehicle could navigate it in the winter. It would likely end up in the fucking lake. The only way I can navigate it is that I have driven it hundreds of times. At night!!! In Winter!!!
Try that in a Google car and I will retract my claim. Yup! It'll end up in the lake, which is only a couple feet from the road's edge. BTW, an oncoming car pushes one to the limit because the road is barely wide enough for two car widths. Thankfully, meeting another car is rare.
I stand by my claim. Autonomous vehicles wouldn't even be able to handle I-80 in the winter. And through Colorado? 
Pipe dreams.
My argument that autonomous vehicles are only tested in fair weather warm climates has not been addressed.
What good is an autonomous vehicle that cannot handle conditions which are prevalent over a vast area of the country's interstate highway system for a significant proportion of the year?
You keep missing the mark, my friend. A fair weather only system is useless given the expense and technology involved.
Instead, there are many technologies which help a driver be safer. Those are available now in high end cars. They should become available for everybody, like anti-lock brakes, which I can personally attest are wonderful on snow and ice covered unpaved roads, and paved roads too.
But a vehicle driving by itself? Dream on unless it is on rails, or something. Who would buy a vehicle that is only useful during warm weather?
Again! Autonomous vehicles are never tested in winter conditions, which happen to occur in a rather vast part of the country.
One should address that before one trots out yet another straw man.
Best regards.
But now only as augmentation, as is demonstrated by high end cars. They have all sorts of tech built in. Eventually that tech will be available in Fords and Chevy, instead of Mercedes and Volvo, who seem to be leading the way.
But autonomous vehicles are a far, far reach beyond tech augmentation. One does not need infrastructure changes or clear weather for augmentation to be useful. It is an add on, not a replacement. There is quite a difference between the two.
That is my point. We may get there eventually, but I would bet that personal transportation will be a thing of the past before autonomous transportation happens.
Just my conjecture.
I only have DU because we recently got cell reception here. Before that, it was dialup.
So it's laughable that one would claim that Grass Lake Road would get RFID any time soon, if ever. It doesn't even have pavement, let alone two lanes. The only reason it isn't wash boarded is that there are residences there with school kids. So they grade the road during the school year and plow it in the winter for the school bus. Plus, there's the beavers dragging tree branches and leaving them across the road. And the always problem of a wash out. Did I say that the roadway is lower than the level of the lake at some points?
But yup! We're gonna have RFID any day now.
Dream on. I would settle for Internet. At least we have cell phones now.
And you think the county is going to place RFID?
These discussions are always good friendly fun.
