Why don't we build some experimental infrastructure that is designed specifically for self driving vehicles?
Hacked billboards make Teslas see phantom objects causing them to swerve or stop
21–30 of 264 posts
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#22Clever but don’t forget you can also ‘hack’ human drivers by showing flashing patterns or just generally distracting them.
Not this easily
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#23Clever but don’t forget you can also ‘hack’ human drivers by showing flashing patterns or just generally distracting them.
If Tesla made humans, that fact would be a problem for them.
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#24This is clever, and Tesla's software is awful, but anyone doing this for real, with intent to divert cars from the road etc. would be committing a pretty serious crime.
The wonderful thing about IoT is the criminal can be on the other side of the world with no chance of any justice coming to them.
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#25Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#26Earlier quoted context omitted.
Can LIDAR reliably detect stop signs? I think it would have to be based on the shape, and LIDAR loses resolution pretty quickly over distance.
You could always map the stop signs in advance
Which would fail in the case of construction workers holding up stop signs or school buses deploying stop signs, to cite just two examples.
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#27States just need to start mandating LIDAR for any level of autonomous certification, in the same way we finally figured out seatbelts were the minimum level of safety for passenger retention.
Why? Cameras can work just as well as your vision. Humans manage just fine and don't have LIDAR. Just because it causes the cars to slow down now doesn't mean it will in the future.
It's a little strange to make a direct comparison between human cognition and computer vision given that the latter lacks the true ability to understand what it sees. You can train the computer for these edge cases, but humans largely do not have them in the first place because we're capable of formulating highly accurate understandings of what we see.
You can of course pick out discrete examples of humans failing tasks like these, but entire classes of computers routinely fail at things like this.
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#28States just need to start mandating LIDAR for any level of autonomous certification, in the same way we finally figured out seatbelts were the minimum level of safety for passenger retention.
While I agree with you on mandating it, a LIDAR "only" gives you a 3D point cloud.
Great for detecting if an object is in your way - but it cannot read a stop sign, you still need a camera system for that - so they still have to harden it against this hack (and who knows how many to come).
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#29Why don't we build some experimental infrastructure that is designed specifically for self driving vehicles?
Having to invest is specialized infrastructure for self-driving cars defeats the whole point of them.
Any money spent on infrastructure for self-driving cars would be better spent on public transportation.
Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop
#30States just need to start mandating LIDAR for any level of autonomous certification, in the same way we finally figured out seatbelts were the minimum level of safety for passenger retention.
Why? Cameras can work just as well as your vision. Humans manage just fine and don't have LIDAR. Just because it causes the cars to slow down now doesn't mean it will in the future.
Your vision is backed by a neural network with around 80 billion neurons implementing a general learning and intelligence model we don't have the first clue how to replicate. We're a bit further away from "Cameras working just as well as your vision" than I think you're anticipating.