U.S. Proposes Car-To-Car Data Sharing Standards (networkworld.com)

Calling it "the next revolution in roadway safety," the U.S. Department of Transportation hopes to standardize "vehicle communications" technology. Slashdot reader coondoggie writes: The idea is to enable a multitude of new crash-avoidance applications that could save lives by preventing "hundreds of thousands of crashes every year by helping vehicles 'talk' to each other," the DOT stated... [D]evices would use the dedicated short range communications to transmit data, such as location, direction and speed, to nearby vehicles. That data would be updated and broadcast up to 10 times per second to nearby vehicles, and using that information, V2V-equipped vehicles can identify risks and provide warnings to drivers to avoid imminent crashes.
Self-driving cars (and human drivers) could be informed when it's safe to enter the passing lane (or when cars move into a vehicle's blind spot), for example, and "often in situations in which the driver and on-board sensors alone cannot detect the threat." Federal agencies estimate it will cost just $350 per vehicle by 2020 (and dropping over the decades to come), and they've also already issued guidelines about securing these systems from unauthorized access.

Comments

Re:automated tracking

[Imrik]

Since this is car to car instead of a centralized system, I find it less worrying in these respects than the efforts to shift from taxing gasoline to taxing road usage.

Another Day, Another Dupe?

[Striek]

Is this the same story as Feds Unveil Rule Requiring Cars To 'Talk' To Each Other, or am I missing something?

Re:Federal Estimates?

[Dorianny]

"Only $350 per vehicle"

So, at least $1000 per vehicle, and probably a lot more?

And then they'll be able to put up a bunch of sensors along the streets to keep track of where anyone goes at any particular time, and do things like monitor adherence to traffic laws.

Of course, they'll say "we would never do that," but we all know how that sort of thing works out in the long run.

Its truly astounding but on this one they actually worked closely with security researchers to make the technology not useful for tracking. The identification number is randomized every 5 minutes and it contains no information that can tie it to the vehicle's VIN. Hopefully the standard will also specify the minimum uncertainty for the randomize routine so we don't end up in the unfortunate situation we have today with many devices that use very predictable randomize functions to generate insecure keys.

Re:Yes. Sounds VERY secure...

[Dutch Gun]

also already issued guidelines about securing these systems from unauthorized access.

Have we learned nothing from the internet and its IoT problem?

At a fundamental level, it's incredibly difficult to prevent unauthorized access to a physical device someone owns, and I deeply dislike relying on a signals from other cars that can be jammed, interfered with, or abused. If the internet has taught us anything, it's that people will figure out how to crack damn near everything, and good things will be abused just because. Someone may try to get cars to react to a phantom obstacle just for the lulz, to be recorded and uploaded to YouTube.

Finally... are we even certain such a system would be of any benefit? Before we start legislating or regulating these sorts of systems into existence, let's allow self-driving technology to mature on its own a bit first, and see if this would even be useful. Otherwise we'll pay an extra $350 (or more likely $1000, as someone else rightly observed) tax for hardware that has no practical purpose.