The CAN Hacking Tool, or CHT. "Auto makers have long downplayed the threat of hacker attacks on their cars and trucks, arguing that their vehicles' increasingly-networked systems are protected from rogue wireless intrusion. Now two researchers plan to show that a few minutes alone … " -- from the Forbes story, "This iPhone-Sized Device Can Hack a Car, Researchers Plan to Demonstrate" by Andy Greenberg (Forbes, 2/05/2014).
Communication Encryption has been around for quite a while in the military for wired connections. And wireless to for communication, say, between jet fighters. So I agree that this problem can be solved in time for V2V communication in the next few years. In fact, I'm hearing that the defense department will benefit from cheaper encryption if it becomes ubiquitous, just as they benefitted from cheaper electronics when smart phones started proliferating. The media is way over-reacting as usual.
Can it stop the car from running or take over the control??
That's the panic. Let's say the car is networked, e.g. to support V2V communications. And let's say that the internal networks will allow an intruder to spoof ABS sensor signals. Some clever software geek may be able to generate a rapid sequence of "wheel stopped" signals to the ABS, preventing the brakes from operating as the driver expected.
The same sort of mechanism might be used on stability control systems, affecting the way the steering wheel might work.
These scenarios can be avoided, of course, but it does take discipline. Some of these back doors might not always be obvious. The urge to "network everything" has to be resisted, for one thing.
An increasing trend of research on the security vulnerability of autonomous cars would enable the manufacturers to come out with more robust security systems, hence this is good.
Reading through the article in Forbes I see that, this "CAN Hacking Tool" would still need physical access to the car by the hackers.
" Their tool, which is about three-quarters the size of an iPhone, attaches via four wires to the Controller Area Network or CAN bus of a vehicle..."
In this case the gadget made by the researchers would accept commands...which are supposed to be hacking commands when the gadget has to be physically hooked to the car's network. Hence I don't see a reason to panic yet as physical access to car's electronics could be restricted. Am I missing anything?
In my opinion, if we can judiciously decide what can be controlled by the V2V communications and what should be lfet only in the control of the native control system and the driver , then a lot of security concerns can be automatically addressed, regarding somebody taking complete control of the car's vital systems and forcing it crash or malfunction,
Even in some earler blogs also I have comented that , a dual control system is required in such cars - A native ( CAN base) system to control the cars vital systems such as Engine control, ABS, Airbags etc. , whereas the V2V system can monitor the car-to-car distance , speed limits, road obstacles and so on and this system should .
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.