[ad_1]
In July 2015, an experiment with a couple of Wired journalists showed how easy it was to hack a Jeep Cherokee and drive it away. The public was stunned – Oh dear! – unexpected discovery and everyone began to whisper about the alleged lack of security of autonomous vehicles. This fear is now so widespread and intense that some have already defined the threat of piracy as the reason why autonomous cars will never become a reality. Even a few accidents can prevent this technology from reaching its full development. But is this fear really justified? Are non-autonomous cars really safer, or is it the other way around?
Why are people so afraid of piracy?
All technologies seem 100% safe when new. But as we learned from e-mails and operating systems in the 90s and early 2000s, nothing is safe once it is made public. This is especially true with autonomous cars because part of the AI that controls them is still partially unidentified. The mathematical model that powers the artificial intelligence of Nvidia's drive systems is not based on instructions provided by programmers or engineers. It's a totally autonomous intelligence, based on in-depth learning, which 'slowly learns' to drive while watching humans do it. In their latest report, published in October 2018, the computer graphic card manufacturer explained how their Drive IX system could track the driver's head and eye movements, improving the integration between humans and machines. Nevertheless, the less we know about the system, the more difficult it is to protect against unwanted intrusions.
The consequences of self-driving car hacking
When hacking takes place in a data center, the worst that can happen is a loss of data. When an autonomous car is hacked, what can happen is a loss of life. However, automakers are accustomed to engineering problems as and when they are discovered, an approach that is not acceptable while so many things are at stake. D & # 39; On the other hand, autonomous vehicles are designed to eliminate most of the millions of deaths on the roads in the world each year, a very real and real threat. Will the dangers of being hacked by an insane cyber criminal outweigh the dangers badociated with human conduct? Some data to be processed will provide the answer.
Technology is moving fast! Stay on the cutting edge with Techopedia!
Join nearly 200,000 subscribers who receive technical information from Techopedia.
The first consideration we must take is that people will not accept autonomous cars if their level of safety is the same as that of human driving. According to a study published by the Society for Risk Analysis, the current risk of death around the world due to human error is already 350 times higher than the frequency accepted by the public. In other words, for autonomous cars to be tolerated, they must at least improve safety on the roads. by two orders of magnitude. However, this may be due to a certain level of perception bias against machine safety. In fact, it is interesting to note what General Motors Co. told the California regulators about their accident reports in September 2018. In the six collisions involving autonomous vehicles, the perpetrators of the accidents were still human drivers.
Another key argument against the safety of autonomous cars is that most car accident statistics focus on real-world collisions. In other words, we collect data and discuss it only when the tragedy has already occurred. But what about billions or billions of accidents that have been avoid? We can not measure the number of non-collisions, so how do we determine the capacity of an AI versus a human at do not crush when things go wrong, for example in bad weather or when you have to drive on a steep slope or dirt road, or when a pedestrian unexpectedly enters the road? Right now, we can not – at least, not reliably. And the situation can get worse if hacking attempts (even the most unsuccessful) can alter the delicate controls of autonomous vehicles. (To learn more about autonomous cars, see The 5 Most Amazing Advances in AI in Autonomous Driving.)
Are autonomous cars more vulnerable to piracy?
Who said that autonomous vehicles are more vulnerable to piracy than traditional cars? The idea that a hacker takes the wheel of the car we drive is certainly terrifying, but it is already possible with non-autonomous cars because of the many vulnerabilities of their Internet software. In 2015, a security breach in FCA's Uconnect allowed pirates to take control of a "traditional" Fiat Chrysler, forcing the automaker to recall more than a million vehicles. Even the experiment described above with the Jeep Cherokee involved a Ordinary, car connected to the internet rather than an autonomous car.
In theory, the inherent interconnectivity between multiple sensors and layers of autonomous vehicle communications could expose them more to cyberattacks because they offer more "entry points". However, the hacking of a connected autonomous car is also much more difficult … for the same reason. Having to access a multi-layered system integrating information from multiple sensors as well as traffic and pedestrian data in real time can be a serious obstacle for hackers. IoT-related solutions can also be applied to enhance their security at an exponential level, such as the integration of secure encryption systems based on quantum mechanics.
Once again, hackers can use these same IoT connections to their advantage to violate autonomous vehicle cyber defenses before they are put in place. Hackers can exploit the vulnerabilities of the production chain and supply chain to infiltrate into an autonomous car even before it is ready. This step is extremely delicate, and the former BlackBerry smartphone maker has announced its commitment to preventing such loopholes with its upcoming software for autonomous vehicle safety, Jarvis.
What are the plans to solve the problem?
What potential countermeasures are the best? Solutions include potential cybersecurity risk mitigation plans in the design and manufacturing processes, as cyber-resiliency must be implemented effectively during the vehicle design phase. Experts have already warned against the propensity of current manufacturers to adapt non-autonomous vehicles to some additional sensors. This may be acceptable now, as engineers are still struggling with prototypes and testing the various features of these vehicles, but this approach is later bound to be largely insufficient to ensure any degree of security.
Other cybersecurity measures can be used beyond the vehicle itself and can apply to any additional technologies that constitute the "environment" in which autonomous cars operate (smart poles, sensors, roads and other infrastructure). For example, a stolen pirated vehicle can be stopped as soon as the GPS detects that it is in a place where it should not be. Finally, as autonomous vehicles begin to replace large scale non-autonomous vehicles, the entire infrastructure of all smart cities will change and security will become an integral part of the network.
Since no hostile hacker has yet targeted autonomous vehicles, no real cybersecurity test has been run to protect the autopilot software in a realistic environment. The computerized learning of the machine requires real "enemies" to train; otherwise, manufacturers simply expose their flanks to threats to which no one is ready. As Craig Smith, director of research at Rapid7, cyberbadytic group Rapid7, said in an interview: "Google has been a target of cyberattacks for years, unlike the auto industry, which has a way to go. to browse." In this regard, automakers appear particularly weaker than other companies because they are not used to preventing problems (especially those that are completely out of their field).
Curiously, however, the solution may come from other industries where engineers already have a significant degree of knowledge of protecting vehicles from malicious attacks. GuardKnox, for example, is a company that can protect fleets of cars, buses, and other vehicles by implementing security technology that has been used to protect Israel. jet fighters. Yes, the F-35I and F-16I fighters, to be precise. Seriously. Jet. Fracking. Affairs. Deal with that, pirates!
The unique and exciting protection solution offered by GuardKnox has long been used by other high-level security systems, such as the Iron Dome and Arrow III missile defense systems. The system applies a formally verified and deterministic communication configuration between the various networks of the vehicle, which blocks any unverified communication. Any external communication attempting to access the vehicle's central gateway shield must be verified, effectively locking the entire system, regardless of the number of vulnerable access points present. Centralization is essential to prevent hackers from accessing the main system of the autonomous car or its systems, such as brakes or wheels, from its communication network. (For more information on calculators, see Your car, Your computer: calculators and the controller network.)
What the future holds
Every new generation of automotive technology has its own hazards and risks to safety. Autonomous cars are no exception, and for the moment we can safely badume that the cybersecurity risks badociated with them are somewhat underestimated. However, they are not at all underestimated. In fact, the current attention paid to these perceived risks only serves to encourage the more in-depth research needed to manufacture the new generation of autonomous vehicles in the safest possible way. As GuardKnox CEO and co-founder Moshe Shlisel clearly pointed out, "manufacturers are now adopting a multi-layered approach to vehicle safety, implementing cutting-edge hardware and software changes to better withstand attacks. malicious. attacks. "
Source link
