As cyber-attacks become more sophisticated not only are computer systems increasingly vulnerable, but so too are more advanced forms of automation. The world of robotics is developing fast, including a new generation of robotic vehicles and drones. The new research demonstrates that alternative safeguards are required in order to minimize the threat of external attacks.
These threats extend to something very sophisticated, like the Mars Rover, to the types of delivery technology that e-commerce sites like Amazon as developing for sending packages to residences.
To highlight the risks, researchers based at the University of British Columbia developed three different types of ‘stealth attack‘. These external inputs caused various machines to crash, miss their targets or to be delayed in completing their pre-programmed tasks.
The main target for the attack is with algorithms used to enable autonomous vehicles to deviate from their intended course. Such deviations are necessary should a robotic vehicle encounter an obstacle in road, for example, or to avoid areas where the wind is strong (in the case of a drone). Algorithms like this help to keep the robot to stay on track while in motion.
As lead researcher, Professor Karthik Pattabiraman notes: “We saw major weaknesses in robotic vehicle software that could allow attackers to easily disrupt the behaviour of many different kinds of these machines.”
He adds: “Especially worrisome is the fact that none of these attacks could be detected by the most commonly used detection techniques.”
This suggests that new security protocols are required to prevent future automated machines from hacking. This is especially so if cyber-criminals fund ways to completely override systems and cause machines to undertake activities that lead to serious economic losses; destroy property; or even take actions leading to human bodily harm
The research has been published in the journal ACSAC, with the research paper titled “Out of control: stealthy attacks against robotic vehicles protected by control-based techniques.”
