Researchers develop AI algorithm to detect and thwart 99% of MitM cyberattacks on military robots.

Problems with MitM attacks on unmanned military robots:

• Man-in-the-middle (MitM) attacks can intercept data between robots and their controllers.

• Attacks aim to disrupt, modify instructions, or even take control of unmanned vehicles.

• Robot Operating System (ROS) is vulnerable due to its high network connectivity.

• Industry 4.0 advancements make robots collaborative but also more susceptible to cyberattacks.

Challenges:

• Determining MitM attacks is tough as robots work in fault-tolerant modes.

• Attacks can happen at multiple system levels, from core to sub-components.

Researchers’ Approach:

• Analyzed robot's network traffic data for compromises.

• Employed node-based methods, packet scrutiny, and flow-statistic-based system.

• Deep learning with a CNN model was used, consisting of multiple layers and filters for high reliability.

Results:

• Researchers created a machine learning algorithm to detect and stop these attacks rapidly.

  • The algorithm was tested on a GVR-BOT replica (U.S. Army's TARDEC).

  • Success Rate: Detected attacks 99% of the time.

  • False Positives: Below 2%.

Future Applications:

• Potential use in advanced robotic systems like unmanned aircraft.

• Researchers plan to test on faster, more complex robotic platforms, e.g., unmanned aerial vehicles.

Source