Australian researchers have developed an algorithm capable of intercepting and shutting down a man-in-the-middle (MitM) cyberattack on an unmanned military robot within seconds.
Using deep learning neural networks to mimic human brain behavior, experts in artificial intelligence from Charles Sturt University and the University of South Australia (UniSA) trained the robot’s operating system to recognize the signature of a MitM eavesdropping cyberattack, where attackers disrupt an ongoing conversation or data transfer.
When tested on a replica of a United States army combat ground vehicle, the algorithm successfully prevented 99% of malicious attacks, with false positive rates below 2%, proving its effectiveness. These results have been published in IEEE Transactions on Dependable and Secure Computing.
Professor Anthony Finn, an autonomous systems researcher at UniSA, states that the proposed algorithm outperforms other recognition techniques used globally to detect cyberattacks. Professor Finn and Dr. Fendy Santoso from Charles Sturt Artificial Intelligence and Cyber Futures Institute collaborated with the US Army Futures Command to replicate a man-in-the-middle cyberattack on a GVT-BOT ground vehicle and train its operating system to identify such attacks.
“The highly networked nature of the robot operating system (ROS) makes it extremely vulnerable to data breaches and electronic hijacking,” says Prof Finn.
“Industry 4, characterized by robotics, automation, and the Internet of Things, requires robots to collaborate, exchanging information with each other via cloud services. However, this also exposes them to cyberattacks,” adds Prof Finn.
Dr. Santoso highlights that despite the widespread usage and benefits of the robot operating system, it largely neglects security issues in its coding scheme due to encrypted network traffic data and limited integrity-checking capability.
“Thanks to deep learning, our intrusion detection framework is highly accurate and robust,” says Dr. Santoso. “It can handle large datasets, making it suitable for safeguarding large-scale and real-time data-driven systems like ROS.”
Prof Finn and Dr. Santoso plan to test their intrusion detection algorithm on different robotic platforms, including drones, which have faster and more complex dynamics compared to ground robots.
