New AI algorithm promises defense against cyberattacks on robots

The study, published inIEEE Transactions on Dependable and Secure Computing, claims that their algorithm outperforms other methods of detecting cyberattacks used globally.

The research was conducted by Professor Anthony Finn from the University of South Australia (UniSA) and Dr Fendy Santoso from Charles Sturt University in collaboration with the US Army Futures Command. They simulated a MitM attack on a GVT-BOT ground vehicle and trained its operating system to respond to it, according to the press release.

According to Professor Finn, an autonomous systems researcher at UniSA, the robot operating system (ROS) is prone to cyberattacks because it is highly networked. He explained that Industry 4, characterized by advancements in robotics, automation, and the Internet of Things, requires robots to work together, where sensors, actuators, and controllers communicate and share information via cloud services. He added that this makes them very vulnerable to cyberattacks. He also said that computing power is increasing exponentially every few years, enabling them to develop and implement sophisticated AI algorithms to protect systems from digital threats.

Dr Santoso, an artificial intelligence and cyber futures expert at Charles Sturt University, said the robot operating system needs adequate security measures in its coding scheme due to encrypted network traffic data and limited integrity-checking capability. He stated that their intrusion detection framework, which leverages the benefits of deep learning, is robust and highly accurate. He also said the system can handle large datasets suitable for securing large-scale and real-time data-driven systems such as ROS.

The researchers plan to test their algorithm on other robotic platforms, such as drones, which have faster and more complex dynamics than a ground robot.

The study was published inIEEE Transactions on Dependable and Secure Computing.

Study abstract:

Safe and secure operations of robotic systems are of paramount importance. Aiming for achieving the trusted operation of a military robotic vehicle under contested environments, we introduce a new cyber-physical system based on the concepts of deep learning convolutional neural networks (CNNs). The proposed algorithm is specifically designed to reduce the cyber vulnerability of the Robot Operating System (ROS), a well-known middleware platform widely used in both civilian and military robots. To demonstrate the efficacy of the proposed algorithm, we conduct penetration testing (real-time man-in-the-middle cyber attack) on the GVR-BOT ground vehicle, a military ground robot, developed by the United States Army Combat Capabilities Development Command (CCDC), Ground Vehicle Systems Center. The cyber attack also exploits the vulnerability of the Robot Operating System (ROS) employed in its onboard computer. We collect experimental data and train our CNN based on two different operating conditions, namely, legitimate and malicious conditions. We normalize and convert the network traffic data in the form of RGB or grayscale images. We introduce two different types of windowing techniques, namely, the independent and overlapping sliding epochs to efficiently feed the network traffic data to our CNN system. Our research indicates the efficacy of the proposed algorithm as our proposed cyber intrusion detection system can achieve reasonably high accuracy of ≥99 % and substantially small false-positive rates ≤ 2 % supported with minimum detection time. In addition, we also compare and demonstrate the relative merits of our proposed algorithm with respect to the performance of some well-known techniques, namely, ‘bag-of-features’ and Support Vector Machine (SVM) algorithms.