AI algorithm detects man-in-the-middle attacks on military driverless vehicles

An artificial intelligence algorithm developed by researchers can detect man-in-the-middle attacks against military unmanned vehicles.

Robot operating system (ROS) is highly networked. Robots need to collaborate. The sensors, controllers, etc. need to communicate and exchange information through cloud services. Therefore, Extremely vulnerable to cyberattacks such as data breaches and electromagnetic hijacking attacks. A man-in-the-middle attack (MitM) is a network attack that can intercept and tamper with communication data between two parties. A man-in-the-middle attack can disrupt the operation of unmanned vehicles, modify transmitted instructions, and even control and guide robots to perform dangerous actions.

Robot systems can be attacked from different levels, including core systems, subsystems, and subcomponents, causing operational problems that prevent the robot from working properly. Researchers from Australia's University of South Australia and Charles Sturt University have developed an artificial intelligence algorithm that can detect and block man-in-the-middle attacks against military unmanned robots. They use machine learning technology to detect man-in-the-middle attacks and detect them within seconds. This will block the attack.

The picture shows the different nodes that man-in-the-middle attacks can attack

Detecting man-in-the-middle attacks against autonomous vehicles and robots is extremely complex, so these systems are Operating in fault-tolerant mode, it is very difficult to distinguish between normal operation and error conditions. Researchers developed a machine learning system that can analyze a robot's network traffic to detect malicious traffic trying to hack into the robot's system. The system uses a node-based approach to scrutinize packet data and a flow statistics-based system to read metadata from the packet header. , using deep learning convolutional neural networks (CNN) to enhance the accuracy of detection results.

The researchers used the GVR-BOT robot to conduct tests, and the experiment showed that 99% of man-in-the-middle attacks can be successfully blocked, with a false positive rate of less than 2%.

The picture shows the sensor data (the attack starts from 300 seconds)

The picture shows the performance test results

Researchers say the system can be improved and used in other robotic systems, such as unmanned aerial systems. Communication between drones is faster and more complex than with land robots.

The above is the detailed content of AI algorithm detects man-in-the-middle attacks on military driverless vehicles. For more information, please follow other related articles on the PHP Chinese website!