Abstract
Robotic Operating System (ROS) is a framework that aims to help developers when building their robotic applications. ROS gives developers access to many drivers, algorithms and even handle communication among the robot’s components. In recent years, ROS has become a crucial tool in the arsenal of robotic developers, and it can be argued that it is the obvious framework of choice. However, ROS does not provide any security features, and it has been demonstrated to be vulnerable to several security threats. To enhance the security of ROS, existing works use software-defined networking (SDN) architecture with pre-defined policies to provide vulnerability analysis. Nevertheless, their capabilities are restricted when handling new attacks or attacks not covered by the pre-defined rules. In this thesis, we propose to enhance the security of ROS with an intelligent, effective threat monitoring and detection system by leveraging SDN and machine learning techniques. The proposed system can be integrated into existing ROS applications as a plug-in. The system architectures would not be affected, and it will provide real-time protection to mitigate malicious behaviors. Our system can also be customized to support the detection of different types of threats.