Message Encryption in Robot Operating System: Collateral Effects of Hardening Mobile Robots

In human-robot interaction situations, robot sensors collect huge amounts of data from the environment in order to characterize the situation. Some of the gathered data ought to be treated as private, such as medical data (i.e., medication guidelines), personal and safety information (i.e., images of children, home habits, alarm codes, etc.). However, most robotic software development frameworks are not designed for securely managing this information. This paper analyzes the scenario of hardening one of the most widely used robotic middlewares, Robot Operating System (ROS). The study investigates a robot’s performance when ciphering the messages interchanged between ROS nodes under the publish/subscribe paradigm. In particular, this research focuses on the nodes which manage cameras and LIDAR sensors, which are two of the most extended sensing solutions in mobile robotics, and analyzes the collateral effects on the robot's achievement under different computing capabilities and encryption algorithms (3DES, AES and Blowfish) to robot performance. The findings present empirical evidence that simple encryption algorithms are lightweight enough to provide cyber-security even in low-powered robots when carefully designed and implemented. Nevertheless, these techniques come with a number of serious drawbacks regarding robot autonomy and performance if they are applied randomly. To avoid these issues, we define a taxonomy that links the type of ROS message, computational units, and the encryption methods. As a result, we present a model to select the optimal options for hardening a mobile robot using ROS.