A modelling approach to privacy and safety issues in cyber-physical systems

Abstract

Safety issues are relevant in the interaction of human and robotic systems, as these might be a risk for humans’ physical integrity. Robotic systems in the armies are a good example of such risk, since these systems are highly dependent on the software integrity that governs their behaviour. But in civil society there are examples of such issues as well. For example, manufacturing systems are connected to each other and/or to communication networks systems, so these being potentially targets of hijacking. Also, medical applications are usually designed to be very accurate/precise, for instance, in surgery. If that accuracy/precision is threatened due to a cyberattack -as some of those robots are connected to communication networks- the patient could even die. In this work, we describe a modelling approach that categorizes the most relevant cybersecurity threats to robots from different analytical viewpoints (e.g., type of human interrelation with robots, industrial sector, etc.). To illustrate the usefulness of this model, the main characteristics of a model addressing the cybersecurity threats to manufacturing robotic systems is shown in this work. Such model has been implemented into a machine-understandable software by using standard knowledge representation methods.