Theory of evidence-based automated decision making in cyber-physical systems

The Smart Grid is a complex cyber-physical system that is evolving rapidly from a relatively isolated to an open and diverse environment. Within this context, enhancing the security of the future Smart Grid becomes a major priority. In this paper we introduce the use of data fusion for automated decision making in cyber-physical systems such as the Smart Grid. One of the most important applications of decision making is in the field of anomaly detection. This can enable the detection of attacks in cyber-physical systems without requiring a complete description of the physical process. The novelty of our approach is that is combines reports of various cyber and physical sensors, rather than focusing on either one single metric, or one singe realm, as was the case of similar techniques. Based on the proposed architecture we implement a new cyber-physical anomaly detection system. We show that data fusion is much more effective if it combines both cyber and physical realms, rather than focusing on the two realms separately.