Data-Driven Modeling of Cyber-Physical Systems using Side-Channel Analysis

Book excerpt: This book provides a new perspective on modeling cyber-physical systems (CPS), using a data-driven approach. The authors cover the use of state-of-the-art machine learning and artificial intelligence algorithms for modeling various aspect of the CPS. This book provides insight on how a data-driven modeling approach can be utilized to take advantage of the relation between the cyber and the physical domain of the CPS to aid the first-principle approach in capturing the stochastic phenomena affecting the CPS. The authors provide practical use cases of the data-driven modeling approach for securing the CPS, presenting novel attack models, building and maintaining the digital twin of the physical system. The book also presents novel, data-driven algorithms to handle non- Euclidean data. In summary, this book presents a novel perspective for modeling the CPS.