A Novel and Unifying View of Trustworthiness in Cyberphysical Systems.

The objective of this work was to study how to assess trustworthiness of computer systems and software used in cyberphysical applications, by developing a multifaceted assessment method and building models based on theories, which rely on non-probabilistic methods and have not been used for such purposes before: (1) Bayesian belief networks, (2) rough sets, (3) Dempster–Shafer theory of evidence, and (4) particle filters. The project objective was accomplished by assuming that trustworthiness is related to the confidence in the results with which related system parameters are evaluated, and then addressing the issue from the perspective of each of the four theories, using a computational model of a cyberphysical system and applying these theories in simple case studies, which can be extended for use in practice.