Self-Explanation in Systems of Systems

Technical systems have reached a complexity rendering their behaviour difficult to comprehend or seemingly nondeterministic. Thus, self-explaining digital systems would be a strong support for tasks like debugging, diagnosis of failures, reliably operating the system or optimization. To be useful, self-explanation must be efficiently computable in a technical system and must be understandable to the addressee. The addressee might be another technical system at the same or another system layer or a human. We provide a conceptual framework for self-explanation including formalization of the inherent concepts of explanation, understandability etc. We instantiate these generic concepts on the example of Mealy machine models of embedded systems and illustrate their use via example from autonomous driving.