DSM-based analysis for the recognition of modeling errors in supervisory controller design

Abstract

The design of supervisory controllers for cyber-physical systems is steadily becoming harder as increasingly more functionality needs to be automated, the systems become larger, and safe operation becomes more important. Model-based systems engineering incorporating formal methods such as supervisory control synthesis can be used to synthesize these supervisory controllers based on models of the uncontrolled system components and models of the control requirements. Although synthesis is an automatic procedure, creating these models is still a manual activity prone to modeling errors. In this paper, we propose to use several DSM-supported analysis techniques to identify potential modeling errors. Analyzing the dependencies between uncontrolled system component models and requirement models with both a domain mapping matrix and a dependency structure matrix reveals potential modeling errors. We present several examples of models from literature to show the potential effectiveness of the DSM-supported analysis of the uncontrolled system and the associated control requirements.