A hybrid modeling and verification paradigm for embedded control systems

Abstract

Contemporary process control includes continuous and discrete components. At the lowest level, continuous PID controllers are used for actuation and control. At a higher level, supervisory control mechanisms are used to select appropriate control algorithms for the different modes of system operation to achieve optimal or near-optimal control. Modeling and analysis of such combined discrete and continuous components requires hybrid modeling techniques. This paper presents a hybrid modeling paradigm, and discusses its execution semantics, which are based on the principles of invariance of state and temporal evolution of state. The modeling and simulation methodology is used to analyze the control behavior of dynamic physical systems, and a model verification technique based on divergence of time demonstrates possible applications in design tasks.