Specifying and verifying real-time systems with timing uncertainty

Abstract

Modeling timing behaviors of systems and verifying timing constraints against the model are major tasks in developing real-time systems. However, it is recognized to be extremely difficult to predict the timing behavior of systems precisely in the requirements phase. Timing uncertainty, if not considered properly in the modeling and verifying steps, may incur subtle, yet critical errors in final products. We propose a method of dealing with such timing uncertainty using an extended duration calculus formalism. The extended duration calculus, named fuzzy duration calculus, allows us to specify uncertain timing behavior of the system using the fuzzy theory. Semantics and proof system of the duration calculus are redefined in order to accommodate fuzzy concepts. Based on the semantics and proof system, we can verify timing constraints against the system model. A portion of nuclear power plant (NPP) control system is employed to demonstrate the feasibility and the effectiveness of our approach.