An Approach to Requirement Analysis for Periodic Control Systems

This paper proposes a requirement analysis approach to periodic control systems that are widely used as one of the real time systems. By regulating the initial requirement documents with key words in natural language, we compile the regulated requirement documents into an intermediate model specified by SPARDL language with formal syntax and semantics. To make the requirement executable, a prototype generation technique is proposed to simulate the system behaviors. To analyze the dataflow relations among modules among the same mode or different modes, we introduce module-level and mode-level dataflow analysis techniques to help system engineers to uncover the potential affections on any two modules. The dataflow analysis techniques are useful especially for module reuse when a new version of the system is developed. We have applied the developed tool based on our approach to the Moon-Exploration Spacecraft Project from Beijing Institute of Control Engineering, and the preliminary experiments are encouraging. We have found both the ambiguity and the inconsistency cases in the requirement documents from the project.