Trustable Formal Specification for Software Certification

Abstract

Formal methods have emerged as a complementary approach to ensuring quality and correctness of high-confidence medical systems, overcoming limitations of traditional validation techniques such as simulation and testing. In this paper, we propose a new methodology to obtain certification assurance for complex medical systems design, based on the use of formal methods. The methodology consists of five main phases: first, informal requirements, resulting in a structured version of the requirements, where each fragment is classified according to a fixed taxonomy. In the second phase, informal requirements are represented in formal modeling language, with a precise semantics, and enriched with invariants and temporal constraints. The third phase consists of refinement-based formal verification to test the internal consistency and correctness of the specifications. The fourth phase is the process of determining the degree to which a formal model is an accurate representation of the real world from the perspective of the intended uses of the model using model-checker. Last phase provides an animation framework for the formal model with real-time data set instead of toy-data, and offers a simple way for specifiers to build a domain specific visualization that can be used by domain experts to check whether a formal specification corresponds to their expectations. Furthermore, we show the effectiveness of this methodology for modeling of a cardiac pacemaker system.