Towards one Model Interpreter for Both Design and Deployment
Résumé
Executable modeling of complex embedded systems is essential for bug discovery and safety validation at early design
stages. A relatively large number of tools enable early design diagnosis and validation by transforming and analyzing themodel into a formal setting. However, this transformation induces a semantic gap rendering diagnosis more difficult. Moreover, on the way to deployment, executable models are transformed into low level executable code. Existence of this second transformation similarly renders diagnosis of the deployed system more difficult,
and also increases validation costs of the approach in the context of critical systems: a non trivial equivalence relation needs to be established between the formally analyzed model and the executable code. In this paper, we introduce a first step towards addressing these problems with a bare-metal UML interpreter, which uniquely defines the executable semantics for both design and deployment. To facilitate the diagnosis and validation process
our interpreter offers a diagnosis interface through which the semantics is shared with diagnosis tools. The tools rely on this interface to interact with (observe and control) the executing model either locally on a PC during early design phases or remotely on the target embedded system after deployment. We illustrate our approach on a railway level crossing system ported to two embedded targets (at91sam7s and stm32), to which we
connect a remote high-level simulator for interactive execution control and exhaustive state-space exploration.