Rule-based model transformation for, and in simulink
Joachim Denil, Pieter J. Mosterman, Hans Vangheluwe · 2014
Over the past decade, the design of embedded sys-tems has come to rely on models as electronic arti-facts that are both analysable and executable. Such ex-ecutable models are at the core of Model-Based De-sign. Simulink ® is a popular Model-Based Design tool that supports simulation of models in various stages of design. While Simulink supports relating the var-ious different models used in design, the technology to do so relies on the underlying Simulink code base. Instead, this paper employs explicit models of the re-lations between the various different design models. In particular, a rule-based approach is presented for model-to-model transformations. The abstraction from the code base provides benefits such as a more intu-itive representation and the ability to more effectively reason about the transformations. The transformation rules and schedules are designed by augmenting stan-dard Simulink model elements (e.g., blocks) for use in model transformation based on the structured RAM-ification approach. The approach is illustrated by the transformation of a continuous-time model, part of an adaptive controller, to a disrete-time counterpart, which is consecutively optimized for simulation. 1.