Slicing for model reduction in adaptive embedded systems development

Abstract

Model-based development of adaptive embedded systems is an approach to deal with the increased complexity that adaptation requirements impose on system design. Integrating formal verification techniques into this design process provides means to rigorously prove critical properties. However, most automatic verification techniques such as model checking are only effectively applicable to systems of limited sizes due to the state-explosion problem. Our approach to alleviate this problem consists of (a) a semantics-based integration of model-based development and formal verification for adaptive embedded systems and (b) an automatic slicing technique of models with respect to properties to be verified. Slicing is carried out on a high-level formal intermediate representation of the models providing a clear separation of functional and adaptation behaviour. The internal model structure can be exploited to identify system parts that are irrelevant for a property. In particular, slicing offers efficient model reductions for the verification of properties of the adaptation behaviour. The overall approach and the slicing techniques have been evaluated together with the development of an adaptive vehicle stability control system.