This paper presents an approach to the efficient verification of embedded systems. Such systems usually operate in uncertain environments, giving rise to a high degree of nondeterminism in the corresponding formal models, which in turn aggravates the state explosion problem. Careful handling of nondeterminism is therefore crucial for obtaining efficient model checking tools. Here, we support this goal by developing a formal computation model and an abstraction method, called delayed nondeterminism, which instantiates nondeterministic values only if and when this is required by the application code. It is shown how this technique can be integrated into our CTL model checking tool [mc]square by introducing symbolic abstract states which represent several concrete states. We also give a simulation relation between the concrete and the abstract state space, thus establishing the soundness of delayed nondeterminism with respect to "path-universal" logics such as ACTL and LTL. Furthermore, a case study is presented in which three different programs are used to demonstrate the effectiveness of our technique. © 2008 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Noll, T., & Schlich, B. (2008). Delayed nondeterminism in model checking embedded systems assembly code. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4899 LNCS, pp. 185–201). https://doi.org/10.1007/978-3-540-77966-7_16
Mendeley helps you to discover research relevant for your work.