by Dominik J. Domis, Christian Schäfer, Mario Trapp
Abstract:
Embedded systems are mainly based on physical models. These models are coarsely approximated and premise various assumptions. Additionally, there is usually not one model, but there are various different physical models the developer has to choose from. Each variant entails different preconditions, advantages, and deficiencies. It is therefore essential to ensure that the approximations of the selected model do not compromise the safety and reliability of the system and that the preconditions are fulfilled under all runtime conditions. Despite the enormous variability, the selection of a variant is still an experienced-based, manual process. As a result, deficiencies are identified only late in the development process, making their compensation very expensive or even impossible. In this paper, we present a physical domain model, which can be used to formally model the physical domain knowledge, including the inherent variability. Based on this model, we will illustrate how different analyses support the developer in exploring the given design space in order to ensure the safety and reliability of the different possible realization concepts.
Reference:
D. J. Domis, C. Schäfer and M. Trapp, "Physical Domain Modeling for the Development of Dependable Embedded Systems", in Proceedings of the 10th international conference on software engineering and applications, A. M. K. Chenk, Ed., Anaheim, CA: ACTA Press, 2006, pp. 150-154.
Bibtex Entry:
@INPROCEEDINGS{Domis2006,
author = {Dominik J. Domis and Christian Schäfer and Mario Trapp},
title = {Physical Domain Modeling for the Development of Dependable Embedded
Systems},
booktitle = {Proceedings of the 10th international conference on software engineering
and applications},
year = {2006},
editor = {Chenk, A. M. K.},
pages = {150-154},
address = {Anaheim, CA},
publisher = {ACTA Press},
abstract = {Embedded systems are mainly based on physical models. These models
are coarsely approximated and premise various assumptions. Additionally,
there is usually not one model, but there are various different physical
models the developer has to choose from. Each variant entails different
preconditions, advantages, and deficiencies. It is therefore essential
to ensure that the approximations of the selected model do not compromise
the safety and reliability of the system and that the preconditions
are fulfilled under all runtime conditions. Despite the enormous
variability, the selection of a variant is still an experienced-based,
manual process. As a result, deficiencies are identified only late
in the development process, making their compensation very expensive
or even impossible. In this paper, we present a physical domain model,
which can be used to formally model the physical domain knowledge,
including the inherent variability. Based on this model, we will
illustrate how different analyses support the developer in exploring
the given design space in order to ensure the safety and reliability
of the different possible realization concepts.},
isbn = {0-88986-642-2},
url = {http://www.actapress.com/Abstract.aspx?paperId=28947}
}