Information Technology Reference
In-Depth Information
tion logics axioms within OWL ontologies supports the automation and proto-
typing of analysis methods. Building upon an open-world assumption, OWL
is very flexible and open so that one cannot rely on implicit assumptions. On
a more technical level, formal reasoning on OWL ontologies may take a lot
of computation time on a state-of-the-art personal computer due to the de-
scription logics axioms. To lower this time, the rule language
SWRL
and the
query language
SPARQL
(SPARQL Protocol and RDF Query Language
9
)
have been used in performance critical situations. Dividing major analysis
steps into several individual steps, the computation time of reasoning can
also be optimized. In terms of the DeSCAS ontologies, large ontologies with
many axioms have been split up into smaller ontologies. However, this results
in several consecutive reasoning steps.
5
Conclusion
To sum up, this paper has presented a methodology for interweaving dif-
ferently geared development activities represented by the three development
streams
functional development and architecture
,
safety measures
,and
human
factors
, which are relevant within the automotive domain. The methodology
forms a development proceeding for the design of safety critical automotive
systems heavily relying on a formal base in contrast to most standard pro-
ceedings available. For this purpose, domain knowledge has been formalized
using OWL ontologies illustrating how design decisions in one development
stream may impact other domains and how this information can be used to
reason about consequences of design decisions related to the current product
development. However, formalization entails additional modeling effort when
it comes to formalizing domain knowledge and standards, since a vast num-
ber of concepts have to be included in the OWL ontologies. On the other
hand, once formalized domain knowledge and standards can be reused in
other projects. The prototype toolchain of DeSCAS which has been used
for the example lane departure warning system clarifies the advantages and
disadvantages of applying OWL ontologies to logical reasoners for formal rea-
soning, and how to overcome problems arising from long computation times
during reasoning.
Nevertheless, future research of DeSCAS will focus on extending and re-
fining the proposed methodology by detailing the modeled domain knowledge
(e.g. analyzing accident statistics to determine the severity of system hazards)
and integrating more analysis methods (such as ASIL decomposition).
References
1. D. Beisel, C. Reuß, E. Schnieder, and U. Becker. Automotive Generic Haz-
ard List. In
Automatisierungs-, Assistenzsysteme und eingebettete Systeme für
Transportmittel (AAET)
, 2010.
9
SPARQL - http://www.w3.org/TR/rdf-sparql-query/
