Ontology-based temporal querying for assessing safety-critical scenarios in road traffic

Abstract

The dissertation is the first to push temporal querying over expressive Description Logics towards practical applicability. Performance is of specific importance, albeit challenging given that the underlying problem is ExpTime-complete. We thus explore multiple approaches and optimizations for enabling well-performing ontology-based temporal data access. While a first procedure compiles the temporal aspects into a finite automaton, we observe that - even in an optimized version - inefficient disjunctive queries are sometimes unnecessarily checked. A second, orthogonal algorithm therefore proposes to skip the automaton and instead operate directly on the query by means of a normal form. As the goal is to establish practical applicability, we finally broaden the evaluation towards embedding temporal querying into a third-party scenario-based test coverage framework.