Enabling Domain-Specific Rule-Based Automation With Semantic Stream Technology
The aim of this study is to consolidate semantics, streaming and user-centered rulebased home automation with the goal of thereby enabling more powerful and more user-friendly home automation systems. To reach this goal this thesis incorporates the rather new field of semantic streaming into home automation taking into account the users' needs and perspective. This thesis contributes a structured literature review on rule-based home automation use cases including their classification by new criteria. As part of the analysis a participatory design study was conducted yielding that users prefer visual languages in the domain of home automation. Furthermore a new pattern in semantic streaming was discovered called dynamic sensor selection. As main contribution of this thesis a home automation system based on semantic streaming called ERAS (Event and Rule Automation System) was developed using two hierarchically-aligned visual languages called Event Language (EL) and Rule Language (RL). ERAS was completely implemented also containing an editor for each of the two visual languages. As existing semantic streaming engines and benchmarks for semantic streaming engines did proof not suitable a custom semantic streaming engine called ECQELS (Extended Continuous Query Evaluation over Linked Stream) was designed and implemented as well as a corresponding benchmark introducing the average response time as relevant feature for comparison of semantic streaming engines. The system developed in this thesis was evaluated by performing the implemented benchmark on ECQELS comparing it to CQELS, probably the most powerful existing semantic streaming engine, which yielded that ECQELS can compete with CQELS for rather simple queries. When executing multiple queries simultaneously ECQELS event outperforms CQELS. Furthermore a user acceptance study was conducted as part of the evaluation comparing the visual language EL and the textual language ECQELS regarding usability. As outcome of the study EL was slightly preferred by the users in most cases but differences could not be proven significantly. On the basis of the results of this research, it can be concluded that the use of semantic streaming technologies in the domain of home automation has the potential to provide great benefits for developing more powerful and nevertheless more userfriendly home automation systems.
|Zugehörige Institution(en) am KIT
||Institut für Telematik (TM)
KITopen ID: 1000056564
||VII, II, 90 S.
||Abschlussarbeit - Master
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page