OoDAAS: Ontology-Driven Analysis for Self-Adaptive Ambient Systems

In the current decade, computing turns towards smart, mobile and ambient devices. As a consequence, applications and services have to be aware in order to react to a highly dynamic context (available resources, environmental conditions, user needs, etc.) and adapt their structures and behaviors. In such context-aware environment (e.g. smart buildings) interoperability is highly recommended and especially for contexts where (i) connected entities in a particular space need to understand each other and (ii) an analysis module can be built based on exchanged data querying. In this paper, we propose an ontology model for smart building that allows monitoring smart entities and enables a rule based analysis. The proposed ontology model defines seven major reusable concepts that provide the basic infrastructure to build an ontology model for smart space environments. As each smart environment is distributed over different locations and may have many kinds of devices etc., we propose domain ontology for the smart building space and implement a prototype for this ontology model using Protégé 3.5 as an ontology editor and Pellet as a reasoning engine. Furthermore, we define smart rules using SWRL (Semantic Web Rule Language) and infer those rules by SQWRL (Semantic Query-Enhanced Web Rule Language) query language in order to analyze the context and generate alerts regarding abnormal behaviors. The proposed ontology model is illustrated in a smart building space where the analysis infers and detects alerting situations, in a wireless sensor network, based on monitoring data collected from sensors.