Obtaining continuous and detailed monitoring of indoor environments has today become viable, also thanks to the widespread availability of effective and flexible sensing technology; this paves the way for the design of practical Ambient Intelligence systems, and for their actual deployment in real-life contexts, which require advanced functionalities, such as for instance the automatic discovery of the activities carried on by users. Novel issues arise in this context; on one hand, it is important to reliably model the phenomena under observation even though, to this end, it is often necessary to craft a carefully designed prototype in order to test and fine-tune the theoretical models. The work described here proposes to use sensor nodes to capture environmental data related to users' activities; the representation of the environment will rely on an ontology expressed in a well-established ad-hoc formalism for sensor devices. An activity model will be produced by analyzing the effect of users' actions on the collected measurements, in order to infer the underlying structure of sensor data via a linguistic approach based on formal grammars. It will finally be shown how such model may be profitably used in the context of a hybrid simulator for wireless sensor networks in order to obtain a scalable and reliable tool. © Springer International Publishing Switzerland 2014.
CITATION STYLE
Ortolani, M. (2014). Extracting structured knowledge from sensor data for hybrid simulation. Advances in Intelligent Systems and Computing, 260, 153–165. https://doi.org/10.1007/978-3-319-03992-3_11
Mendeley helps you to discover research relevant for your work.