Abstract |
Ambient Intelligence (AmI) has been one of the greatest visions of the Information Society.
It brings intelligence to people’s everyday environment making it autonomously
adaptable, proactive and context-sensitive to the users’ needs. In parallel, the evolution of
Ambient Assisted Living (AAL) field has attracted considerable attention in the scientific
community and it constitutes one of the most significant applications of AmI technologies.
AAL gives the opportunity to elderly to lead independent lives, confiding in part the care of
their health and their living in smart devices that operate in people’s most intimate environment,
their home.
The objective of the current thesis is to develop the concept of an integrated contextaware
system that combines technologies of the semantic web, with those of Ambient Intelligence,
in a way of assisting an old person with his/her daily routine. Particularly, we
exploit advantages of ontologies and semantics for creating a formal representation of context
information, which in our domain includes environment entities, user profiling and
health aspects. This information constitutes the knowledge base of our system. Three types
of reasoning are applied upon this knowledge: a first level rule-based reasoning for inferring
additional information about the context, a case based reasoning for recognizing the current
activity and a final reasoning step that undertakes to offer personalized assistance in the form
of rules expressed in a high-level rule language. The rules about the assistance conclude to
actions related with the recognized activity, with aspects of the person’s health, or with
emergency situations. Furthermore, we present a use case scenario that captures the functionality
supported by the current approach. A subset of this has also been demonstrated with
realistic conditions and users showing that it can perform in real time with compensatory
results.
|