· · · Repository hosted by TU Delft Library

This paper presents an approach to structuring knowledge and reasoning for high-level interpretation of sensor data in e.g. independentliving applications. The main contribution is to use generalized events, described in terms of ‘space-time chunks’, as a unifying and simplifying structuring principle. We use reasoning with ontologies and rules in combination with a database system, and also incorporatenumerical computation. We show that an easy to use modeling formalism is obtained, and that reasoning is feasible at the time of servicerequest, by using R-entailment, which enables efficient exploitation of ontologies and rules in the presence of RDF data. Two applications were built and evaluated using the approach described in this paper, both of which are related to monitoring well-being of elderly people, and both of which use simple, low-cost sensors.

An awareness system was designed to provide peace of mind and a sense of connectedness to adults who care for an elderly parent living alone. The iterative design of the Aurama awareness system showed andour empirical research, including field trials ranging from four tosix months, confirm the potential of awareness systems to support both generations suggesting that future research should examine, firstly, how to convey long-term trends regarding the wellbeing of the elderly and, secondly, how to intertwine the communication of awareness information with expressive forms of communication. We further explore the role that formal care providers can play when caring for the elderly using the awareness system. We discuss implications of our studies for the design of ambient intelligent systems supporting awareness between elderly, their adult children and care providers, and derive several lessons learned in conducting long-term field trials.