Towards an Interactive Personal Care System driven by Sensor Data

This demo presents an integrated application, exploiting technologies derived from various branches of Artificial Intelligence. The prototype, when triggered by an appropriate event, interacts with a person and expects them to perform a simple acknowledgement gesture: the following actions, then, depend on the actual recognition of this gesture. In order to achieve this goal, a combination of data stream processing and rule based programming is used. While the more quantitative components exploit techniques such as Clustering or (Hidden) Markov Models, the higher levels, more qualitative and declarative, are based on well known frameworks such as Event Calculus and Fuzzy Logic.