Towards a System Architecture for Recognizing Domestic Activity by Leveraging a Naturalistic Human Activity Model
Résumé
Existing activity recognition approaches in the smart home domain suffer from poor human activity models. Combining expertise from cognitive ergonomics and ubiquitous computing, we discuss the hard technical challenges to address when leveraging a realistic model of human activity. We present the architecture of a prototype smart home system that we are developing and show the gap that exists between our current capabilities in terms of contextual-knowledge extraction and the complexity of the targeted activity recognition. To fill this gap, we propose and discuss the integration of PHATT, an existing algorithm for plan recognition, into our system in order to mine additional information from the dynamics of context.