A Probabilistic Non-monotonic Activity Qualifier

The International Classification of Functioning, Disability and Health (ICF) defines Functioning and Disability as the results of the interaction between the health conditions of a person and his/her environment. It considers a set of components and qualifiers to evaluate activity and participation. In this paper, we interpret a performance quantifier under a human activity recognition process. To this end, we introduce a novel definition of an activity which is based on ICF guidelines. This definition gives place to a probabilistic non-monotonic activity qualifier. In order to recognize an activity according to our novel activity's definition, we explore non-monotonic reasoning technics to capture domain knowledge in terms of action specification languages. By considering an action specification language, called CTAID, and Answer Set Programming, we propose and develop a system called ActRec system which takes background information into consideration and recognize activities according to our suggested definition. Moreover, we show that by aggregating our probabilistic non-monotonic activity qualifier, we are able of detecting complex activities, e.g., long-term activities. We illustrate our approach in the context of an ambient assisted living environment called As-A-Pal