Data driven modeling of the simultaneous activities in ambient environments Case study: automatic data driven anomaly recognition for interruptive activities in smart home

Resident of a smart home, who may be an old person or an Alzheimer patient needing permanent assistance, actuates the world by realizing activities, which are observed through the embedded sensors of smart home. Typically, this person may sometimes forget completion of the activities; may realize the activities of daily living incorrectly, and may enter to dangerous states. In order to provide automatic assistance for the smart home resident through the embedded electronically controllable actuators and make the smart home resident able to live independently at home we propose to calculate a possibilistic logical space for correct realization of activities, which may be represented in form of a multivariable problem. Regardless from the physical entity (modality and location) of the intelligence source and the quantity of individuals who perform the activities; per each possible goal or activity, we consider a unique source of intelligence (for example a social mind) who directs the order of fuzzy events that occur in the ambient environment, then the plan behind world actuations is modeled applying extensions of the fuzzy logic. The main key point that we deal with is the analysis of the observations in order to make inferences about possible simultaneous activities that may be planned and realized by one or more individuals; so that we can reason in the cases the parallel activities are interrupted.