The demographic change in the European society leads to an increased demand for assistance and support systems for elderly people. To cope with the upcoming superannuation of the population a variety of projects for developing Ambient Assisted Living (AAL) solutions were funded. The usage of multiple sensor sources can increase the accuracy and fault tolerance of an ambient system, but increases the effort and complexity for decision making as well. In ambient assisted environments, where the decision making process has to consider different user specific behavior, complexity becomes an issue. Therefore, we propose the usage of a semantic layer, which encapsulates user specific information and enriches sensed user interactions with semantics. Within this paper we present an approach, which (i) allows to verify sensor events in a user specific context during decision making and (ii) ensures that actions are triggered corresponding to the user's needs. To demonstrate the feasibility of our approach we implemented the use case of detecting and handling a critical household event by using a simulation environment to evaluate different scenario set ups.