Situation recognition and interpretation based on multisensor data is an important research challenge in the situation awareness field. Existing research has developed techniques concerned with accurate and reliable situation recognition via sensor driven detection of events in an environment. However, real world applications of situation awareness require perception of a situation's meaning, knowledge of expected changes and their relevance to environments inhabitants. Recognizing the significance and implications of situations in complex real world scenarios is challenging, but is essential for designing applications for real world environments. This paper presents a novel knowledge driven approach to situation awareness. Within it we extend established data driven methods of situation recognition by utilizing domain knowledge across the entire situation life cycle. We utilize ontologies for explicit representation of environmental and application context as well as situation modeling. We explore the link between low-level environment context and highlevel application knowledge using a generic situation model. We exploit semantic reasoning to provide situation recognition and interpretation and demonstrate delivery of application oriented situation awareness in a smart environment. Finally, a case study-based scenario is utilized in order to demonstrate the system's operation.