Advanced fuzzy inference engines in situation aware computing

We focus on the very important family of context-aware applications. Context-aware computing relies on tasks like capturing/sensing environmental parameters (e.g., lightness, location), classifying context, and inferring further knowledge about that context (determine the situation the user is currently in). However, the relevant applications have to deal with the inherent imperfection of context sensing for decision making. We propose an extension of context representation and inference for situation-aware applications. Our model relies on Fuzzy Set Theory to accommodate the imperfection of sensed context. Based on this model, we have developed three fuzzy inference engines, which rely on advanced semantics (specialization, mereological and compatibility relations). We have evaluated the proposed engines through a series of experiments involving real users. Our findings indicate the strong points of the proposed context classification and inference processes.