Context Information Management for Pervasive Computing

Abstract. Pervasive Computing systems have to deal with the con-textual information (context), which characterizes the current situation of the involved entities (e.g., users, mobile devices, environment, etc.). This dissertation studies context management issues related to the ca-pability of a pervasive system on adapting its behavior to the involved entities context. Specifically, the interaction between the user and such system has to be less intruding as long as the latter recognizes the current user situation and adapts its functions accordingly. Such issues comprise the concept of Context Awareness. The dissertation focuses on context knowledge representation and reasoning as well as on approximate rea-soning (Fuzzy Sets Theory). The management of lower level environmen-tal information that emanates from sensors is also of great importance and is achieved with novel data fusion and decision fusion techniques that are proposed. Moreover, have been studied issues regarding collab-orative context awareness and reasoning as well as bio-mimetic contex-tual dissemination. Adaptive algorithms are proposed so that is rendered possible the efficient dissemination of information in distributed environ-ments. The objective is the minimization of communicating costs and the enhancement of context quality. Consequently, have been designated issues such as context discovery, context representation and inference, context fusion and collaborative context awarenes