An Agent-based Context Awareness Management Framework

Context awareness drives adaptability of pervasive computing systems both in virtual and physical environments that change dynamically. This paper presents a context awareness management framework based on intelligent software agents. The fundamental element of this framework is our RAP (Resources, Actors, Policies) context model which defines two ways for context information representation: set based and ontology based. The RAP context model set based representation is used by the context management agents to evaluate the conditions under which management processes should be executed. The ontology based representation is used for reasoning and learning purposes, to infer new context related information. The context model management infrastructure is implemented using intelligent software agents that generate and administrate the context model artifacts at run time. The proposed context management framework is tested and validated in our DSRL (Distributed Systems Research Laboratory) smart environment infrastructure. Keywords-intelligent agents; context awareness; context management; smart environments.