Enhancing applicability of context-aware systems using agent-based hybrid inference approach

Context awareness promises to enhance the everyday life of mobile users. It will allow people to access information and services related to their current situations without the need to manually configure and instruct computer systems. To make this possible, we are proposing an agent-based, context aware framework that makes use of sensing technologies and wireless networking. In specific, the framework addresses the challenge of finding a common understanding of context and uses an approximate reasoning mechanism to provide higher level of context information used in personalizing services in the environment. The paper argues that, in order to facilitate context awareness, a set of core services is required. These services constitute of meta-context ontology, an inference engine based on fuzzy logic, and an inductive learning mechanism. Combined, these functionalities will provide a reconfigurable, interoperable, and adaptable framework that support context aware applications and the spontaneous situations occurring from explicit and implicit users' interactions with the environment.