A framework of context-awareness for ubiquitous computing middlewares

In upcoming fature, ubiquitous computing environments allow users to provide appropriate services or information which intelligently interact with omnipresent computing resources as user's needs in anywhere and any-time fashion. So ubiquitous computing environments need context-aware systems to gather and analyze context information from situations or surroundings. Also the systems should have abilities to handle and generate context information by combining, inferencing and learning of sensed contexts. But the previous context-aware systems are lack of systematic handling of contexts in various types, and only a few have learning mechanisms to acquire individual users 'preference or users'interests. In this paper, we propose a general framework for ubiquitous computing middlewares which easily handles and generates various types of contexts from diverse sensors. The framework is designed to sense, combine, infer, and learn context information in order to provide users with various services in changing environments. We partially implemented the proposed framework for a music recommendation system as a prototype.