Extraction of implicit context information in ubiquitous computing environments

Abstract

The evolution of low-cost, networked sensors, often directly internet-enabled, is bringing truly ubiquitous smart environments into daily life. The more ubiquitous middleware platform is intelligent, the greater context information flood problem has been caused. Hence, there have been increasing demands for efficient methods of discovering desirable knowledge from a large collection of context data. But unfortunately, current ubiquitous middleware platforms do not employ appropriate data mining techniques to meet such growing demands. Therefore, this paper aims to propose a new design of ubiquitous middleware platform that enhances context awareness in evolving pervasive environments. We achieve this goal first by incorporating a mining module into our previously suggested middleware platform CALM (Component-based Autonomic Layered Middleware) and then by instantiating the module with an efficient mining algorithm. © Springer-Verlag Berlin Heidelberg 2006.