Context-aware and resource efficient sensing infrastructure for context-aware applications

Middleware for wireless sensor networks and middleware for context-aware applications both provide information abstraction and programming support for gathering, pre-processing, and managing sensor data. However the former mostly concentrates on optimising the operations of the resource constrained hardware and simplifying access to the raw sensor data while the latter focuses on gathering sensor data, pre-processing it to the abstract context information required by the applications and providing reasoning on this data. In this paper, we explore the idea of enhancing middleware for context-aware applications with solutions from sensor networks middleware to allow resource efficient and context-aware management of sensing infrastructure. The decisions on which sensor data needs to be delivered to the middleware for evaluation are based on current contextual situations. The approach allows to trade the level of confidence in context information for resource efficiency in context provisioning without a detrimental effect on the functionality of context-aware applications.