CDTOM: A Context-driven Task-oriented Middleware for Pervasive Homecare Environment

Abstract

With the growing number of the elderly, we see a greater demand for home care, and the vision of pervasive computing is also floating into the domain of the household that aims to build a smart home which can assist inhabitants (users) to live more conveniently and harmoniously. Such health-care pervasive applications in smart home should focus on the inhabitant's goal or task in diverse situations, rather than the various complex devices and services. The core challenge for homecare design is to perceive the environment and assess occurring situations, thus allowing systems to behave intelligently according to the user's intent. Due to the dynamic and heterogeneous nature of pervasive computing environment, it is difficult for an average user to obtain right information and service and in right place at right time. This paper proposes a context-driven task-oriented middleware (CDTOM) to meet the challenge. The most important component is its task model that provides an adequate high-level description of user-oriented tasks and their related contexts. Leveraging the model multiple entities can easily exchange, share and reuse their knowledge. Based on the hierarchy of task ontology, a novel task recognition approach using CBR (case-based reasoning) is presented and the performance of task recognition is evaluated by task number, context size and time costing. Moreover, a dynamic mechanism for mapping the recognized task and services is also discussed. Finally, we present the design and implementation of our task supporting system (TSS) to aid an inhabitant's tasks in light of his lifestyle and environment conditions in pervasive homecare environment, and the results of the prototype system show that our middleware approach achieves good efficiency of context management and good accuracy of user's activity inference, and can improve efficiently quality of user's life.