Pervasive computing is allowing healthcare to move from care by professionals in hospital to self-care, mobile care, and at-home care. The pervasive healthcare system, CARA(Context Aware Real-time Assistant), is designed to provide personalized healthcare services for chronic patients in a timely and appropriate manner by adapting the healthcare technology to fit in with normal activities of the elderly and working practices of the caregivers. This paper presents a fuzzy-logic based context model and a related context-aware reasoning middleware that provides a personalized, flexible and extensible reasoning framework for CARA. It provides context-aware data fusion and representation as well as inference mechanisms that support remote patient monitoring and caregiver notification. Noteworthy about the work is the use of fuzzy-logic to deal with the imperfections of the data, and the use of both structure and hierarchy to control the application of rules in the context reasoning system. Results are shown for the evaluation of the fuzzy-logic based context reasoning middleware under simulated but realistic scenarios of patient monitoring. The results indicate the feasibility of the system for e_ective at-home monitoring.
