The quality of data in pervasive environments is the focus of this paper, with an emphasis on medical applications. The paper asserts that the next generation of pervasive computing environments will require intelligent real-time context-sensitive data management systems, such as health-related systems that gather data from multiple sensors in real time, including mobile devices. Pervasive environments generate large datasets, and in the case of medical systems the massive amount of data often must be processed in real time. The authors describe a model for management systems to support the collection and management of data, and appropriate data infrastructure to increase the quality of data. This management system model is sensitive to the context of the data coming from continually validated sensors, some of which may be mobile and wireless.

The authors present two components of data management system architecture: data validation and data consistency. The data management system consists of agent-based software components that deliver an increase in the quality of service through patient sensor validation and context-sensitive data consistency from the various input sources. These agent-based components are used to capture the environment’s ability to link the middleware-layer software to the front-end applications. The system is made aware of the environment of the data collection devices by including such components as accuracy of sensor location, monitoring, and uncertainty modeling for components to reduce, for instance, false alarms in life support systems. The software agents are shown to be capable of proactive and reactive executions to increase the quality of the data provided to the medical environment. The model was tested against a number of accepted medical test cases. In particular, the validation of patient sensor readings, including wireless mobile sensors, is shown to achieve results comparable to standard bedside devices with no wireless mobile sensors.

The paper requires the reader to have significant knowledge of the software models and environments that are used when creating software for pervasive computing environments.