Context as autonomic intelligence in a ubiquitous computing environment

This paper presents the ANS architecture that uses ubiquitous computing to monitor medical patients at home. Since there is no notion of the patient carrying out maintenance of such a system, it must be self-managing or autonomic. In the ANS, sensors such as temperature, location, etc., use a form of logic to abduce a context, i. e., the state/quality of a given device or its function. Our contribution lies in the emergent autonomicity of the architecture driven by its ability to derive the most appropriate source for a particular application. This is done by allowing the application to define mathematically its own notion of what level of service provides the best satisfaction and is based on Quality of Context (QoC) attributes that describe each alternative. The ANS framework is lightweight and can provide real-time adaptation, which is necessary in resource-starved ubiquitous computing environments that support medical applications.