A quality-aware approach for selecting context information from redundant context sources
Abstract
Sensor-rich ubiquitous computing systems could face many inconsistent decisions while selecting context data from redundant sources to compose the context, since low-quality sensor data can be selected to accordingly adapt the environment and services based on context changes. These systems need mechanisms to verify the quality of collected sensor data in order to select reliable context information and reduce the probability of making erroneous context-dependent decisions. Quality of Context (QoC) indicators linked to collected context data can be used to deal with this challenge, offering means to improve the decision-making support of context management frameworks. This paper presents a quality-aware approach for selecting reliable context data from redundant context sources based on the QoC indicator probability of correctness. The proposed approach was embedded in the decision-making support of the context fusion layer of our Context Management Framework (CxtMF), which was developed to provide QoC-enriched context information for ubiquitous healthcare systems.