A Model for Reasoning about Situations Using Uncertain Context in a Mobile Peer-to-Peer Environment

The proliferation of mobile computing, sensing and networking technologies have enabled mobile devices to individually sense context information from their environment and develop situation awareness capability. To have a better understanding of the environment, independent devices can collaboratively capture context information, and exchange the information by establishing a mobile peer-to-peer environment (MP2P). To develop the situation awareness capability for mobile devices in this environment, we propose CoMoS (Context Mobile Spaces) as a theoretical model to represent context and reason about situations for ubiquitous application in an MP2P environment in this report. To integrate multiple context information and infer the most probable situations, a data fusion technique based on the Demspter-Shafer (DS) theory of evidence was applied in the proposed model. To deal with issues of reliability of context information used for situation reasoning, several models to estimate quality of context information were proposed. Finally, we demonstrated and evaluated the capability and performance of the proposed model, as the basis of situation reasoning through the use of context information captured by multiple devices in our test bed environment.