Situation Awareness Based on Dempster-Shafer Theory and Semantic Similarity

Abstract

In pervasive computing environment the low-level context data provided by the sensors are usually meaningless, and thus higher-level context needs to be extracted. Situation is the semantic interpretation of low-level context, permitting a higher-level specification of human behavior in the scene and the corresponding system service. Context modeling and reasoning are the two key parts in the situation awareness. In this paper we present a multiple level architecture for context modeling, and a reasoning approach based on the Dempster-Shafer Theory (DST) and semantic similarity. The Dempster-Shafer theory is employed to analyze low-level context and eliminate the conflict among different sensors. Semantic similarity is used to reason out the higher-level context information based on the ontology. Computer simulation reveals that the proposed approach allows more efficient and accurate reasoning of higher-level context information compared to the existing approach.