A novel context inconsistency elimination algorithm based on the optimized Dempster-Shafer evidence theory for context-awareness systems

With the fast advancement of the Internet of things (IoT), context-awareness systems (CASs) have been widely used in many different fields, such as digital home and smart healthcare. However, low quality of context (QoC) usually causes the CASs to make inappropriate decisions, therefore, the context inconsistency has become an urgent problem that needs to be resolved. Although many researchers have adopted some QoC parameters to solve this problem, they do not take sufficient account of the relationships between context information sources and the impacts of the uncertainty of context information on the credibility of context information sources. In this paper, different distance measurement methods are utilized to divide context information sources into credible context information sources and incredible context information sources, on this basis, the Deng entropy is introduced to construct a new discounting factor in order to assign different discounting factors for different kinds of context information sources and a novel context inconsistency elimination algorithm based on the optimized Dempster-Shafer (D-S) evidence theory is proposed. The experimental results demonstrate that the proposed algorithm based on the Cosine distance can obtain 94.33% context-judge rate under high precision configuration of sensors. Besides, under low precision configuration of sensors, compared to the correlation coefficient based on generalized information quality (CIQ)-weighted algorithm which has the highest context-judge rate among other inconsistency elimination algorithms, the proposed algorithm based on the Cosine distance can solve more inconsistent context information.