A new context correctness measure CMoC and corresponding context inconsistency elimination algorithm

Context-aware systems (CASs) have become increasingly prevalent in modern-day settings such as digital home and smart healthcare, primarily owing to the expansion of the Internet of Things (IoT). High-quality contexts play a vital role in the effectiveness of CASs, especially in applications involving multiple sensors. When multiple sensors provide identical information, utilizing various sources of context can enhance the dependability of said contexts through the provision of mutual support. However, there is a high possibility that multiple contexts conflict, which can lead to a problem of context inconsistency. Various methods that include quality of context (QoC) parameters have been employed to solve the issue of context inconsistency. One of the limitations of these methodologies is their inadequate consideration of the context correctness. Moreover, when attempting to address context inconsistency through the utilization of QoC parameters, the conventional Dempster-Shafer (D-S) evidence theory may result in significant conflicts. To tackle the aforementioned concerns, a comprehensive measure of correctness (CMoC) is introduced to offer a more efficient assessment of the correctness of contexts. In light of the limited flexibility in assigning probabilities to contexts using D-S evidence theory, the two-dimensional mass function (TDMF) is introduced. This method allows for assigning probabilities to contexts from various sources. On this basis, a context inconsistency elimination algorithm based on CMoC and TDMF is proposed. The effectiveness of the proposed algorithm has been verified through experimental analyses conducted across various dimensions. Based on the experimental findings, it has been demonstrated that the proposed algorithm’s context-judge rate can achieve 97.08%, provided that the high precision configuration of sensors, as outlined in this paper, is utilized. This achievement surpasses that of other inconsistency elimination algorithms by at least 2.05%.