Conflicting data fusion method based on differences between fuzzy information of the single subset of focal elements

Abstract

Abstract The diversity of information, the complexity of environment and the limitations of sensors generally make information sources with strong uncertainty and high conflict. However, in Dempster-Shafer (D-S) evidence theory, the conflict coefficient k cannot effectively measure the degree of conflict between two bodies of evidence (BoEs). This paper quantifies the conflict in terms of the difference in ambiguity information between single subset focal elements. A novel fuzzy Chi-Square distance is proposed, in which the differences of elements in both the belief and plausibility are comprehensively considered to avoid the problem of missing information. The metric properties (boundedness, symmetry, nondegeneracy, and triangle inequality) of the fuzzy Chi-Square distance are proved in detail. And this paper investigates the resistance to disturbance of the new conflict metric Next, a new conflicting data fusion method is derived, which can obtain more accurate data fusion results and have strong robustness. Finally, the analysis of numerical examples and practical applications verifies the effectiveness and superiority of the method proposed in this paper.