Combine Conflicting Evidence Based on the Belief Entropy and IOWA Operator

Evidence theory is widely used in information fusion. However, how to combine highly conflicting evidence is still an open issue. In this paper, a modified average method is proposed to address this issue based on the belief entropy and induced ordered weighted averaging operator. One of the advantages of the proposed method is that both the uncertainty and reliability of evidence are considered. In addition, it provides a right for the decision maker to combine the evidence based on the requirements for the precision of the results. A numerical example is shown to illustrate the use of the proposed method and an application based on sensor fusion in fault diagnosis is given to demonstrate the efficiency of our proposed method.