Abstract: The evidence ambiguity can fundamentally affect the fusion results, but no impeccable method has been found to measure the evidence ambiguity in evidence theory. To suppress the counterintuitive results generated in the combination of high conflicting evidences, many scholars have proposed modified combination approaches based on the correction of original evidence. However, these methods do not take evidence ambiguity into consideration. Fuzzy entropy method can effectively evaluate the ambiguity (uncertainty). Given the evidence uncertainty, this paper proposes a new weighted evidence fusion algorithm based on evidence distance and fuzzy entropy theory. The coefficients of uncertainty of each evidence can be obtained by the fuzzy entropy method. Then, the weights acquired by evidence distance are amended by the coefficients of uncertainty to obtain the synthesis weights. The obtained results show the effective results of the proposed method.

Key words: evidence theory, fuzzy entropy, Jousselme distance, conflict of evidence