What is it about?
A fault diagnosis decision fusion method based on closeness and DS theory is proposed to solve the error of fusion results caused by the conflict between diagnosis information. The membership function of normal distribution is used as the evidence closeness, and then the weight of evidence is determined according to the harmonic average, and then the weight of evidence and BPA are modified. Finally, the Dempster’s rule is used for decision fusion.
Why is it important?
Comparing with other methods, the method in this paper takes into account the overall average distribution degree of evidence in the identification framework, reduces the conflict of evidence while retaining important diagnostic information, and the algorithm has better decision fusion accuracy and lower algorithm complexity.