Collaborative Fault Diagnosis Decision Fusion Algorithm Based on Improved DS Evidence Theory

Abstract

DS evidence theory has in obtaining a correct diagnosis when confronted with highly conflicting evidence, a collaborative fault diagnosis decision fusion algorithm based on an improved version of DS evidence theory is proposed. The algorithm builds upon the closeness of certain kinds of evidence produced by existing DS evidence theory algorithms. According to the importance of the diagnostic information, weights are assigned to reduce the conflicting information while retaining the important diagnostic information. Simulated example shows that the algorithm could reduce the impact of conflicts in diagnostic information and improve the accuracy of the decision fusion process.