An evidential sensor fusion method in fault diagnosis

Abstract

Dempster-Shafer evidence theory is widely used in information fusion. However, it may lead to an unreasonable result when dealing with high conflict evidence. In order to solve this problem, we put forward a new method based on the credibility of evidence. First, a novel belief entropy, Deng entropy, is applied to measure the information volume of the evidence and then the discounting coefficients of each evidence are obtained. Finally, weighted averaging the evidence in the system, the Dempster combination rule was used to realize information fusion. A weighted averaging combination role is presented for multi-sensor data fusion in fault diagnosis. It seems more reasonable than before using the new belief function to determine the weight. A numerical example is given to illustrate that the proposed rule is more effective to perform fault diagnosis than classical evidence theory in fusing multi-symptom domains.

Weighted Evidence Combination Based on Distance of Evidence and Entropy Function

“…To preprocess data, such as Murphy's simple average in [23], Deng et al 's weighted average on the basis of distance of evidences in [26], and Han et al 's modified weighted average in [27], and to amend the combination rule, such as Smets' unnormalized combination rule [24], Dubois and Prade's disjunctive combination rule [25], and Yager's combination rule [3], are two main methods to fix up highly conflicting evidences combination efficiently. It should be pointed that there are some other combination rules to address the dependence evidence issue, which may be another alternative to handle conflicts [31][32][33][34].…”

A New Interval Numbers Power Average Operator in Multiple Attribute Decision Making

“…Three interval number decision matrices e k = (a (k) ij ) 4×4 are given as Equation [25][26][27]. Three interval number decision matrices e k = (a (k) ij ) 4×4 are given as Equation [25][26][27].…”

“…How to aggregate the given data efficiently and obtain the final decision properly has been discussed heatedly. [22][23][24][25] Yager 6 introduced the power average operator with the consideration of the relationship between the values to be aggregated. 20,21 It is difficult to give merely clear numbers under the influence of complexity of the object things, limitation of our cognition, and fuzziness of human minds.…”

Physarum solver: a bio-inspired method for sustainable supply chain network design problem

“…Besides, some available theories, e.g. evidence theory, probabilistic graphs, have been leveraged to address the uncertainty arising in various optimization problems (Jensen and Nielsen 2013;Jiang et al 2016a, b;Ning et al 2016;Yao et al 2013).…”