An Information Entropy-Based Method of Evidential Source Separation and Refusion
Yuze Wang, Jindong Zhang, Jiale Qiao · IEEE Sensors Journal · 2019
D-S theory is widely used in decision fusion of multi-source information and uses multiplication rule to realize numerical fusion between different sensors. But it is also because of multiplication rules that D-S theory is difficult to apply to data fusion between multi-sensor information with complex correlation or failure. Therefore, based on D-S theory, this paper proposes a method to eliminate the correlation between evidence sources by using physical quantities (obtained by sensor detection). The algorithm convolutes the sensor's evaluation information entropy with the sensor's Gauss anomaly distribution function to deal with the anomaly. The algorithm expands the application of D-S evidence theory in multi-sensor information fusion, and has the ability to deal with sensors that report anomalies.