Evidence Combination Based on Singular Value

Evidence theory is an efficient tool to fuse multi-source information. However, when Dempster’s rule of combination is applied to combine highly conflicting evidences, counter-intuitive results may emerge. In order to resolve such a problem, a new weighted averaging evidence combination method based on singular value was put forward. Minimum singular value can measure the conflict between two evidences more accurately than other common methods. Thus, using singular value to determine the weights of evidences is more reasonable. Based on the weighted averaged evidences and Dempster’s rule, the final results are acquired. The numerical example illustrates that the proposed method can fuse highly conflict evidences effectively and has faster convergence speed and better focusing degree than some other existing combination methods.