A Novel Batch-Mode Evidence Combination Approach

Dempster's rule of combination in Dempster Shafer evidence theory is widely used in many applications of multi-sensor fusion to implement the combination of multiple pieces of evidence given by independent information sources. However, Dempster's rule of combination sometime may bring out counter-intuitive or illogical results especially when the evidences to be combined are highly conflicting. Such counter-intuitive results may arouse many controversies about its validity. To solve this problem, several modified or improved evidence combination approach were proposed. In this paper, a novel batch-mode evidence weighted combination approach is proposed. For all the bodies of evidence to be combined, the pair-wise combinations are implemented at first. Then the weighted average of the pair-wise combination results can be obtained as the final combination result, where the weights are generated according to the distance of evidence. Some numerical examples are provided and the experimental results derived show that the proposed approach is rational and effective.