Abstract: In order to better solve the fusion problem of high conflict evidence, an improved D-S algorithm fusing three pieces of evidence directly was proposed. Firstly, the calculation method of the support of each focal element under the identification framework is given according to the function of evidence supporting measurement of similarity. Secondly, a weighted assignment algorithm of conflicting information based on D-S evidence combination rule is put forward according to the properties of the conflicting factor generated in the direct fusion process of three-dimensional evidence and the support degree of each focal element. Finally, simulation is implemented under the background of multi-sensor multi-target recognition system. Analysis and simulation show that the improved D-S algorithm based on direct fusion of three-dimensional evidence has a strong anti-interference performance, and it can fuse various types of conflicting information effectively and improve the target recognition probability.

Key words: D-S combination rule of evidence, target recognition, high conflicting information, evidence supporting measurement of similarity