D-S证据理论可以有效表示随机性和模糊性的不确定信息,可以在不需要先验信息的情况下对证据进行组合,这些优点使得该理论被广泛应用于信息融合的各个领域。但是在实际应用中,D-S证据理论中的组合规则无法有效融合冲突信息,为此研究人员在组合规则改进或是数据模型修改两个方面提出了很多改进方法,但却忽视了对冲突本身的研究。首先分析了证据理论中经典冲突系数的不足,而后引入证据距离,结合经典冲突系数提出了一个改进的二元组来描述证据冲突,新的二元组冲突表示可以更好地度量证据之间的冲突,而且对Dempster组合规则的适用性具有指导意义。算例说明了新的冲突表示方法的有效性。 D-S evidence theory can effectively represent random information and ambiguity uncertainty information, and can combine evidence without prior information. These advantages make the theory widely used in all fields of information fusion. However, in practical applications, the combination rules in DS evidence theory can not effectively fuse the conflict information. Therefore, researchers have proposed many improvement methods in terms of combinatorial rule modification or data model modification, but ignored the research on conflict itself . Firstly, the insufficiency of classic conflict coefficient in evidence theory is analyzed, and then the evidence distance is introduced. Combined with classical conflict coefficient, an improved two-tuple is proposed to describe the evidence conflict. The new two-tuple conflict representation can better measure the difference between evidence Conflicts, but also the applicability of the Dempster combinatorial rules are instructive. The example illustrates the validity of the new conflict representation method.