论文部分内容阅读
证据理论是常用的一种决策级信息融合方法,能有效处理不确定性信息.然而该方法在融合冲突证据时却可能会得到与常理相悖的结论.针对这一问题,对已有解决方法进行了总结,在此基础上提出一种新的折扣方法,综合考虑了证据之间的相互冲突情况和各证据自身的区分能力.鉴于冲突、区分能力等概念的定义具有模糊性,引入模糊集合予以表示,定义了它们的模糊隶属函数,并设计了相应的模糊推理规则以获取折扣值.算例表明,新方法对一致性证据影响甚微,而对冲突证据能够进行有效处理,得到合理的融合结果.
Evidence theory is commonly used as a decision-level information fusion method, which can effectively deal with the uncertainty information.However, the method may be contradictory to the common sense when the evidence of conflict is merged.To solve this problem, The author puts forward a new discount method based on this, which takes into account the mutual conflict between evidences and the distinguishing ability of each evidence.With the ambiguity of the definition of concepts such as conflict and differentiation ability, fuzzy sets are introduced to give The fuzzy membership functions are defined, and the corresponding fuzzy inference rules are designed to obtain the discount value.Examples show that the new method has little effect on the consistency evidence, and the evidence of conflict can be effectively processed to obtain a reasonable fusion result.