摘要
针对标准D-S证据理论中存在的冲突证据合成问题,提出了一种冲突证据加权的方法。该方法将各个传感器的基本概率赋值映射到多维空间中的某个点,计算任意两点的欧氏距离,利用平均距离确定证据的权重,采用哈夫曼树对证据加权平均后再利用D-S合成规则实现信息的融合。实例论证了该方法的有效性,它能有效解决冲突证据合成的问题。
In order to solve the problems to conflict evidence combination in standard D-S theory,this paper proposed a novel approach based on evidence weighted.Firstly,it mapped basic probability assignment（BPA） to a point in the multidimensional space.Then,it computed the Euclidean distance for two arbitrary points and obtained the weight of every evidence.After weighted averaging the evidence by constructing the Huffman tree,it used the D-S combination rule to achieve information fusion.The validity of the approach was demonstrated by simulation example.The proposed approach can solve the problem to conflict evidence combination.
出处
《计算机应用研究》
CSCD
北大核心
2012年第5期1815-1817,共3页
Application Research of Computers
基金
甘肃省自然科学基金资助项目(1010RJZA046)
甘肃省财政厅资助项目(0914ZTB148)
关键词
证据理论
证据加权
欧氏距离
冲突证据
哈夫曼树
evidence theory
evidence weighted
Euclidean distance
conflict theory
Huffman tree
