自动化学报Issue(5)：914-927,14.DOI:10.16383/j.aas.2015.c140681
一种基于马尔科夫链的冲突证据组合方法
A Method of Conflictive Evidence Combination Based on the Markov Chain
摘要
Abstract
Aiming at the problem that highly conflictive evidence can not be processed by Dempster rule in intelligent information processing, a method of conflictive evidence combination based on Markov chain is proposed by considering the high-efficiency anti-interference performance for the sequentiality of sequential evidences. At first, the deterministic state description in the classic Markov chain is extended to nondeterministic state description. And then, the past evidences are sampled sequentially according to the sliding window width l, which could be amended according to the weight computed by utilizing the similarity measure. A Markov model is established on these past evidences amended so that a transition probability matrix could be obtained, which is used to compute the evidential representative. Finally, this representative is combined with itself for l−1 times according to the Murphy0s combination method. Of course, this method also fits parallel fuse in a step. Through simulation experiments, the comparisive analysis show that the new method0s advantage is obvious. That is to say, it efficiently solves the problem of the combination of conflictive evidences; moreover, it keeps robustness and sensibility of combinational result.关键词
证据推理/冲突/马尔科夫链/状态不确定性/组合规则Key words
Evidence reasoning/conflict/Markov chain/state-uncertainty/combination rule引用本文复制引用
李新德,董清泉,王丰羽,雒超民..一种基于马尔科夫链的冲突证据组合方法[J].自动化学报,2015,(5):914-927,14.基金项目
国家自然科学基金(60804063,61175091),航空基金(20140169002),江苏省“青蓝工程”资助计划,江苏省“六大高峰人才”资助计划资助Supported by National Natural Science Foundation of China (60804063,61175091), Aeronautical Science Foundation of China (20140169002), Qing Lan Project of Jiangsu Province, and Six Major Top-talent Plan of Jiangsu Province （60804063,61175091）
