摘要
针对Dempster方法在合成高度冲突的证据时结果与直观结论相悖的问题,提出了一种基于支持度的改进证据合成方法。首先根据证据之间的相似性程度判断是否存在证据冲突,对于总数在3条以上且存在冲突的证据组合,应用各证据的支持矩阵,通过其最大特征值对应的特征向量导出各证据的支持度系数,由此对基本概率分配(basic probability assignation,BPA)函数进行转化,再采用Dempster方法进行合成。仿真计算结果表明,该方法能有效处理冲突证据组合,克服了一票否决问题,相对于其它几种典型改进算法具有更快的识别速度,能进一步减小识别结果的不确定性,并保留了Dempster方法满足交换律、结合律的良好特性。
To solve the problem of the unrealistic results arising in using Dempster’s method for the combi-nation of highly conflicting evidences,an improved evidence combination method is presented which is based on support degree.First of all,the similarities between evidences are used as an approach to judge whether con-flicts exist.If there are more than 3 evidence combinations with conflicts,the characteristic vector correspon-ding to the maximum characteristic value of the support matrix is used so as to obtain the evidences’support de-gree coefficients.Then,BPA functions are transformed by means of the support degree coefficients.Finally, the Dempster’s method is used for the combination.The stimulation results show that the proposed method can deal with conflicting evidence solve the problem caused by one vote veto effectively,combinations calculate faster than other improved algorithms,reduce uncertainty in recognizing results,and retain the advantages of Dempster’s method.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2014年第8期1483-1487,共5页
Systems Engineering and Electronics
基金
国家自然科学基金(61074191)
全军军事学研究生重点课题(2011JY001-018)资助课题
关键词
信息融合
冲突证据
支持度
特征值
information fusion
conflicting evidence
support degree
characteristic value
