学术
摘 要:
在分析D—S证据组合公式所存在问题的基础上，根据聚类分析思想，提出了对证据进行聚类加权的证据组合方法。该方法将较为相似的证据归为一类，根据聚类结果对证据有区别地进行组合，不同类之间的证据根据聚类距离进行加权后组合，同一类证据按一种加权证据组合公式组合。应用到目标识别系统后使其性能在一定程度上得到改善。最后通过计算实例表明算法是有效的。[著者文摘]
文章出处:
《现代防御技术》-2008年36卷1期 -80-83页
栏目信息:
文献标识码:
A
文章编号:
1009-086X（2008）-01-0080-04
Multiple Source Sensor Evidence Combination Research Based on Evidence Cluster
XIE Bin, LIU Fu-xian, XING Qing-hua （ AFEU, the missile institute, Shaanxi Sanyuan 713800, China）
Abstract:
Base on analyzing the disadvantage of D - S theory, employing the idea of clustering analysis,a new evidence combination method was proposed. This method range the similar evidence in one catalog, recombine it according to the clustering result. The evidence in different catalog is recombined after giving them a weight modulus which is consulted to the clustering distance; The evidence in the same catalog is recombined by the original combination formula. Checking this method on the target recognition system, it worked well. At last this method was proved to be effective in the final example calculation.[著者文摘]
Key words:
D - S evidence theory ; evidence clustering ; target recognition
更多>>免费文章
cqvip.com