摘要
针对信任函数理论中经典Dempster组合规则难以有效融合高冲突证据并存在焦元基模糊问题,提出了一种基于信任函数理论的修正融合目标识别算法。修正融合算法在对相容命题进行组合时,考虑了焦元基的影响,使基本信任质量合理地向基数较小的焦元命题聚焦,以避免焦元基模糊问题;在对冲突命题进行组合时,对命题进行倾向性分析并对局部冲突采用局部分配的策略,以有效融合高冲突证据。算例与仿真比较分析验证了此修正融合目标识别算法的合理有效性和优越性。
The Dempster＇s rule within the belief function theory can produce anti-intuitive results to combine high conflictive evidence. Moreover, its ignorance of the focal element cardinality leads the confusion problem. To account for these two problems, this paper proposed a modified fusion algorithm for target recognition. For the combination of consistent propositions, the basic belief mass can be reasonable transferred to the focal element with little cardinality （through considering the influence of their cardinality） to avoid the confusion problem in the proposed algorithm; and for the combination of inconsistent propositions, it adopts a local re- distribution strategy for local conflict based on their preference. The effectiveness and advantage of the proposed algorithm are veri- fied by several calculable examples and simulation results.
出处
《电子技术应用》
北大核心
2015年第6期84-87,共4页
Application of Electronic Technique
基金
国家自然科学基金(61401486)
关键词
信息融合
信任函数理论
高冲突证据
组合规则
information fusion
belief function theory
high conflictive evidence
combination rule
