可拓学理论在信息融合中的应用研究

Information fusion algorithm based on extenics and D-S evidential reasoning

基本概率分配函数的选取是实际应用D-S证据推理的信息融合算法的难题之一。探讨了可拓集理论和关联函数的本质,提出对关联函数进行归一化处理,并以归一化后的数据作为证据体对分类命题的可信度分配,利用Dempster组合规则生成总体可信度分配,实现多源信息的融合。仿真计算结果表明,该算法是可靠的和有效的。

The selection of Basic Probability Assignmen（tBPA） function is difficult in the information fusion algorithm of D-S evidential reasoning.This paper studies the theory of extenics set and the essence of extenic correlation function,puts forward the standardization of extenics correlation function and considers the BPA function of evidential body for classified proposition.Dempster combination method is applied to get total confidence assignment,thereby multi-resource information fusion is realized.The results of simulating test show that this algorithm is dependable and effective.