基于加权IFS证据合成的多源信息融合方法

Multi-source information fusion based on weighted IFS evidence combination algorithm

摘要:

为权衡信息融合系统对各条证据的可信程度，提出基于直觉模糊交叉熵的证据折扣方法，用于直觉模糊证据合成规则。并根据确定性理论，对直觉模糊决策中的记分函数进行改进，得到基于确定性因子的记分函数形式，用于直觉模糊证据合成结果判定。最后通过与其他算法的对比实验，验证算法在多传感器目标识别应用中是有效的，且具有较高的计算效率。

Abstract:

A method for determining the discounting factor of each intuitionistic fuzzy evidence based on intuitionistic fuzzy cross-entropy is proposed to treat the reliability among the evidences for different sources and avoid the conflicts. It can be used in the intuitionistic fuzzy evidence combination to improve the efficiency of D-S combination algorithm. The score function simply used in the intuitionistic fuzzy decision is improved according to the certainty factor to discuss the final results of the combination. Experiments comparison with other approximations indicates the proposed algorithm is of effectiveness and high-performance in the multi-source information fusion.