基于直觉模糊集的时域证据组合方法研究

证据理论已广泛应用于时空信息融合领域，由于时域信息融合表现出明显的序贯性和动态性，为实现基于证据理论的时域信息融合，有效处理时域冲突信息，结合证据可靠性评估和证据折扣的思想，在直觉模糊框架内提出了一种基于复合可靠度的时域证据组合方法.首先定义一种基于可靠度的直觉模糊数排序方法，在此基础上提出一种基于直觉模糊多属性决策的证据可靠性评估方法;然后，基于此方法对时域信息序列中相邻时间节点的证据可靠性进行评估，得到时域证据的相对可靠性因子;最后，结合由时域证据可靠度衰减模型得到的实时可靠性因子，得到时域证据的复合可靠性因子，再基于证据折扣运算和Dempster证据组合规则提出一种基于复合可靠度的时域证据组合方法.数值算例和仿真表明，该方法具有较强的时间敏感性，充分体现了时域信息融合的动态性特点，可以较好地处理时域证据中的冲突信息，基于该方法构建的融合识别系统具有较强的抗干扰能力.

Combination of Temporal Evidence Sources Based on Intuitionistic Fuzzy Sets

Evidence theory has been widely used in spatial and temporal information fusion. The sequential and dynamic characteristics of temporal fusion calls for a new combination rule of temporal evidence sources. In this paper, temporal evidence combination is analyzed in the framework of evidence reliability and evidence discounting. A method of temporal evidence combination is proposed based on the composite reliability factor of temporal evidence. A ranking method for intuitionistic fuzzy values is firstly presented, followed by the presentation of evidence reliability evaluation based on intuitionistic fuzzy multiple criteria decision making. Then the relative reliability factors of evidence sources in neighboring time nodes are evaluated. By combining the relative reliability factor and real-time reliability factor yielded by the credibility decay model, a composite reliability factor is obtained. Finally, according to evidence discounting and Dempster's combination rule, a method of temporal evidence combination based on the composite reliability factor is proposed. Numerical examples and simulation demonstrate that the proposed method is time sensitive, which can reflect the dynamic nature of temporal information fusion. Moreover, it is illustrated that this method can deal with conflict in temporal fusion well. The proposed temporal evidence combination rule can enhance the anti-interference performance of the target identification fusion system.