基于粗糙集和证据理论的目标身份融合方法

Method of target identity fusion based on rough set and evidence theory

在多传感器目标身份融合中为了解决传统证据理论存在基本概率赋值主要依靠专家利用经验给出、证据合成时将不同重要性证据同等对待等问题,提出一种基于证据权重的基本概率赋值方法.通过决策表得出传感器的证据权重与决策可信度,利用证据权重调节信度分配比例,计算出基于证据权重的基本概率赋值,从而在后续证据组合时有效增加高权重传感器影响力.实例分析表明,该方法身份融合是准确、有效的.

In multi-sensor target identity fusion, to solve the problems in the traditional evidence theory that the basic probability assignment relies mainly on the experts’giving by experiences and evidence synthesis make no discrimination on those evidences that are of different importance, this paper puts forward a basic method of probability assignment based on evidence weight. By the decision tables, we can obtain the evidence weight and decision credibility of sensors, and compute the basic probability assignment based on evidence weight, using the weight to adjust the allocation proportion of the credibility, thus improving effectively the influences of high weight sensor in the following evidence combination. By analyzing the examples, it is shown that this proposed method is of accurate identity infusion and effective.