针对多传感器目标识别中证据存在非一致性问题, 提出一种面向目标识别的基于证据置信熵与相似性的多传感器证据融合算法。首先, 利用传感器证据非一致性不确定度与非特异性不确定度, 提出了一种基于置信熵的传感器证据不确定性度量模型; 在此基础上, 结合传感器证据距离与证据冲突, 设计了一种基于置信熵与相似性的传感器证据权重生成方法; 最后, 构建了一种多传感器证据融合模型。仿真结果表明, 所提出的多传感器信息融合方法进行目标识别时, 相较于传统算法具有更好的有效性。

Aiming at the inconsistency of evidence in multi-sensor target recognition,a multi-sensor evidence fusion algorithm based on evidence belief entropy and similarity is proposed for target recognition.Firstly,a measurement model of sensor evidence uncertainty based on belief entropy is introduced by using the inconsistent uncertainty and nonspecific uncertainty of sensor evidence.On this basis,a method of generating sensor evidence weight based on confidence entropy and similarity is designed by combining the distance and conflict of sensor evidence.Finally,a multi-sensor evidence fusion model is constructed.The simulation results show that the proposed method is more effective than the traditional algorithm in target recognition.