开放世界下的雷达辐射源融合识别算法研究

针对开放世界下雷达辐射源目标库不完备的现状,对开放世界下雷达辐射源识别的信息融合模型进行了研究。建立了不同类型雷达特征参数的模糊隶属度函数,采用待测样本与模型样本匹配的方法生成广义概率指派函数,并采用修正的广义证据理论算法融合多参数信息获取识别结果。实验数据表明:该模型能在雷达辐射源目标库不完备的情况下,识别已知目标和判别未知目标,判别结果相对基于经典D-S证据理论和原始广义证据理论的融合识别方法更加可靠、有效,在雷达识别的场景中具有潜在应用价值。

Fusion Algonithm of Radar Emitter Recognition in Open World

Target recognition is an important link in the chain of information processing for air defense. The exact recognition of radar emitter under uncertainty is significant to target recognition. In open world, the radar library is usually incomplete, and traditional target recognition method cannot identify new targets. To address this issue, in this paper, a fusion model of radar emitter recognition is studied in open world. Firstly, the operation of fuzzification for characteristic parameters of radars with different features are made and basic probability assignment is obtained by mode recognition. Finally, the evidence provided by different parameters are fused by generalized evidence theory. The experimental results demonstrate that the proposed method is more reliable and effective in the ability of recognizing the known target and distinguishing the unknown target, compared to the methods based on the traditional D-S evidence theory and generalized evidence theory. This work has a potential application value in the radar identification scenario.