基于PDW多特征融合的辐射源信号分选方法

Radiation signal sorting method based on PDW multi-feature fusion

针对现有辐射源信号分选技术混叠信号处理困难、分选精度不高的问题,提出了一种针对脉冲信号描述字的多特征融合分选方法,采用分治思想,具备两级分选架构。第一步通过构建时空密度聚类模型,分离出在时频域上混叠的辐射源信号;第二步通过多参数交并比方法,进一步提升算法精度。与传统方法相比,所提方法能够有效分选出时频域混叠脉冲序列,受辐射源脉冲重复间隔变化造成的影响较小,相比传统方法具有更高的准确性。

In view of the difficulty to deal with aliasing signals and low accuracy of sorting in the field of radiation signals sorting,this paper proposes a multi-feature fusion method,focusing on pulse description word(PDW),which is a two-step process architecture based on divide-conquer method.In the first step,the problem of radial source signal aliasing in the time-frequency domain is solved by the proposed spatial-time density clustering model.The second step is to further improve the accuracy of sorting through the intersection over union method.Compared with the traditional methods,this method can effectively separate the pulse signal of the mixed time-frequency domain with higher sorting accuracy.It is robust against the variation of pulse repetition interval(PRI)of radiation source.