基于独立分量分析的多目标识别后置处理方法

A Post-processing Method for Multi-target Identification Based on Independent Component Analysis

针对独立分量分析在多目标混合信号分离中存在排序不确定的问题,提出一种能量域后置处理方法。该方法依据基于目标辐射噪声的幅频特性不同特点,可以有效降低分离后混合信号排序的混乱概率,提高分离后混合信号排序的一致性。数值仿真处理结果表明,该方法有效克服混合信号分离后排序混乱现象,增强了独立分量分析在实际信号分离中的有效性。

In view of the independent component analysis in the multi - objective mixed signal separation exists sort of uncertain prob- lems, this paper proposes a post - processing method based on energy domain. The method based on the different characteristics of tar- get radiated noise amplitude frequency characteristics that can effectively reduce probability of confusion after sorting mixed signal sepa- ration, improve consistency of the sort of mixed - signal separation. Numerical simulation data processing results show that this method is effective to overcome tthe confusion sorting mixed signal separation, enhance the effectiveness of independent component analysis in the actual signal separation.