特定辐射源识别的频域鲁棒ε-混合模型方法

Specific emitter identification based on frequency domain robust ε-contaminated method

特定辐射源识别(specific emitter identification,SEI)是电子战领域的热点问题,而利用辐射源中频波形分类是SEI的一种自然选择。针对多径环境下的SEI波形检验遇到的难题,采用鲁棒的假设检验方法进行已知辐射源信号的检验。该算法利用多径信号与直达波信号间的频域局部相关性,在频域建立ε混合模型,并且将局部频带内的多径干扰表示成与信号频谱幅度有关的噪声,得到一种鲁棒的频域相关-限幅检验器。通过仿真验证了该检验器性能,并说明了该方法通过频域选择获得性能改善的机理。

Specific emitter identification （SEI） is a hot point in the field of electronic war. The use of emitter＇s intrapulse is a nature choice in SEI. A robust hypothesis testing method is proposed for the SEI application in the multi-channel environment. The algorithm uses the local correlation between multi-channel signals and di rect signals in frequency domain, constructs a ε-contaminated model which models the multi-channel disturb as a signal-dependent noise. The derived detector is a correlation-limiter in frequency domain. The simulation proves its performance, and indicates its mechanics of frequency selection.