基于CEEMD的压缩感知降噪在雷达信号识别中的应用研究

The CEEMD and Compressed Sensing Theory De-Noising Method and Its Application to Radar Emitter Signal Recognition

面对日益复杂的电磁对抗环境,如何有效从复杂噪声背景下提取雷达辐射源信号,成为目前热点难题。为此,提出了基于CEEMD的压缩感知理论降噪方法。原始信号经CEEMD处理,对分解得到的本征模态分量通过计算排列熵确定噪声分量和信号分量,在对噪声分量进行有用信号提取时,采用正交匹配追踪算法替代传统阈值处理方法以恢复信号稀疏性,达到降噪目的,该方法无需信号先验知识,克服了传统阈值选取主要依靠主观判断的问题。通过仿真信号分析和雷达辐射源信号识别应用研究,结果表明,所提方法较CEEMD传统阈值方法对低信噪比条件下雷达辐射源信号识别率有了较大的提升,而且减少了有用细节流失。

Facing the increasingly complex electromagnetic countermeasure environment, how to extract radar emitter signal from complex noise background effectively has become a hot issue. Therefore, a noise reduction method based on CEEMD is proposed. The original signals are decomposed by CEEMD,and the noise component and signal component are determined by calculating the permutation entropy of the decomposed components. When extracting the useful signals from the noise components, the orthogonal matching pursuit algorithm is used to replace the traditional threshold processing method to restore the signal sparsity and achieved the purpose of noise reduction. This method does not require the prior knowledge of signal and overcomes the problem of relying on subjective judgment in the traditional threshold selection. The results of simulation signal analysis and radar emitter signal recognition show that the proposed method can improve the recognition accuracy of radar emitter signal effectively, and reduce the loss of useful details.