遗传算法优化变分模态分解提取舰船辐射噪声特征线谱方法

特征线谱提取是舰船目标识别的一个重要研究环节，常采用传统的DEMON谱分析方法，处理过程中，一般对舰船噪声时域信号未予抑噪，低信噪比情况下，传统DEMON谱分析性能差。对此，提出一种采用遗传算法优化变分模态分解方法，用于分解舰船噪声原时域信号，获得抑制噪声后的舰船噪声重构信号，进而有效提取了舰船目标噪声幅度调制特征线谱。该方法首先采用遗传算法优化变分模态分解的两个关键输入参数(分解所取模态个数和惩罚因子)，对变分模态分解得到的各阶固有模态分量加以判别，去除噪声主导分量，保留信号主导分量，使重构舰船噪声信号显著抑制了干扰噪声，然后对降噪后的重构信号进行频谱分析，获得目标噪声调制特征线谱。理论分析、仿真和实验数据处理结果表明，相比传统DEMON谱分析法，基于遗传算法优化变分模态分解的舰船噪声特征线谱提取方法具有更好的噪声抑制能力，所获取的舰船噪声幅度调制特征线谱信噪比明显高于传统DEMON方法，具有一定优势，前景良好。

Feature line spectrum extraction of ship radiated noise based on variable mode decomposition optimized by genetic algorithm

Feature line spectrum extraction is an important research direction of ship target recognition. Aiming at the weakness of traditional DEMON spectral analysis in noise suppression, a method for extracting the characteristic line spectrum of ship radiated noise based on genetic algorithm (GA) optimized variational mode decomposition (VMD) is proposed. Firstly, the genetic algorithm is used to optimize the two key parameters of VMD, and the inherent mode function (IMF) of each order is obtained through VMD decomposition. Secondly, the dominant component of the signal is retained and the dominant component of the noise is removed by distinguishing the components of each order, so as to achieve the suppression effect of noise interference. Finally, the target noise characteristic line spectrum is obtained by spectral analysis of the reconstructed signal after noise reduction. Simulation and experimental data processing results show that compared with DEMON spectral analysis method, the feature line spectrum extraction method based on GA-VMD has better noise suppression ability, and the feature line spectrum extraction effect of ship radiated noise is better.