四阶谱在水中目标识别中的应用

The fourth-order spectrum application on underwater targets recognition

由于水声信号通常是非平稳、非高斯的,因此对非高斯特征的提取就显得特别重要。而四阶谱具有抑制高斯噪声的能力,本文用标准的一维滞后窗构造三维滞后窗,并将其应用于四阶谱的非参数估计中。仿真实验证明其能明显地抑制四阶谱估计的方差。另外,提取了水声信号的四阶谱特征,识别实验表明其能提高识别率。推导的离散序列非参数四阶谱递推算法使四阶谱的运算速度明显提高。

The extracting of non-Gauss characteristics is very important for the hydro-acoustics signal being non-stabilization and non-Gauss property normally,so the fourth-order spectrum can be used to deal with the hydro-acoustics signal as it having the capability of repressing the Gauss noise.The three-dimensional lag window is formed through the standard one-dimensional lag window in this paper.And it is applied to the non-parametric estimation of fourth-order spectrum,simulation results show that it restrains the variance of estimation distinctly.Furthermore,the characteristics of fourth-order spectrum was extracted,the identifying experiment indicates that this method increased the ratio of identification.The discrete sequence of non-parametric fourth-order spectrum recursive algorithm is induced,and the computing speed is improved evidently.