针对传统非高斯概率分布模型在海洋混响建模中稳健性不足的问题，首先采用高斯-污染混合分布模型对混响建模，该分布模型是高斯分布和未知的污染分布的线性组合。然后基于最大熵原则，在均值约束条件下确定了最大熵污染分布为拉普拉斯分布，并应用期望最大化（Expectation Maximization，EM）算法估计高斯-拉普拉斯混合分布模型的概率密度参数。最后通过真实海洋混响建模，验证了高斯-拉普拉斯混合分布模型的稳健性。
Aiming at the problem that the traditional non Gaussian probability distribution model is not robust enough in marine reverberation modeling, firstly, the Gaussian pollution mixed distribution model is used to model the reverberation, which is a linear combination of Gaussian distribution and unknown pollution distribution. Then based on the maximum entropy principle, the maximum entropy pollution distribution is determined as a Laplacian distribution under the mean constraint condition, and the expectation maximization (EM) algorithm is used to estimate the probability density parameters of the Gaussian Laplacian mixture distribution model. Finally, the robustness of the Gaussian Laplacian mixed distribution model is verified by modeling the real ocean reverberation.
2025,47(23): 141-146 收稿日期：2025-3-17
DOI：10.3404/j.issn.1672-7649.2025.23.022
分类号：U674.7；TJ630
基金项目：国家自然科学基金资助项目（11974090）
作者简介：代振(1991-),男,博士,讲师,研究方向为水声信号处理
参考文献：
[1] KAY S, SALISBURY J. Improved active sonar detection using autoregressive prewhiteners[J]. The Journal of the Acoustical Society of America, 1990, 87(4): 1603–1611.
[2] 赵航芳, 祝献, 宫先仪. 混响背景下的信号检测[J]. 哈尔滨工程大学学报, 2004, 25(1): 34-37.
ZHAO H F, ZHU X, GONG X Y. Detection of signals in reverberation noise[J]. Journal of Harbin Engineering University, 2004, 25(1): 34-37.
[3] ABRAHAM D A, GELB J M, Oldag A W. Background and clutter mixture distributions for active sonar statistics[J]. IEEE of Oceanic Engineering, 2011, 36(2): 231–247.
[4] 许彦伟, 何鑫彪, 刘明刚, 等. 干扰背景非瑞利混响参数估计[J]. 声学学报, 2023, 48(4): 733-742.
XU Y W, He X B, LIU M G, et al. Parameter estimation for non-Rayleigh reverberation under background interference[J]. Acta Acustica, 2023, 48(4): 733-742.
[5] 卫红凯, 王平波, 蔡志明, 等. 混响的混合高斯概率密度建模[J]. 声学技术, 2007, 26(3): 514-518.
WEI H K, WANG P B, CAI Z M, et al. Gaussian mixture model for reverberation[J]. Technical Acoustics, 2007, 26(3): 514-518.
[6] 王平波, 卫红凯, 娄良轲, 等. 海洋混响数据的SαS分布建模[J]. 哈尔滨工程大学学报, 2021, 42(1): 55-60.
WANG P B, WEI H K, LOU L K, et al. Oceanic reverberation probability density modeling based on symmetric alpha-stable distribution[J]. Journal of Harbin Engineering University, 2021, 42(1): 55-60.
[7] 李宗吉, 张西勇. 基于 α 稳定分布模型的水声信号白化滤波方法[J]. 舰船科学技术, 2014, 36(8): 53-56.
LI Z J, ZHANG X Y. Whitening filtering method for underwater acoustic signal based on α-stable distribution model[J]. Ship science and technology, 2014, 36(8): 53-56.
[8] 孙疆僵. 基于统计分析的混响抑制方法研究[D]. 哈尔滨: 哈尔滨工程大学, 2022.
[9] LI X T, SUN J, WANG SH Y, et al. Near-optimal detection with constant false alarm ratio in varying impulsive interference[J]. IET Signal Processing, 2013, 7(9): 824-832.
[10] 周访滨, 朱建军, 陈永奇, 等. 污染正态分布的熵估算[J]. 中南大学学报(自然科学版), 2013, 44(3): 1269-1274.
ZHOU F B, ZHU J J, CHEN Y Q, et al. Entropy estimation of contaminated normal distribution[J]. Journal of Central South University (Science and Technology), 2013, 44(3): 1269-1274.
[11] 孙永厚, 周洪彪, 黄美发. 基于最大熵的测量不确定度的贝叶斯评估方法[J]. 统计与决策, 2008, 32(12): 143-145.
SUN Y H, ZHOU H B, HUANG M F. Bayesian evalua–tion method of measurement uncertainty based on maxi–mum entropy[J]. Statistics and decision making, 2008, 32(12): 143-145.
[12] 任博, 岳珠峰, 崔利杰, 等. 混合数据信息下不确定性描述的改进最大熵函数法[J]. 西北工业大学学报, 2022, 40(2): 450-457.
REN B, YUE Z F, CUI L J, et al. An improved entropy-based representation for mixed uncertaintry about intervals and point data[J]. Journal of Northwestern Polytechnical University, 2022, 40(2): 450-457.
[13] YANG M S, LAI C Y, LIN C Y. A robust EM clustering algorithm for Gaussian mixture models[J]. Patter Recognition, 2012, 45(11): 3950-3961.
[14] 梁志贞, 张磊. 面向 Kullback-Leibler 散度不确定集的正则化线性判别分析[J]. 自动化学报, 2022, 48(4): 1033-1047.
LIANG Z Z, ZHANG L. Regularized linear discriminant analysis based on uncertainty sets from Kullback-Leibler divergence[J]. Acta Automatica Sinica, 2022, 48(4): 1033-1047.
