摘要
传统频域滤波与时域分析方法难以有效处理非平稳干扰的时变特性与多源耦合问题。分析船舶局域网络干扰的时域、频域特征,构建基于小波分析的“信号预处理—特征提取—源定位”一体化框架,提出自适应小波阈值去噪算法,提取能量熵、频率重心、时域突变点密度构成多维度特征向量,实现干扰类型的精准分类。结合小波时频局部化特性与TDOA技术,构建干扰源时空定位模型,仿真试验表明,该方法对高斯噪声、脉冲干扰和混合干扰的信噪比提升分别达15.8、18.3、16.2 dB,验证了本文方法的有效性与工程应用价值。
Traditional frequency-domain filtering and time-domain analysis methods are difficult to effectively handle the time-varying characteristics and multi-source coupling problems of non-stationary interference.Analyze the time-domain and frequency-domain characteristics of the interference in the ship's local area network,construct an integrated framework of"signal preprocessing-feature extraction-source location"based on wavelet analysis,propose an adaptive wavelet threshold denoising algorithm,extract energy entropy,frequency center of gravity,and density of time-domain mutation points to form multi-dimensional feature vectors,and achieve precise classification of interference types.By combining the wavelet time-frequency localization characteristics with TDOA technology,a spatio-temporal location model of the interference source was constructed.Simulation experiments show that the signal-to-noise ratio improvement of this method for Gaussian noise,pulse interference and mixed interference reaches 15.8,18.3,16.2 dB respectively,verifying its effectiveness and engineering application value in the mining of interference information in ship networks.
作者
李慧姝
任娟慧
任波
LI Huishu;REN Juanhui;REN Bo(Department of Computer and Information Engineering,Shanxi Institute of Energy,Jinzhong 030600,China;Taiyuan University of Technology,Taiyuan 030024,China;Hydrology and Water Resources Survey Station of Shanxi Province,Taiyuan 030001,China)
出处
《舰船科学技术》
北大核心
2025年第16期173-176,共4页
Ship Science and Technology
基金
山西省教育科学“十四五”规划课题(GH-230252)。
关键词
小波分析
噪声处理
源定位
信息挖掘
wavelet analysis
noise processing
source localization
information mining
作者简介
李慧姝(1981-),女,硕士,副教授,研究方向为人工智能及信息化。
