文章摘要
邓兵.一种非线性动力学降噪法应用于海底探雷抗混响的探讨[J].声学技术,2001,(3):110~113
一种非线性动力学降噪法应用于海底探雷抗混响的探讨
A study on nonlinear dynamics noise-reduction method to detect mines lying on the seafloor and anti-reverberation
投稿时间:2000-07-04  修订日期:2001-04-27
DOI:
中文关键词: 动力学模型  FIR滤波器  神经网络  径向基
英文关键词: dynamical model  FIR filter  neural network  Radial Basis
基金项目:
作者单位
邓兵 海军航空工程学院电子工程系, 山东烟台264001 
摘要点击次数: 1140
全文下载次数: 769
中文摘要:
      文章所述降噪法建立在背景信号能够用低维动力学系统建模的基础之上,利用待分离信号与干扰背景的频谱差异,通过连续拟合的两层RBF网来构造非线性逆滤波器以达到信号与背景分离的目的。本文给出了分别以仿真混沌信号和实测混响信号为背景与单频脉冲信号相分离的示例,证明了该算法在混响背景下实现信号分离的可行性(分离前后信混比提高了约16dB),分析了算法有效使用所必需的频谱差异条件(不小于0.15个圆周频率),为抑制混响提供了一条新的思路。
英文摘要:
      A nonlinear dynamic noise reduction method is described in this paper,which achieves the separation between the useful signal and the noise by the construction of nonlinear inverse to linear filter in virtue of successive approximation using Radial Basis Function,based on the spectrum difference between the aforementioned signals.Two examples with sine pulse as the useful signal are show in the paper. One is the artificial chaotic signal, and another is the reverberation. These examples confirm the feasibility about this method with reverberation as the background(the improvement on SRR reaches 16dB before and after the usage of this method).The analysis of the condition of spectrum difference has been done to ensure this method's validity(≥0.15 cycle frequency).In this paper a new approach about anti reverberation is provided.
查看全文   查看/发表评论  下载PDF阅读器
关闭