基于神经网络解卷积的脉冲响应函数提取

来源期刊:中南大学学报(自然科学版)2011年第z1期

论文作者:周凯 王晓东 叶庆卫 周宇

文章页码:67 - 71

关键词:脉冲响应函数;广义逆;相关系数法;神经网络

Key words:impulse response function; generalized inverse; correlation coefficient; neural network

摘    要:在分析经典的脉冲响应信号时域法(相关系数法)的基础上,提出采用径向基网络进行脉冲响应信号提取,以增强噪声抵抗能力,使得在较强噪声情况下还能提取出较理想的脉冲响应信号。在径向基网络样本设计中,提出了对输入信号样本矩阵采用矩阵LU变换来提高网络的训练性能,LU分解也增强了网络的泛化能力。试验结果表明:基于径向基网络的脉冲响应信号提取在较强噪声环境下的表现优越于经典的时域法,并且算法时间也较快,能够适用于实际工程应用。

Abstract: A method was proposed to extract the impulse response signal based on radial basis network in order to enhance capability of noise resistance and extract the desired one. In the sample design of RBF network, signal sample matrix was inputted to matrix LU to enhance the network training performance, and LU decomposition to improve the generalization ability. Simulation results show this proposed approach is superior to the classical time-domain method in very noisy environment and time consumption, which is capable of practical engineering application.

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号