基于自适应小波神经网络实现鱼雷超声引信信号消噪  

Noise Reduction of Torpedo Ultrasonic Fuze Signal Processing Based on Self-Adaptive Wavelet Neural Network

作　　者：尹熙鹏[1] 樊养余[1] 段哲民[1] 程伟[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710072

出　　处：《鱼雷技术》2009年第5期19-22,共4页Torpedo Technology

基　　金：国家自然科学基金资助项目(60874101)

摘　　要：将小波理论和人工神经网络技术同时引入鱼雷超声引信目标信号处理中,抑制信号噪声,并给出了自适应小波神经网络的构造方法。采用正交的Daubechies小波系作为小波元,并对尺度参数进行优化选取,以实现更好的拟合信号。仿真结果表明,该自适应小波神经网络对引信目标信号的处理具有消噪效果好,失真小的特点,有利于防止出现误判和漏判目标的现象。To remove the noise signal effectively in processing torpedo ultrasonic fuze signal, wavelet theory and artificial neural network are introduced simultaneously, and the construction method of self-adaptive wavelet neural network is given. Better fitting signal is achieved by adopting orthogonal Daubechies wavelet group and optimizing the scale parameter. Simulation result shows that this wavelet neural network can effectively suppress noise from ultrasonic fuze signal with less distortion, leading to reduction of target misjudgment and misdetection of torpedo.

关 键 词：鱼雷 超声引信 自适应 小波分析 神经网络 

分 类 号：TJ431.7[兵器科学与技术—火炮、自动武器与弹药工程]