摘要
分析了一般情况下的信号与噪声分离的理论和方法问题 ,重点讨论了利用BP神经网络从强噪声背景下的提取弱信号的方法 ,给出了具体的BP训练算法和提取弱信号的公式。
The general issue of separating signal from noise background is first analyzed,then the stress is put upon the extraction of weak signal from strong noise background by BP neural networks.There are two main steps:fist train the networks to be stabilized by using pure noise and noise mixed with weak signal respectively,then extracting weak signal by making use of different weights and different outputs in two trained networks.The concrete BP training algorithm and the formula of extracting weak signal are presented.By assuming the activation function of neurons to be linear,simulation is carried out.The results show the presented method is practical and of good effect.The ergodic noise and the signal of low frequency are required.
出处
《济南大学学报（自然科学版）》
CAS
2001年第3期195-197,共3页
Journal of University of Jinan(Science and Technology)
