基于神经网络的自适应噪声相消

The Adaptive Noise Cancelling Based on Neural Network

要用常规自适应滤波器实现噪声相消，处理是很繁杂，也是很困难的。而神经网络具有高度的并行运算能力和非线性变换特性，特别是多层前向ＢＰ网可以实现复杂的非线性映射，以及很强的自适应和容错能力，为解决以上问题展示了十分诱人的前景。本文采用ＢＰ网络实现自适应噪声相消，实际应用效果是良好的。

It is difficult for conventional adaptive filter to cancel noise. Because neural network has the features of a high degree of concurrenent calculation and nonlinear mapping, especially, the abilities of adapting and fault compatibility, it makes adaptive noise cancelling possible. We successfully apply BP network in noise cancelling. The effect of practical application is satisfactory.