基于神经元网络的自适应噪声抵消系统研究与实现

Research on Adaptive Noise Cancellation System Based on Neural Network

自适应噪声抵消技术是自适应滤波器最普遍的应用之一,它是一种在未知信号和噪声的先验知识条件下,能够很好地消除背景噪声影响的信号处理技术,具有很高的应用价值;但是,在很多情况下,噪声环境非常复杂,往往是非线性的,而目前所使用的自适应滤波器均属线性滤波器,滤波后会使原始信号产生失真;由于神经网络具有非线性等优点,可以很好的逼近非线性函数,所以提出了基于神经网络的自适应噪声抵消器;仿真结果表明,该方法可以效地实现噪声的抵消;最后提出应用DSP实现语音信号自适应噪声抵消的具体方案。

Adaptive noise cancellation technique is one of the most common application of adaptive filter. It is a very well signal processing technique which can eliminate the effects of background noise when priori knowledge of the noise and signal is unknown. So it has very high application value. In many cases, noise environment is very complex and often nonlinear though most adaptive filters used in the present are all linear. The original signal will be distorted after filtering by these filters. Neural network can approximate to nonlinear functions very well as its＇ nonlinear advantage, etc. So, in this paper, a neural network is applied in a adaptive noise canceller. Noise can be cancelled effectively, which is proved by simulation. Finally, a concrete scheme for audio signal to realize the adaptive noise canceling system by using DSP is brought forth.