神经网络自适应有源噪声对消器

Neural Networks Adaptive Active Noise Cancellor

本文提出一种新颖的非线性自适应滤波器－递归神经网络（ＲＮＮ）自适应滤波器．讨论ＲＮＮ滤波器的构成及算法，实现了ＲＮＮ自适应有源噪声对消器．ＲＮＮ自适应有源噪声对消器性能优于内线性滤波器组成的线性对消器，能适应各种噪声环境．仿真结果充分证明了这一点．

A novel nonlinear adaptive filter- recurrent neural networks (RNN) adaptive filter is provided in the paper. Its structure and algorithm are investigated, and RNN adaptive active noise cancellor is realized~. RNN adaptive active noise cancellor has better anti-noise performances than linear cancellor which is made up of linear filter and can adapt to different kinds of noise environments Simulative has fufficiently verified this.