Multilayer perceptron neural networks for active noise cancellation

Some experiment results of multilayer perceptron neural networks for active noise cancellation (ANC) are presented in this paper. Active noise cancellation is an approach to noise reduction in which a secondary noise source destructively interferes with the unwanted noise. Conventional ANC systems apply a digital filter to cancel the noise and suffer from the fact that each noisy environment needs to be treated individually. In this paper, we introduce a three-layer MLP neural network to replace the aforementioned filter. We apply this architecture to three different broad-band-noise environments and achieve 20 dB noise attenuation