Artificial neural net based noise cancellor

This paper presents a new method for noise cancellation with an Artificial Neural Network. The network used is a feedforward one with three layers. The backpropagation and Stastical Cauchy's learning algorithms are employed for adaptation of the internal parameters of the network. The constrained tangent hyperbolic function is used to activate the neurons and to provide the desired non-linearity. Promising simulation results for noise cancellation intensify the validity of superseding the proposed scheme for many existing techniques. To demonstrate the effectiveness, the proposed method is applied to different, input conditions with varying SNRs. With incomplete signal samples the net is found to produce output having a striking resemblance with that of the desired ones. A performance comparision of the two algorithms is presented in the paper for better appraisal.