A nonlinear noise active control algorithm based on deep neural network

To solve the problem that the traditional filter-x least mean square (FxLMS) algorithm and the correlation variable step FxLMS algorithms cannot handle nonlinear interference, a nonlinear noise active control algorithm based on deep neural network (DNN) is proposed. First, nonlinear interference between the acoustic channel and the secondary loudspeaker is introduced into the active noise control system. Second, the training objectives and loss functions of the deep neural network are set, and the network parameters are trained. Finally, a simulation of the active noise control in the local workspace of a tufted carpet loom is carried out to verify the performance of the DNN algorithm. The results show that the DNN algorithm is effective in dealing with nonlinear disturbances.