Nonlinear active noise control based on neural networks compensation

Most existing active noise control (ANC) methods usually require the precise model of secondary path, thus the performance may be deteriorated if the identified secondary path is inaccurate. In this paper, a novel adaptive control scheme based on the neural networks feedforward compensation is presented for nonlinear active noise control. By using the nonlinear approximation ability of neural networks in the control design, the nonlinear secondary path (NSF) is regarded as a plant to be controlled while the nonlinear primary path (NPF) is considered as an unknown noise model. The proposed scheme does not require the information of the primary and secondary path compared with classical noise control approaches. In addition, the stability of the closed-loop system is proved by Lyapunov theory. The simulation results demonstrate the validity of the proposed method.