Adaptive ANFIS-based filter for active control of sinusoidal primary noise in nonlinear path

Linear active noise control (ANC) system is promising in suppressing broadband and narrowband noise in a linear acoustic site. However, in some situations, the acoustic systems are nonlinear which make the primary noise being nonlinear distortion at the canceling point, and the linear ANC system maybe can not work well for these cases. This paper presents an adaptive network-based fuzzy inference system (ANFIS)-based nonlinear filter for active noise control, which is expected to be successfully applied to situations with nonlinear distortion. In the proposed adaptive control algorithm, the parameters are updated by gradient descent method. In the simulations, the secondary-path FIR model coefficients are estimated in advance by off-line system identification method, and one case for sinusoidal primary noise with nonlinear distortion is presented, which verifies the effectiveness of the proposed system in suppressing sinusoidal noise generated by rotating machines, e. g., engines in vehicles.