NONLINEAR NARROWBAND ACTIVE NOISE CONTROL USING VOLTERRA FILTER

Narrowband active noise control (ANC) has been widely applied to reduce narrowband noise generated by rotating machines like fans, engines, motors, cutting machines etc. So far, many algorithms and structures have been developed for linear narrowband ANC (NANC) systems. However, nonlinear distortions may exist in real-life sound fields. The conventional linear NANC systems will suffer from performance degradation due to the nonlinear distortions. This paper presents a new structure for an NANC system that uses an adaptive Volterra filter to suppress a noisy sinusoid. The reference sinusoidal noise is distorted by a nonlinear part of the primary path, and the conventional linear NANC is awkward in handling the nonlinearity. The filtered-X least mean square (FXLMS) algorithm is adpoted to update both the linear and the nonlinear adaptive filters of the proposed system. Numerical simulations are provided to show that the new nonlinear NANC system is effective in suppressing sinusoidal noise even if the primary path has both linear and nonlinear characteristics.