Nonlinear Active Noise Control using NARX model structure selection

In this work a novel nonlinear active noise control (NANC) scheme is proposed to deal with reference and error microphone saturation issues. Polynomial NARX models are used in the adaptive controller for enhanced model flexibility. A suitable model selection approach is used off-line to find an accurate and compact model structure for the adaptive controller. The controller parameters are successively updated with an appropriate adaptive algorithm based on the error gradient and on the residual noise. Some simulation experiments are provided to show the effectiveness of the proposed approach compared to existing alternatives.