An Enhanced Impulse Noise Control Algorithm Using a Novel Nonlinear Function

In many scenarios where impulse noise occurs, an active noise control (ANC) system using a filtered-x least mean square (FxLMS) algorithm will have undesirable effects. To solve this problem, a novel adaptive algorithm is proposed that takes the Gaussian error function as the cost function to nonlinearly transform the residual error. To attenuate the different levels of impulse noise, the parameter k is used as a factor that transforms the error signal by different degrees. To further improve the efficiency of the algorithm, we propose a variable step size strategy based on normalized variable step size, combined with the sine function. Not only the random impulse noise signals generated by the Chambers-Mallows-Stuck method are adopted for simulation, but also the noise signals collected from real vehicles are used in this work. Finally, the simulations are carried out and the results show that the proposed algorithm offers perfect performance in dealing with different levels of impulse noise.