A filtered-x VSS-NLMS active noise control algorithm robust against impulsive noise and noisy inputs

In this paper, we proposes an active acoustic noise cancellation algorithm which actively cancel out the acoustic noise using a variable step-size normalized least mean square (NLMS) adaptive filter. One of the difficulties of active noise cancellation (ANC) system is that the system has a very long echo tails. So, the FIR filter coefficients are very long. Because the ANC system is also accepts electric signals through a microphone, the ANC system is very sensitive to noise. The Huber function is applied to prevent wrong filter coefficients update by impulsive noise. The Huber function uses the error signal as it is when the error signal is within a certain range. If the error signal is outside a certain range, this function does not update the coefficients. The variable step-size which obtained by mean square deviation (MSD) analysis, speeds up the convergence speed of the algorithm and makes the steady-state error smaller. Simulation results show that the proposed algorithm has much better performance than other algorithm in noisy environment.