An Improved Variable-Step FXLMS for Active Noise Control in High-Noise Environment

This work proposes an improved variable-step FXLMS (IVS-FXLMS) adaptive algorithm for active noise control, which considers the effect of secondary path. A lower bound for the step size is obtained to establish a minimum adaptation-level and a range of step size values for breaking the trade-off between misadjustment and convergence time. Computer simulations are performed for a comparative assessment of the performances of the various adaptive algorithms based on the mean-square-error criteria, with respect to the proposed algorithm, in a high-noise environment. Comparison is done in terms of noise attenuation, reproduction of the original signal and mean square error. It is observed from the obtained results that the FXLMS algorithm, due to its ability to take secondary path into consideration, performs better than the LMS and NLMS adaptive algorithms. However, the proposed algorithm exhibits an improved performance over the FXLMS algorithm by effectively attenuating noise at the output and reproducing the original acoustic signal, which was corrupted with noise. Moreover, the proposed algorithm also incurs the least mean square error of all the considered algorithms.