A variable momentum factor filtered-x weighted accumulated LMS algorithm for narrowband active noise control systems

In this paper, a filtered-x weighted accumulated least mean square (FXWALMS) algorithm is proposed for a typical narrowband active noise control (NANC) system by introducing the momentum LMS (MLMS) algorithm into the conventional filtered-x LMS (FXLMS) algorithm. The algorithm uses the new cost function to derive the updating equation for the discrete Fourier coefficients (DFCs) of the secondary source. In this way, the proposed algorithm achieves fast convergence and tracking capabilities at the expense of degrading the steady-state performance. To remedy the problem of the poor steady-state performance for the FXWALMS algorithm, a scheme of using a variable momentum factor is proposed to improve the overall performance of the system. In addition, compared to the filtered-x recursive least squares (FXRLS) algorithm, the performance of the proposed algorithm is better in terms of the computational complexity and tracking capabilities. Computer simulations were conducted to demonstrate the superior performance of the proposed algorithm in both stationary and non-stationary scenarios.