A new variable step size method for online feedback path modeling in active noise control systems

In this paper, a new variable step size method for online acoustic feedback path modeling (and neutralization) in single channel active noise control (ANC) system is proposed. The proposed method uses three filters. (1) A least mean square based disturbance estimation filter (2) A variable step size LMS based feedback path modeling filter and (3) An FxLMS based noise control filter. In the proposed method, the convergence behavior of the feedback path modeling filter is improved by using a variable step size LMS (VSS-LMS) algorithm. In addition to this, the disturbance estimation (prediction) filter further speeds up the convergence of the feedback path modeling filter. Computer simulations show that for narrow band input signal, the feedback path modeling filter, the noise control filter and the residual error converge faster than the existing methods.