Online secondary path modeling in active noise control system using PBS-LMS algorithm

In this paper, a Parallel Binary Structured Least Mean Square (PBS-LMS) based adaptive filter is used in the online secondary path modeling of an active noise control (ANC) system. The objective is to improve the performance of the ANC system, at a reasonable computational cost. In the proposed method, the PBS-LMS algorithm is used in the adaptation process of the modeling filter. The s steps look ahead characteristic of PBS-LMS algorithm improves the online secondary path modeling. This modeling filter generates a better anti-noise signal with respect to the primary noise. This leads to a faster convergence of the main ANC adaptive filter; and consequently it improves the noise cancellation performance of the ANC system. Computer simulations show that the proposed ANC system results in higher convergence speed and lower residual noise level, compare to the other LMS based ANC systems.