Active Noise Control of 3D enclosure System using Co-FXLMS Algorithm

摘要

The LMS(Least-Mean-Square) type algorithm is the most widely used algorithm for active noise control. The LMS algorithm can easily obtain the complex transfer function in real-time, so modified LMS algorithms that can improve performance have been developed. Especially, the Filtered-X LMS (FXLMS) algorithm has been applied to Active Noise Control (ANC) and Active Vibration Control (AVC). However, the step size of FXLMS algorithm is fixed. So, small step size ensures the stability, but reduces the convergence speed and big step size enhances the convergence speed, but reduces the stability. Therefore, it is necessary to develop the LMS-type algorithm that contains variable step size. Thus Correlation FXLMS (Co-FXLMS) algorithm was proposed to improve the performance of ANC. The Co-FXLMS algorithm is realized by using the estimation of the cross correlation between the error signal and the filtered input signal. When the correlation is high, the gain is also high, and the Co-FXLMS algorithm is in an “active” state. When the error signal and the filtered input signal are uncorrelated, the gain is close to zero, and the Co-FXLMS algorithm is put in an “asleep” state. In this paper, Co-FXLMS algorithm is applied to the experiment on the ANC of the 3D enclosure system.