Abstract
Abstract
In the adaptive noise cancellation system, the traditional fixed step size LMS algorithm is used to filter the noise signal, but the noise signal cannot be completely cancelled out. In order to realize the two-stage filtering, this paper proposes to combine two fixed step size LMS algorithms and to combine a fixed step size LMS algorithm with a variable step size LMS algorithm based on Bessel function. The combinations of different LMS algorithms are applied to the adaptive noise cancellation system, and the MATLAB simulation experiment is performed on the noisy signal. Theoretical analysis and simulation results show that the two-stage filtering realized by combination of different LMS algorithms is obviously better than the one-stage filtering realized by a single fixed step LMS algorithm, and the performance in terms of fast convergence speed, tracking speed and steady state error is improved.