A Novel Cancellation Path Modeling Signal for Active Noise Control

A new on-line cancellation path modeling signal for active noise control is proposed in this paper. The signal is based on the spectrum of the disturbance signal, and is especially useful for the cancellation path modeling in active noise control when the disturbance signal is tonal. It is shown that using the proposed signal as the perturbation signal in modeling can significantly reduce the modeling error than that of using the random signal. This is because the modeling error is mainly caused by the spectrum of the modeling signal close to the frequency of disturbance signal, and the spectrum of the proposed signal around the frequency of the disturbance signal is much less than that of random signal. Some experimental results are also given to illustrate the above discussions. In the experiments, the same level of the proposed signal and the random signal are used separately as modeling signal with the same convergence coefficient, and it is found that the maximum phase difference between the true canceling path transfer function and the obtained canceling path transfer function at the disturbance frequency is 180 degrees with random signal, and reduced to only 15 degrees with the proposed signal. The new proposed modeling signal can greatly increase the performance of an active noise control system.