Improving active noise control without secondary path modeling using subband phase estimation.

The problem of the secondary path variation on feedforward active noise control (ANC) systems can be solved by applying a phase shift on the reference signal. The existing algorithms choose the phase shift from several fixed candidates by monitoring the residual noise power, which has the problem of deteriorated convergence speed, especially when the difference between the chosen phase shift and the true phase value is large. In this paper, an improved ANC algorithm is proposed by having a better estimation of the phase shift in each subband with the least squares method. Simulation results validate the faster convergence speed of the proposed algorithm, and its computational load is discussed.