Online acoustic feedback mitigation with improved noise-reduction performance in active noise control systems

The main focus of this paper is the acoustic feedback path neutralisation during online operation of the single-channel active noise control (ANC) systems. Invasive techniques, in which additive auxiliary noise is injected for online feedback path modelling (FBPM), degrade the noise-reduction performance of ANC systems. In the existing methods for online FBPM, additive auxiliary noise with fixed variance is injected during all operating conditions of ANC systems. In this paper, a scheduling strategy is proposed to have time-varying gain for additive auxiliary noise. The time-varying gain is computed on the basis of (i) the convergence status of the FBPM filter, and (ii) the power of the interference term in the error signal of the FBPM filter. Simulation results show that, compared to the existing methods, the proposed method can better achieve the conflicting requirements of fast convergence of the FBPM filter, and reduced steady-state power of the residual error at the error microphone. For sake of completeness, appendix presents typical simulation results when both the secondary path and feedback path are simultaneously identified during online operation of ANC systems.