Abstract:
The performance of conventional active noise control (ANC) systems degrades under non-Gaussian conditions. In the recent past, generalized filtered Maximum correntropy criterion (FxGMCC) algorithms has been widely used to tackle such non-Gaussian noises. However, noise reduction performance of MCC based algorithms degrades due to the high steady-state misalignment. To overcome this limitation, a filtered-x exponential hyperbolic cosine adaptive filter (FxEHCAF) algorithm proposed in this paper. Later, an affine combination technique has been applied and the corresponding affine FxEHCAF (A-FxEHCAF) algorithm has been developed here showing faster convergence rate in transient state along with low residual error in steady-state. Simulation study shows that the proposed FxEHCAF and A-FxEHCAF algorithms provide enhanced noise control performance over other existing algorithms.