A Filtered-X Combined Step-Size Affine Projection Tanh Algorithm for Active Noise Control

This study introduces the combined step-size affine projection tanh algorithm (CSS-APTA), which addresses the trade-off problem between convergence speed and steady-state misalignment inherent in fixed step-size adaptive filtering algorithms. Furthermore, the research presents a novel approach to active noise control by incorporating the combined step-size filtered-x affine projection tanh algorithm (CSS-FxAPTA). The primary focus of the proposed algorithm is on enhancing robustness in scenarios involving correlated input signals and impulsive noise. Simulation results validate that the proposed algorithm achieves a suitable convergence rate while significantly reducing steady-state misalignment compared to existing methods, thereby delivering superior overall performance.