A multi-strategy enhanced algorithm for active impulsive noise control

Existing active impulsive noise control (AINC) algorithms often face a trade-off between convergence and steady-state error, and their hyperparameters are usually set empirically, which limits practical applications. To address these issues, this paper proposes a multi-strategy enhanced algorithm. First, the fractional lower-order stochastic gradient descent filtered-x least hyperbolic tangent (FoFxLHT) algorithm is introduced to improve robustness against impulsive components. Then, a competitive combination structure (C-FoFxLHT) is designed to effectively coordinate convergence and steady-state performance. Finally, an online tuna swarm optimization (TSO) module is integrated to achieve adaptive tuning of the fractional-order parameter, forming the complete TSO-C-FoFxLHT algorithm with enhanced self-adaptability. Simulation and experimental results demonstrate that the proposed algorithm outperforms several recent and conventional AINC methods in terms of convergence, average noise reduction, and engineering practicality.