Active noise control (ANC) in the presence of impulsive/non-Gaussian disturbances remains challenging for the conventional adaptive algorithms. In recent years, information theory-based maximum correntropy criterion (MCC) and its variants have gained significant attention in environments with non-Gaussian disturbances. However, the MCC-based algorithm’s performance suffers from high steady-state misalignment, which limits the noise control performance. In order to overcome this limitation, this paper proposes a filtered-x logistic distance metric adaptive filter (FxLDMAF) for robust ANC. It optimizes a logistic distance–based objective whose smooth saturation curtails the influence of significant instantaneous errors while maintaining sensitivity around the origin, thus promoting stable coefficient updates. Performance is examined under a symmetric α -stable interface and additional non-Gaussian conditions, including Laplacian, Uniform, and Binary noise. Across all scenarios, simulations show that FxLDMAF achieves lower steady-state misalignment and superior noise control performance compared to benchmark approaches (RFxLMS, FxMCC, and FxGMCC). The results indicate that FxLDMAF is well-suited for real-time ANC deployments operating in unpredictable, heavy-tailed acoustic environments.
Patnala et al. (Sun,) studied this question.