Design and Validation of a Single-Channel Active Noise Control System with Multi-Loudspeaker Control Based on Ear Localization

To address the degradation in noise reduction stability caused by changes in ear position, this paper proposes a three-loudspeaker adaptive active noise control (ANC) method based on ear recognition perception. A sliding window approach is employed to accelerate ear position detection, and a single-channel input structure is used to design three sets of secondary path models. The system selects the optimal loudspeaker in real time based on both average sound pressure level error and detected ear position. Additionally, the classical FxLMS algorithm is enhanced by integrating the Whale Optimization Algorithm (WOA) for global step-size optimization, improving the convergence speed and stability of the adaptive algorithm. Simulation and experimental results demonstrate that the proposed WOA-FxLMS method outperforms the conventional FxLMS algorithm in objective evaluation metrics. Compared with other improved FxLMS variants, it offers strong global search capability, fewer parameter requirements, and faster convergence. These results confirm the effectiveness of the ear-aware three-loudspeaker layout.