A Stabilized Hybrid Active Noise Control Algorithm of GFANC and FxNLMS with Online Clustering

TL;DR

The paper tackles the challenge of achieving both fast response and low steady-state error in active noise control by merging Generative Fixed-filter ANC ($GFANC$) with Filtered-x Normalized Least Mean Square ($FxNLMS$) in a stabilized, dual-rate framework. A frame-rate CNN predicts weight vectors for $M$ sub-control filters derived from a single pre-trained broadband filter, while a sampling-rate $FxNLMS$ continuously refines the generated filter; an online clustering module governs weight updates to prevent destabilizing reinitializations. The approach decomposes the broadband filter into $M$ sub-filters, uses a frame-rate weight predictor $ extbf{g}'$, and employs dynamic clustering to decide when to switch to a new weight vector, thereby improving stability without sacrificing adaptability. Results show that the proposed method delivers fast noise reduction with very low steady-state error and enhanced stability, outperforming $GFANC$, $FxNLMS$, $SFANC$, and $SFANC$–$FxNLMS$, and requiring only a single pre-trained broadband filter. This work offers a practical, robust solution for real-time ANC in vehicle and cabin environments and reduces training requirements for broadband filters.

Abstract

The Filtered-x Normalized Least Mean Square (FxNLMS) algorithm suffers from slow convergence and a risk of divergence, although it can achieve low steady-state errors after sufficient adaptation. In contrast, the Generative Fixed-Filter Active Noise Control (GFANC) method offers fast response speed, but its lack of adaptability may lead to large steady-state errors. This paper proposes a hybrid GFANC-FxNLMS algorithm to leverage the complementary advantages of both approaches. In the hybrid GFANC-FxNLMS algorithm, GFANC provides a frame-level control filter as an initialization for FxNLMS, while FxNLMS performs continuous adaptation at the sampling rate. Small variations in the GFANC-generated filter may repeatedly reinitialize FxNLMS, interrupting its adaptation process and destabilizing the system. An online clustering module is introduced to avoid unnecessary re-initializations and improve system stability. Simulation results show that the proposed algorithm achieves fast response, very low steady-state error, and high stability, requiring only one pre-trained broadband filter.

Figures (5)

• Figure 1: Block diagram of the GFANC–FxNLMS algorithm: a CNN and online clustering provide the weight vector to generate the control filter at the frame rate, while FxNLMS continuously optimizes the generated filter at the sampling rate.
• Figure 2: Online clustering decides if the CNN-predicted weight vector updates the current one, preventing unnecessary filter re-initializations in FxNLMS and enhancing stability.
• Figure 3: Frequency spectra of a pre-trained broadband control filter and the $M$ sub control filters decomposed from it.
• Figure 4: Error signal comparison of the GFANC–FxNLMS algorithm without and with online clustering.
• Figure 5: Noise reduction performance of different ANC algorithms on (a)-(d): the vehicle noise, (e)-(h): the $100$-$1200$ Hz noise.