Abstract
Active noise control (ANC) represents an efficient technology for enhancing the noise suppression performance and ensuring the stable operation of multi-sensor systems through generative model-enhanced data representation and dynamic information fusion across heterogeneous sensors due to the complexity of the real-world environment. To address problems caused by a nonlinear noise source, a novel adaptive neuro-fuzzy network controller is proposed for feedforward nonlinear ANC systems based on a variable step-size filtered-x least-mean-square (VSS-LMS) algorithm. Specifically, the LMS algorithm is first introduced to update the weight parameters of the controller based on the adaptive neuro-fuzzy network. Then, a variable step-size adjustment strategy is proposed to calculate the learning gain used in the LMS algorithm, which aims to improve the nonlinear noise suppression performance. Additionally, the stability of the proposed method is proven by the discrete Lyapunov theorem. Extensive simulation experiments show that the proposed method surpasses the mainstream ANC methods with regard to nonlinear noise.