Improving ANC Computational Efficiency: A Novel Fx LMS Algorithm with Dynamic Response Control

This paper presents a novel variant of the Fx LMS algorithm. This variant optimizes computational efficiency by utilizing a calculated imperfect secondary path model with a pure-delay impulse response, simplifying the filtered reference signal calculation and enhancing the convergence rate. Additionally, the algorithm includes a control parameter to adjust its dynamic response, further improving its convergence behaviour. We provide a theoretical framework based on root locus theory to optimize this parameter and the algorithm's performance. The theoretical contributions are substantiated through computer simulations, validating the enhanced effectiveness of the proposed approach in ANC applications.