Convex combination of functional link artificial neural networks and generalized Fouries series for active noise control system
Authors: Dinh Cong Le
Jounal of science and Tachnology-Da Nang University
: No 10.1.2022. : 82-86
Publishing year: 10/2022
Abstract - In this study, we propose a novel controller for active
noise control (ANC) application based on convex combination of
Functional Link Artificial Neural Networks (FLANN) and Even
Mirror Fourier (EMF) filters, and is named Convex-FLANNEMF (C-FLANN-EMF). The C-FLANN-EMF controller
adaptively combines the outputs of the component filters using
the random gradient algorithm and the nonlinear exponential
function to minimize residual noise. Therefore, it can overcome
the disadvantages of the FLANN and EMF controllers, when the
nonlinearity that exists in the components of the active noise
control system is an unknown priori. Many simulation results
have shown that the noise cancellation performance of the
proposed C-FLANN-EMF controller is more effective than the
FLANN and EMF component controllers in different nonlinear
scenarios of the ANC system.
Active noise control; FLANN; EMF; Convex combination; Nonlinear controller.
