An Intermittent FxLMS Algorithm for Active Noise Control Systems With Saturation Nonlinearity

Abstract

In practical active noise control systems, saturation nonlinearity is one of the most common nonlinear effects, which can occur at the electronic circuits in the secondary path. Adaptive algorithms designed for linear systems, such as the filtered-x least mean square (FxLMS) algorithm, may diverge in nonlinear systems with a saturated secondary path and result in limited noise reduction performance. Various nonlinear algorithms with improved stability have been proposed; however, a trade-off between noise reduction performance and computational complexity exists when selecting a satisfactory algorithm. This paper therefore proposes a simple algorithm that can achieve significant noise reduction with low computational complexity. The proposed algorithm adopts the same update equation as the FxLMS algorithm, but uses an intermittent update strategy to avoid unnecessary fluctuations of the weight vector when the output of the control filter lies in the saturation region. Simulation results under different conditions demonstrate the effectiveness and advantages of the proposed algorithm.