Adaptive Impulsive Noise Control Algorithms based on The Fractional Lower-order Statistics

Noise control of signals is a key challenge problem in signal enhancement, signal recognition, communication, radar, and sonar applications. The most widely used method is adaptive linear filtering method, which can adaptive change filter parameters with the stochastic property of the stationary Gaussian noise. The representative algorithms of this include least mean square (LMS) adaptive filter and recursive least squares (RLS) adaptive filter. The conventional adaptive filtering algorithms suffer from severe degradation in impulsive noise environments. Although the fractional lower-order statistics (FLOS) based adaptive methods are robust to the impulsive noise, the complexity of these algorithms is a main problem in real world implementation. In this paper, two adaptive algorithms for impulsive noise reduction are proposed. The proposed methods have the capability of steady-state in the presence of impulsive noise. Simulation results illustrate an improvement in terms of convergence and steady-state performance.