A Robust Affine Projection Algorithm Against Impulsive Noise
Jae Jin Jeong
Abstract
This letter proposes a prefiltered observation-based affine projection algorithm (APA) to achieve robustness against outliers. The conventional robust algorithm for correlated input signal, which is called the affine projection sign algorithm (APSA), was developed by using the L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> -norm of the error signal. However, it suffers from slow tracking speed for a system change suddenly, because it can not distinguish between the time-varying system and occurrence of the impulsive noise. To overcome this problem, the proposed algorithm is induced from the matrix inequalities in no impulsive interference. The proposed algorithm determines the necessity for updating the weight vector in the noise. Hence, it has robustness and tracking ability because it discriminates between the time-varying system and occurrence of the impulsive noise. Simulations in a system identification scenario show that the proposed algorithm surpasses the APA and APSA in terms of convergence rate and tracking performance under the impulsive noise environment.