Robust FIR System Identification for Super-Gaussian Noise Based on Hyperbolic Secant Distribution

Abstract

In this paper, we present a finite impulse response (FIR) system identification algorithm based on the hyperbolic secant (sech) distribution. In our algorithm, we assume that the noise signal follows a super-Gaussian distribution. Then, to estimate the FIR coefficients, we derive the optimization algorithm based on the majorization-minimization (MM) algorithm. In our algorithm, the scale parameter is also estimated as well as the FIR coefficients. Our simulations demonstrate the proposed method is effective in the FIR system identification in super-Gaussian noise environments.