Adaptive Impulse Noise Suppression in OFDM Systems

Abstract

This paper presents a novel approach to adaptive impulse noise suppression in OFDM system. Our method is based on the analytically modeled threshold whose adaptation, in a contrast to known approaches, does not require computation of the impulse noise parameters in each iteration. The proposed method utilizes a priori knowledge in frequency domain in order to reconstruct samples corrupted by impulse noise. A priori information in frequency domain includes zeros in guard bands, pilot symbols as well as parameters of the data carriers constellation. The proposed algorithm reconstructs signal samples corrupted by impulse noise based on hard decisions in the frequency domain and detects the noise impulses through comparison of the reconstructed signal and the received signal. Decision regarding the impulse sample is based on the threshold whose value is decreasing exponentially from iteration to iteration. The exponent (decreasing rate) is determined based on the reliability of impulse calculated from pdf of random variable calculated from a difference between reconstructed signal and the received signal. The performances of the proposed system are evaluated through simulation which confirms theoretical analysis. Ill. 7, bibl. 11, tabl. 2 (in English; abstracts in English, Russian and Lithuanian).