Analyze the Efficiency of Blind Signal Extraction Algorithms in a Background of Impulse Noise Based on the Maximization of the Absolute Value of the Kurtosis.


In this paper, analyzed the efficiency of algorithms of blind pulse signal extraction in a background of impulse noise based on the maximization of the absolute value of the kurtosis. Synthesized blind separation algorithms with fixed point and it is considered in combination with the gradient. The convergence of these algorithms is shown for zero and nonzero initial conditions. A lemma and two theorems are formulated Allowing to prove the blind allocation of the signal and to determine the number of decisions with regard to signal extraction. Modeling established that the fixed point algorithm based on the maximization of the absolute kurtosis value is more efficient and allows to separate the pulse desire signal with the signal-to-noise ratio of 30 dB more than the gradient algorithm with the same objective function. Computer modeling of AbsoKurt and AbsoKurtFP algorithms Carried out in Simulink using Matlab programing.