Using Blind Signal Separation in the Task of Detecting FSK Signals

The article examines the efficiency of the blind source separation algorithm in the task of FSK signal detection. Classical algorithms of blind source separation and criterias of independence of signals are examined. The model of the second and fourth cumulants blind source separation algorithm is described. To eliminate the trivial result, additional boundaries are used on signal variance. A system of equalities is provided to find the ratios of the stirring matrix. The family of uncorrelated solutions of the unmixing matrix is presented. It is shown how to estimate solution which leads to fourth cumulants of output signals equal to zero. Examined in the paper FSK modulation and demodulation model is presented. The efficiency of using the BSS method in the task of receiving the FSK signal with the white Gaussian noise was shown.