U.E. SATTAROVA
TECHNOLOGY AND SOFTWARE FOR DETERMINING CORRECT NORMALIZATION OF CORRELATION FUNCTIONS
With the traditional approach, it is assumed that the transition to normalized estimates of the correlation functions of noisy signals makes it possible to simplify calculations, move to a dimensionless value, and thus compare processes with different values of variances. It is demonstrated in this paper that the obtained results still contain significant errors from the noise even when the classical conditions are satisfied. The technology for corrected calculation of normalized correlation functions of noisy signals is proposed in the paper, which makes it possible to significantly reduce the errors from the noise even if the classical conditions are not satisfied.
Keywords: random signal, noise, noisy signal, normalized estimates of the correlation function, robust estimates
TECHNOLOGY AND SOFTWARE FOR DETERMINING CORRECT NORMALIZATION OF CORRELATION FUNCTIONS
With the traditional approach, it is assumed that the transition to normalized estimates of the correlation functions of noisy signals makes it possible to simplify calculations, move to a dimensionless value, and thus compare processes with different values of variances. It is demonstrated in this paper that the obtained results still contain significant errors from the noise even when the classical conditions are satisfied. The technology for corrected calculation of normalized correlation functions of noisy signals is proposed in the paper, which makes it possible to significantly reduce the errors from the noise even if the classical conditions are not satisfied.
Keywords: random signal, noise, noisy signal, normalized estimates of the correlation function, robust estimates