Abstract
The results of the control-theoretic analysis of the least-mean-square (LMS) algorithm proposed by the author (1990) to include the case of nonwhite input data are extended and redefined. It is shown that, under some mild assumptions, the contribution of the cross power spectra (between the plant variation and the gradient noise) to the total error is negligible. >
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have