Abstract
In this paper, the proportionate normalized least mean square (PNLMS) and its variants, µ-law PNLMS (MPNLMS), improved PNLMS (IPNLMS), and filter PNLMS (FPNLMS) algorithms are simulated for different types of input signals for a sparse system identification problem. The input signals such as a Gaussian random signal, band-limited signal, a speech signal, a color signal, and a uniform random signal are analyzed in this paper. The input is compressed by a wavelet transform before application. The simulations are carried out to examine the behavior of Gaussian noise on a sparse physical system. The simulation further justifies the robustness of the FPNLMS algorithm to the change in the input signal and noise level.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have