This paper describes an improved version of the previously proposed fast converging algorithm, the proportionate normalized least mean squares (PNLMS), for network echo cancellers. We introduce a simple analysis of the PNLMS convergence behavior to show why after the fast initial convergence, it slows down. Also, the method has worked out to overcome this deficiency is presented. The improvement is shown by simulation results.
Từ khóa
#Echo cancellers #Equations #Convergence #Signal analysis #Algorithm design and analysis #Stability #Speech #Digital signal processing #Eigenvalues and eigenfunctions #Adaptive filters
Tài liệu tham khảo
haykin, 1996, Adaptive Filter Theory
10.1109/78.236504
gay, 1998, An efficient, fast converging adaptive filtcr for network echo ’, Proc Assilomar Conf
10.1109/89.861368