Proportionate adaptive algorithms for network echo cancellation Export

@article{Deng06, abstract = {By analyzing the coefficient adaptation process of the steepest descent algorithm, the condition under which the fastest overall convergence will be achieved is obtained and the way to calculate optimal step-size control factors to satisfy that condition is formulated. Motivated by the results and using the stochastic approximation paradigm, the /spl mu/-law PNLMS (MPNLMS) algorithm is proposed to keep, in contrast to the proportionate normalized least-mean-square (PNLMS) algorithm, the fast initial convergence during the whole adaptation process in the case of sparse echo path identification. Modifications of the MPNLMS algorithm are proposed to lower the computational complexity.}, author = {Deng, Hongyang and Doroslovacki, M.}, booktitle = {Signal Processing, IEEE Transactions on [see also Acoustics, Speech, and Signal Processing, IEEE Transactions on]}, citeulike-article-id = {2767471}, citeulike-linkout-0 = {http://dx.doi.org/10.1109/TSP.2006.872533}, citeulike-linkout-1 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1621408}, doi = {10.1109/TSP.2006.872533}, journal = {Signal Processing, IEEE Transactions on [see also Acoustics, Speech, and Signal Processing, IEEE Transactions on]}, keywords = {repeater}, number = {5}, pages = {1794--1803}, posted-at = {2008-05-07 22:49:00}, priority = {0}, title = {Proportionate adaptive algorithms for network echo cancellation}, url = {http://dx.doi.org/10.1109/TSP.2006.872533}, volume = {54}, year = {2006} }

TY - JOUR ID - Deng06 L3 - citeulike-article-id:2767471 N2 - By analyzing the coefficient adaptation process of the steepest descent algorithm, the condition under which the fastest overall convergence will be achieved is obtained and the way to calculate optimal step-size control factors to satisfy that condition is formulated. Motivated by the results and using the stochastic approximation paradigm, the /spl mu/-law PNLMS (MPNLMS) algorithm is proposed to keep, in contrast to the proportionate normalized least-mean-square (PNLMS) algorithm, the fast initial convergence during the whole adaptation process in the case of sparse echo path identification. Modifications of the MPNLMS algorithm are proposed to lower the computational complexity. IS - 5 JF - Signal Processing, IEEE Transactions on [see also Acoustics, Speech, and Signal Processing, IEEE Transactions on] EP - 1803 TI - Proportionate adaptive algorithms for network echo cancellation VL - 54 T2 - Signal Processing, IEEE Transactions on [see also Acoustics, Speech, and Signal Processing, IEEE Transactions on] SP - 1794 KW - repeater AU - Deng, Hongyang AU - Doroslovacki, M PY - 2006/// UR - http://dx.doi.org/10.1109/TSP.2006.872533 ER -