TY - GEN
T1 - Convergence analysis of the NSRLMMN algorithm
AU - Ulla Faiz, Mohammed Mujahid
AU - Zerguine, Azzedine
PY - 2012
Y1 - 2012
N2 - In this work, the ε-normalized sign regressor least mean mixed-norm (NSRLMMN) adaptive algorithm is proposed. The proposed algorithm exhibits increased convergence rate as compared to the least mean mixed-norm (LMMN) and the sign regressor least mean mixed-norm (SRLMMN) algorithms. Also, the steady-state analysis and convergence analysis are presented. Moreover, the proposed ε-NSRLMMN algorithm substantially reduces the computational load, a major drawback of the ε-normalized least mean mixed-norm (NLMMN) algorithm. Finally, simulation results are presented to support the theoretical findings.
AB - In this work, the ε-normalized sign regressor least mean mixed-norm (NSRLMMN) adaptive algorithm is proposed. The proposed algorithm exhibits increased convergence rate as compared to the least mean mixed-norm (LMMN) and the sign regressor least mean mixed-norm (SRLMMN) algorithms. Also, the steady-state analysis and convergence analysis are presented. Moreover, the proposed ε-NSRLMMN algorithm substantially reduces the computational load, a major drawback of the ε-normalized least mean mixed-norm (NLMMN) algorithm. Finally, simulation results are presented to support the theoretical findings.
KW - Adaptive filters
KW - LMF
KW - LMS
KW - Least Mean Mixed-Norm (LMMN)
KW - Sign regressor LMMN algorithm
UR - https://www.scopus.com/pages/publications/84869752894
M3 - Conference contribution
AN - SCOPUS:84869752894
SN - 9781467310680
T3 - European Signal Processing Conference
SP - 235
EP - 239
BT - Proceedings of the 20th European Signal Processing Conference, EUSIPCO 2012
ER -