Análise em Regime Permanente de um Algoritmo NLMS de Passo Variável
Eduardo Vinicius Kuhn, José Gil Fausto Zipf, Rui Seara
Evento: XXXV Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2017)
Keywords: Variable step-size NLMS algorithm adaptive filtering stochastic modeling
AbstractThis paper focuses on the steady-state analysis of a variable step-size normalized least-mean-square (VSS-NLMS) algorithm from the literature, which uses a step-size adjustment rule with high immunity against the measurement noise. Specifically, model expressions are derived aiming to predict the excess mean-square error (EMSE) as well as the mean and mean squared step-size values in steady state. Based on the model expressions, some important characteristics of the algorithm are discussed. Simulation results are presented confirming the accuracy of the obtained model expressions.