Algoritmo com Ganho Proporcional ao Desvio Médio Quadrático dos Coeficientes Baseado na Autocorrelação do Erro

Fábio Luis Perez, Francisco das Chagas de Souza, Rui Seara

DOI: 10.14209/sbrt.2012.147

Evento: XXX Simpósio Brasileiro de Telecomunicações (SBrT2012)

Keywords: Error autocorrelation echo cancellation adaptive filtering system identification.

Abstract

This paper presents a new adaptive algorithm alternative to the mean-square weight deviation-proportionate ( -proportionate) z2 algorithm. The latter combines a mean-square weight deviation-proportionate gain distribution with a uniform one. Such a strategy is strongly dependent on the measurement noise power level that, in practice, is not readily available. Aiming to circumvent such dependence, a new strategy of gain distribution based on error autocorrelation is discussed here. The proposed algorithm outperforms the original one (z2-proportionate). Through simulation results, the effectiveness of the new algorithm is assessed.

Download