New variant of NLMS/F algorithm with low computational cost
Adaptive filters are used in a wide variety of signal processing applications (e.g., acoustic echo cancellation, system identification, channel equalization, etc.). Adaptive algorithms are an essential part of adaptive filters since they update the filter coefficients to model the desired response....
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Artículo publishedVersion |
| Lenguaje: | Español |
| Publicado: |
FIUBA
2022
|
| Materias: | |
| Acceso en línea: | http://elektron.fi.uba.ar/index.php/elektron/article/view/163 http://repositoriouba.sisbi.uba.ar/gsdl/cgi-bin/library.cgi?a=d&c=elektron&d=163_oai |
| Aporte de: |
| Sumario: | Adaptive filters are used in a wide variety of signal processing applications (e.g., acoustic echo cancellation, system identification, channel equalization, etc.). Adaptive algorithms are an essential part of adaptive filters since they update the filter coefficients to model the desired response. Therefore, adaptive algorithms must have low computational cost and high speed of convergence. In this paper, a new variant of the Normalized Least-Mean-Fourth (NLMF) algorithm based on set membership is presented, in addition, a method to automatically adjust the step size is presented. To evaluate its performance, the algorithm was simulated in system identification and acoustic echo cancellation applications. The results demonstrate that the proposed algorithm improves the convergence speed and exhibits low computational cost compared to the conventional NLMS/F algorithm. |
|---|