Algoritmo de mínimos cuadrados con error codificado para filtrado adaptivo

Analysis of a modified least mean square algorithm to reduce its design complexity for digital adaptive filter implementation, is presented. Such a modification consists of changing the error codification of the algorithm, because such error is an integer number. Results obtained during the testing of the method in applications to adaptive filters such as linear prediction and system identifier, show that the speed convergence increases. This allows the algorithm to be applied to adaptive filter for which high speed is required. Also, the proposed modification is compatible with existing adaptive filters, since error codification can be separately done.