Title :
Efficient μ-law improved proportionate affine projection algorithm for echo cancellation
Author :
Yang, Jian ; Sobelman, Gerald Edward
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Minnesota, Minneapolis, MN, USA
fDate :
1/1/2011 12:00:00 AM
Abstract :
For echo cancellation, the proportionate affine projection algorithm (APA) has been proposed to take advantage of the sparseness of the echo path. In this reported work, the μ-law proportionate technique is applied to the APA combined with a recently proposed structure having a memory of the proportionate step-size control matrices. Simulation results show that this new algorithm, termed the μ-law memorised improved proportionate APA (MMIPAPA), as well as its segmented implementation (SMIPAPA), yields a further performance improvement while maintaining low computational complexity.
Keywords :
adaptive filters; computational complexity; echo suppression; μ-law improved proportionate affine projection; adaptive filtering; computational complexity; echo cancellation; proportionate step-size control matrices;
Journal_Title :
Electronics Letters
DOI :
10.1049/el.2010.7937