Abstract—Adaptive algorithms used in acoustic echo
cancellers (AEC) must present high convergence speed and low
misadjustment, therefore algorithms like the affine projection
(APA) are frequently used, nevertheless its high computational
complexity could be a problem for hardware implementations.
In this paper, we present an acoustic echo canceller using the
Error Coded Affine Projection Algorithm, which is a
modification that assigns a code to the error and establishes a
threshold to update the filter coefficients, in such way the
coefficient update is not performed during each iteration.
Results show the efficiency of the algorithm in terms of good
convergence behavior, final misadjustment and a considerable
reduced number of updates, making this a great option for AEC
implementations.
Index Terms—Acoustic echo canceller, Adaptive filters,
Affine projection algorithm, Computational cost.
The authors are with the Instituto Politécnico Nacional, ESIME
Culhuacan, Mexico City, Mexico (e-mail: javaloso@ipn.mx,
fserranoo@ipn.mx, gavalos580@hotmail.com).
[PDF]
Cite:J. G. Avalos-Ochoa, F. A. Serrano-Orozco, and G. Avalos-Arzate, "Acoustic Echo Canceller Using the Error Coded Affine Projection Algorithm," International Journal of Information and Electronics Engineering vol. 6, no. 5, pp. 317-320, 2016.