Title :
Step-size control for acoustic echo cancellers applying an RBF network
Author_Institution :
Signal Theory, Darmstadt University of Technology, Merckstr. 25, D-64283 Darmstadt, Germany
Abstract :
The control of an acoustic echo canceller (AEC) is an essential part of hands-free telephone sets. Due to the fact that no single estimator is yet known to reliably control the AEC, various estimators should be implemented. Nevertheless, the combination of several estimators is quite difficult and usually determined heuristically. In this paper, an approach for automatic combination of estimators, based on a radial-basis function (RBF) network, is presented. Expert knowledge of estimators is no longer required and the heuristics involved are clearly reduced. Results of the proposed method in comparison to a conventional method are shown at the end of the paper.
Keywords :
Acoustics; Kernel; Noise measurement; Radial basis function networks; Reliability; Training; Vectors;
Conference_Titel :
Signal Processing Conference, 2000 10th European
Print_ISBN :
978-952-1504-43-3