Title :
On the performance of adaptive sign algorithm with smoothed gradient estimate
Author_Institution :
Zhuhai Educ. Coll., Guangdong, China
Abstract :
The effects of the smoothed gradient estimate on the performance of the adaptive sign algorithm applied to adaptive echo cancelling are discussed in this paper. One of the smoothing schemes of the gradient estimate is analyzed in detail including the mean and the second moment convergence behavior of weights. It indicates that the smoothed gradient estimate does enlarge the allowable adaptive step size and improve the steady state error without affecting the convergence speed. Whether to adopt the gradient smoothing in practice depends on the conclusion by weighing the performance improved and the amount of computation added
Keywords :
adaptive filters; adaptive signal processing; convergence of numerical methods; echo suppression; gradient methods; parameter estimation; smoothing methods; statistical analysis; adaptive echo cancelling; adaptive sign algorithm; adaptive step size; mean; performance; second moment convergence; smoothed gradient estimate; steady state error; Adaptive filters; Adaptive systems; Convergence; Educational institutions; Least squares approximation; Noise cancellation; Signal processing algorithms; Smoothing methods; State estimation; Steady-state;
Conference_Titel :
Signal Processing Proceedings, 1998. ICSP '98. 1998 Fourth International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-4325-5
DOI :
10.1109/ICOSP.1998.770318
