Title :
The λ²-LMS algorithm for separating sinusoids using the intermediate-converged adaptive filter solution
Author :
Glover, John R., Jr. ; Wang, Cheng-Xun
Author_Institution :
University of Houston, Houston, TX, USA
fDate :
4/1/1982 12:00:00 AM
Abstract :
A new procedure is proposed for automatically separating sinusoids of different, but unknown, frequencies, based only on the fact that they differ in amplitude. The procedure involves taking the intermediate solution from the LMS adaptive solution rather than the converged steady-state solution. In addition, a new algorithm which converges at a rate proportional to the squares of the signal eigenvalues is described and gives superior performance when taking the intermediate adaptive solution.
Keywords :
Adaptive filters; Adaptive signal detection; Convergence; Decorrelation; Delay; Eigenvalues and eigenfunctions; Interference cancellation; Least squares approximation; Narrowband; Noise cancellation;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
DOI :
10.1109/TASSP.1982.1163871