Title :
A family of pseudo-least squares estimation algorithms without division
Author :
Ling, Fuyun ; Manolakis, Dimitris ; Proakis, John G.
Author_Institution :
Codex Corporation, Marshfield, MA
Abstract :
In this paper, we show that it is possible to approximate true division by dividing a number by a power-of-two in the LS lattice algorithm that is based on the new direct-coefficient-update form of the exact least squares (LS) lattice algorithm. The resulting pseudo-LS lattice algorithm is modestly slower in initial convergence than the true LS lattice algorithm, while it has a steady state error similar to the latter. The results are extended to a family of LS estimation algorithms covering a broader scope of applications.
Keywords :
Adaptive algorithm; Adaptive filters; Degradation; Digital signal processing; Filtering algorithms; Lattices; Least squares methods; Modems; Signal processing algorithms; Steady-state;
Conference_Titel :
Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP '86.
DOI :
10.1109/ICASSP.1986.1168747