Title :
Optimal Robust Adaptive LMS Algorithm without Adaptation Step-Size
Author :
Abu-Ella, Omar Ali ; El-Jabu, Bashir
Author_Institution :
Dept. of Elec. Eng., Univ. of the 7th of October, Misurata
Abstract :
The main problem in the LMS algorithm is to find the most suitable step-size. Here, we propose an algorithm to solve this. While the conventional LMS uses a pre-determined constant step-size, this algorithm uses stimulating input data and error sample to determine the optimum step-size, a method that makes the algorithm more adapted with time variant channel, improves the convergence characteristics, and utilizes the training period without any increase in the complexity cost.
Keywords :
direction-of-arrival estimation; least mean squares methods; time-varying channels; LMS algorithm; complexity cost; error sample; optimal robust adaptive algorithm; optimum step-size; pre-determined constant; time variant channel; Adaptive control; Array signal processing; Communication industry; Convergence; Cost function; Least squares approximation; Programmable control; Robustness; Shape control; Telecommunications; Adaptive Beamforming; Beamforming; LMS; Optimum Step-Size;
Conference_Titel :
Millimeter Waves, 2008. GSMM 2008. Global Symposium on
Conference_Location :
Nanjing
Print_ISBN :
978-1-4244-1885-5
Electronic_ISBN :
978-1-4244-1886-2
DOI :
10.1109/GSMM.2008.4534613
