A Study in the Effect of Step-Size of LMS Algorithm in Sparse Channel Estimation

Abstract

LMS algorithm for the estimation of a channel which impulse response is characterized by a large number of inactive tap coefficients is used as a base to this paper due to its relatively low computational complexity, good stability properties, and relatively good robustness against implementation errors.Detection, including tap decoupling technique is incorporated with LMS algorithm under colored input conditions to improve and increase the performance of LMS algorithm for channel estimation .This technique allows discrimination between active and inactive taps of the unknown channel and reverses the effect of the colored input signal. The main focuses of this paper is on studying the effect of the step-size (μ) of LMS algorithm, and the tap decoupling technique under colored input conditions.