Analytical Performance of the LMS Algorithm on the Estimation of Wide Sense Stationary Channels

Estimation of an asymptotically stationary AR channel using fBm based LMS algorithm

“…3, it is evident that a lower value of MSWE can be achieved with the FB-VSLMS algorithm as compared to the approach presented in [4].…”

“…It is to note that for asymptotic stationarity of an AR process |p i |<1 for all i and because the step-size parameter (γ) is also small in the applications of wireless communications [4], we assume…”

“…In this paper, we have investigated the performance of this algorithm to estimate or track a class of time-varying channels that are represented by asymptotically stationary AR channels. These AR channels [3] are important because Rayleigh channel model is generally characterized by an asymptotically stationary AR-2 channel model [4].…”

A novel variable step size adjustment method based on channel output autocorrelation for the LMS training algorithm

“…., L, in update Equation (4) and is quite sensitive to k , the step-size parameter defines the convergence rate of the estimation algorithm. The single step size for all coefficients is the most critical part of the LMS algorithm that it is subjected to many studies to adjust the step size adaptively [8][9][10][11][12][13][14][15][16][17][18]. In order to guarantee a stable operation in all VSS-LMS algorithms, a sufficient condition for the step-size parameter is [4,9] 0< k < 2 3tr[R] (5) where R is the input autocorrelation matrix.…”

“…They define the optimum step size opt as the step size value that minimizes the maximum absolute Eigen value of the transition matrix A for a particular mean square analysis. Galdino et al [18] analytically evaluated the performance of the LMS algorithm on the estimation of time-varying channels, using the estimation error correlation matrix, the mean square weight error (MSWE) and the mean square estimation error (MSE) as parameters. Galdino et al [18] analytically evaluated the performance of the LMS algorithm on the estimation of time-varying channels, using the estimation error correlation matrix, the mean square weight error (MSWE) and the mean square estimation error (MSE) as parameters.…”