Abstract: The householder RLS (HRLS) algorithm is an O(N2)
algorithm which recursively updates an arbitrary square-root of the
input data correlation matrix and naturally provides the LS weight
vector. A data dependent householder matrix is applied for such
an update. In this paper a recursive estimate of the eigenvalue
spread and misalignment of the algorithm is presented at a very low
computational cost. Misalignment is found to be highly sensitive to
the eigenvalue spread of input signals, output noise of the system and
exponential window. Simulation results show noticeable degradation
in the misalignment by increase in eigenvalue spread as well as
system-s output noise, while exponential window was kept constant.
Abstract: We present a discussion of three adaptive filtering
algorithms well known for their one-step termination property, in
terms of their relationship with the minimal residual method. These
algorithms are the normalized least mean square (NLMS), Affine
Projection algorithm (APA) and the recursive least squares algorithm
(RLS). The NLMS is shown to be a result of the orthogonality
condition imposed on the instantaneous approximation of the Wiener
equation, while APA and RLS algorithm result from orthogonality
condition in multi-dimensional minimal residual formulation. Further
analysis of the minimal residual formulation for the RLS leads to
a triangular system which also possesses the one-step termination
property (in exact arithmetic)
Abstract: In this paper an efficient incomplete factorization preconditioner is proposed for the Least Mean Squares (LMS) adaptive filter. The proposed preconditioner is approximated from a priori knowledge of the factors of input correlation matrix with an incomplete strategy, motivated by the sparsity patter of the upper triangular factor in the QRD-RLS algorithm. The convergence properties of IPLMS algorithm are comparable with those of transform domain LMS(TDLMS) algorithm. Simulation results show efficiency and robustness of the proposed algorithm with reduced computational complexity.