convergence rate; vector correlation matrix; input signal vector
The standard LMS algorithms for adaptive beamforming though straightforward, suffer from the main draw back of having slow, convergence rate. This is normally attributed to the wide range of spread of eigenvalues of the input signal vector correlation matrix. By properly pre-whitening the input signal vector these eigenvalues can be equalized thereby improving the convergence rate. An Escalator structure recently proposed for linear prediction problems is used to pre-whiten the input signal vector. This whitened signal vector is now fed to a beam former with desired main beam shape. The overall adaptive array is computationally efficient and exhibits faster convergence rate than the conventional LMS algorithms.