A novel variable step size LMS algorithm based on decorrelation

With highly correlated input signal, the variable step size (VSS) algorithms always suffer from a low convergence speed. One way to overcome this problem is to decorrelate the input signal before passing it to the equalizer. In this paper, we propose a novel variable step size least mean square (LMS) algorithm based on decorrelation. In this algorithm, the update of the coefficients is matched to the correlation properties of input. The computer simulation results are consistent with the theoretical analysis and also show that the algorithm proposed can achieve a faster convergence rate and a smaller mean square error (MSE).