An Uncorrelated Variable Step-Size LMS Adaptive Algorithm
In this paper, the authors discuss and analyze two variable step-size adaptive algorithms in which the performance is reduced when input signal are highly correlated with each other. As a result, they propose an uncorrelated variable step-size LMS adaptive algorithm in order to overcome that shortcoming. The main idea of this algorithm is to pre-treat input signal by de-correlation then use this pretreated signal to update weight vector of the adaptive filter. Due to the theory of correlation, the algorithm can achieve good convergence rate. Finally, the simulations are consistent with the theoretical analysis.