A robust step size adaptation scheme for LMS adaptive filters

It is shown that the known gradient step size adaptation algorithm for adaptive filters has a limited application area: it is disabled when the conditions of an independence of input signal samples are violated. The reason of this statement is the nonequivalence of the mean square error (MSE) and output error minimization relatively to the step size for correlated input samples. A robust algorithm for an adaptive step size is proposed. Its efficiency is demonstrated by simulations.

[1]  B. Widrow,et al.  Stationary and nonstationary learning characteristics of the LMS adaptive filter , 1976, Proceedings of the IEEE.

[2]  V. J. Mathews,et al.  Stochastic gradient adaptive filters with gradient adaptive step sizes , 1990, International Conference on Acoustics, Speech, and Signal Processing.

[3]  V. John Mathews,et al.  A stochastic gradient adaptive filter with gradient adaptive step size , 1993, IEEE Trans. Signal Process..

[4]  Scott C. Douglas.  Generalized gradient adaptive step sizes for stochastic gradient adaptive filters , 1995, 1995 International Conference on Acoustics, Speech, and Signal Processing.

[5]  Anthony G. Constantinides,et al.  The LMK algorithm with time-varying forgetting factor for adaptive system identification in additive output-noise , 1996, 1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings.

[6]  Ling Qin,et al.  Convergence analysis of a variable step-size normalized LMS adaptive filter algorithm , 1996, 1996 8th European Signal Processing Conference (EUSIPCO 1996).