Comparison of LMS and NLMS adaptive filters with a non-stationary input

The tracking performances of the LMS and NLMS algorithms are compared when the input of the adaptive filter is nonstationary. The analysis is done in the context of tracking a Markov plant. A periodic scenario of the variation of the input power is considered. The steady-state peak mean square deviation is used as the tracking performance measure. It is found that one algorithm outperforms the other depending on the values of the rate of variation of the input power, the minimum input power, the noise variance, and the mean square plant parameter increments.